This section reproduces the release notes for new features and incompatible changes in prior releases of Apache Kudu.
The list of known issues and limitations for prior releases are not reproduced on this page. Please consult the documentation of the appropriate release for a list of known issues and limitations. |
Support for Python 2.x and Python 3.4 and earlier is deprecated and may be removed in the next minor release.
Clients can now require authentication and encryption instead of depending on server-side settings (KUDU-1921).
Kudu Masters now automatically attempt to add themselves to an existing cluster if there is a healthy Raft quorum among Kudu Masters.
A new tool kudu master unsafe_rebuild
is added to reconstruct the master
catalog from tablet metadata collected from tablet servers. This can be used
in emergencies to restore access to tables when all masters are unavailable.
A new tool kudu table set_replication_factor
is added to alter the
replication factor of a table. The tool immediately updates table metadata in
the master, and the master will asynchronously effect the new replication
factor. Progress can be monitored by running ksck
.
It’s now possible to require a minimum replication factor for a Kudu table.
This can be achieved by customizing the setting for the newly introduced
--min_num_replicas
kudu-master’s flag. For example, setting
--min_num_replicas=3
enforces every newly created table to have at least 3
replicas for each of its tablets, so there cannot be a data loss when just a
single tablet server in the cluster fails irrecoverably. For the sake of
backward compatibility, --min_num_replicas
is set to 1 by default.
It’s now possible to track startup progress on the /startup
page on the web
UI. There are also metrics added to track the overall server startup progress
as well as the processing of the log block containers and starting of the
tablets (KUDU-1959).
A new tool kudu table add_column
is added to add columns to existing tables
using the CLI
(KUDU-3339).
A new tool kudu tserver unregister
is added to remove a dead tablet server
from the cluster without restarting the masters
(KUDU-2915).
Kudu will now more aggressively fsync consensus-related metadata when metadata is configured to be on an XFS mount. This may lead to increased contention on the device that backs metadata, but will prevent corruption in the event of an outage (KUDU-2195).
A clearer message is logged when the Ranger subprocess crashes, to specify a problem with the Ranger client.
Two new flags have been introduced for the kudu table scan
and kudu perf
table_scan
CLI tools: --row_count_only
and --report_scanner_stats
. With
these new flags, the above mentioned CLI tools allow to issue scan requests
equivalent to running “SELECT COUNT(1) FROM <table_name>” from impala-shell.
These new provisions are useful in detecting and troubleshooting scan
performance issues.
Added replica selection configuration knob for the kudu table scan
and kudu
perf table_scan
CLI tools: it’s controlled by the --replica_selection
flag.
To improve security, the following flags are now marked as sensitive and will be redacted in the logs and WebUI when the redaction is enabled:
--webserver_private_key_file
--webserver_private_key_password_cmd
--webserver_password_file
The logic to select the effective time source when running with
--time_source=auto
has been updated. The builtin
time source would be
auto-selected if a Kudu server runs with --time_source=auto
in an
environment where the instance detector isn’t aware of dedicated NTP servers
AND the --builtin_ntp_servers
flag is set to a valid value. Otherwise, if
--builtin_ntp_servers
flag is set to an empty or invalid value, the
effective time source becomes system
for platforms supporting the
get_ntptime()
API, otherwise the catch-all case selects the system_unsync
as the time effective source.
It is now possible to print or edit PBC files in batch using the kudu pbc
CLI tool, and also to format its JSON input/output as “pretty”.
Client connection timeout is now configurable in the Java client (KUDU-3240).
A new /healthz
endpoint is now available on the kudu-master and tablet-server
embedded web servers for liveness checks
(KUDU-3308).
Hive Metastore URI is now logged to the console when connecting via kudu hms
CLI tool (KUDU-3189).
It is now possible to start up a master when there is an additional master address present in the master addresses flag (KUDU-3311).
Table entity is now accessible in KuduWriteOperation
in the C++ client,
making understanding errors on the client side easier
(KUDU-2623).
The rebalancer tool now doesn’t move replicas to tablet servers in maintenance mode (KUDU-3328).
Improved the performance of the run length encoding (RLE).
Log4J used in Ranger subprocess was upgraded to 2.17.1 which contains patches go several security vulnerabilities (CVE-2021-44832, CVE-2021-45105, CVE-2021-45046, and CVE-2021-44228).
Kudu servers previously crashed if hostnames became unresolvable via DNS (e.g. if the container hosting a server were destroyed). Such errors are now treated as transient and the lookups are retried periodically. See KUDU-75, KUDU-1620, and KUDU-1885 for more details.
Fixed an issue in Kudu Java client where concurrent flushing of data buffers could lead to errors reported as 'java.lang.AssertionError: This Deferred was already called' (KUDU-3277).
Fixed Kudu RPC negotiation issue when running with cyrus-sasl-gssapi-2.1.27-5
and newer versions of the RPM package. A failed RPC connection negotiation
attempt would result in an error logged along with the full connection
negotiation trace: Runtime error: SASL(-15): mechanism too weak for this
user: Unable to find a callback: 32775
(KUDU-3297).
Fixed crash in kudu-master and kudu-tserver when running with kernel where the
getrandom(2)
API is not available (versions of Linux kernel prior to 3.17).
Fixed bug which could lead to exhaustion of the address space for the outgoing connections on a busy Kudu cluster (KUDU-3352).
Fixed a bug in the Java client where a malformed tablet server ID in the scan token causes connection failures and timeouts in some cases (KUDU-3349).
Fixed a bug where the rebalancer failed with -ignored_tservers
flag
(KUDU-3346).
Kudu 1.16.0 is wire-compatible with previous versions of Kudu:
Kudu 1.16 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Rolling upgrade between Kudu 1.15 and Kudu 1.16 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
Kudu 1.0 clients may connect to servers running Kudu 1.16 with the exception of the below-mentioned restrictions regarding secure clusters.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.16 and versions earlier than 1.3:
If a Kudu 1.16 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.16 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
The Kudu 1.16 Java client library is API- and ABI-compatible with Kudu 1.15. Applications written against Kudu 1.15 will compile and run against the Kudu 1.16 client library and vice-versa.
The Kudu 1.16 C++ client is API- and ABI-forward-compatible with Kudu 1.15. Applications written and compiled against the Kudu 1.15 client library will run without modification against the Kudu 1.16 client library. Applications written and compiled against the Kudu 1.16 client library will run without modification against the Kudu 1.15 client library.
The Kudu 1.16 Python client is API-compatible with Kudu 1.15. Applications written against Kudu 1.15 will continue to run against the Kudu 1.16 client and vice-versa.
Please refer to the Known Issues and Limitations section of the documentation.
Kudu 1.16.0 includes contributions from 17 people, including 5 first-time contributors:
Riza Suminto
Zoltan Chovan
kedeng
khazarmammadli
yejiabao
Thank you for your contributions!
The kudu-mapreduce
integration has been removed in the 1.15.0 release.
Similar functionality and capabilities now exist via the Apache Spark, Apache Hive, Apache Impala,
and Apache NiFi integrations. See KUDU-3142
for details.
Support for Python 2.x and Python 3.4 and earlier is deprecated and may be removed in the next minor release.
Kudu now experimentally supports multi-row transactions. Currently only INSERT
and
INSERT_IGNORE
operations are supported.
See here for a
design overview of this feature.
Kudu now supports Raft configuration change for Kudu masters and CLI tools for orchestrating addition and removal of masters in a Kudu cluster. These tools substantially simplify the process of migrating to multiple masters, recovering a dead master and removing masters from a Kudu cluster. For detailed steps, see the latest administration documentation. This feature is evolving and the steps to add, remove and recover masters may change in the future. See KUDU-2181 for details.
Kudu now supports table comments directly on Kudu tables which are automatically synchronized when the Hive Metastore integration is enabled. These comments can be added at table creation time and changed via table alteration.
Kudu now experimentally supports per-table size limits based on leader disk space usage or number
of rows. When generating new authorization tokens, Masters will now consider the size limits and
strip tokens of INSERT
and UPDATE
privileges if either limit is reached. To enable this feature,
set the --enable_table_write_limit
master flag; adjust the --table_disk_size_limit
and
--table_row_count_limit
flags as desired or use the kudu table set_limit
tool to set limits per
table.
It is now possible to change the Kerberos Service Principal Name using the --principal
flag. The
default SPN is still kudu/_HOST
. Clients connecting to a cluster using a non-default SPN must set
the sasl_protocol_name
or saslProtocolName
to match the SPN base
(i.e. “kudu” if the SPN is “kudu/_HOST”) in the client builder or the Kudu CLI.
See KUDU-1884 for details.
Kudu RPC now supports TLSv1.3. Kudu servers and clients automatically negotiate TLSv1.3 for Kudu
RPC if OpenSSL (or Java runtime correspondingly) on each side supports TLSv1.3.
If necessary, use the newly introduced flag --rpc_tls_ciphersuites
to customize TLSv1.3-specific
cipher suites at the server side.
See KUDU-2871 for details.
TLS ciphers renegotiation for TLSv1.2 and prior protocol versions is now explicitly disabled. See KUDU-1926 for details.
The location assignment for Kudu clients is now disabled by default since it doesn’t bring a lot
of benefits, but rather puts an extra load to Kudu masters. This change reduces the load on Kudu
masters which is essential if too many clients run in a cluster. To enable the location assignment
for clients, override the default by setting --master_client_location_assignment_enabled=true
for
Kudu masters.
The behavior of the C++ client replica selection for closest replica, the default, was updated to match the behavior of the Java client. Instead of picking a random replica each time, a static value is used for each process ensuring that the selection remains deterministic and can benefit from better caching. See KUDU-3248 for details.
The Web UI /rpcz endpoint now displays information on whether an RPC connection is protected by TLS, and if so, provides information on the negotiated TLS cipher suite.
Tooling requests and C++ client requests bound for leader masters will now be retried in the event the masters cannot be reached.
Cluster tooling will now validate that the master argument contains no duplicate values. See KUDU-3226 for details.
The error message output by Kudu Java client in an attempt to write into a non-existent table partition now contains the table’s name. See KUDU-3267 for details.
Fixed a bug in the Kudu tablet servers that could result in a crash when performing an incremental backup of rows that had many batches of updates. See KUDU-3291 for more details.
The Kudu Java client will now retry scans bound for tablets hosted on quiescing tablet servers at replicas on other tablet servers. See KUDU-3213 for more details.
Fixed a race between the scheduling of a maintenance op and the destruction of a tablet. This could previously lead to a crash. See KUDU-3268 for more details.
Fixed crash in Kudu C++ client introduced with KUDU-1802. See KUDU-3254 for details.
Fixed bug in Kudu Java client which manifested in AUTO_FLUSH_BACKGROUND
sessions hung in a call
to KuduSession.flush()
method. Another sign of the bug were stuck data ingest workloads based on
Java client (e.g., kudu-spark applications) with
"java.lang.AssertionError: This Deferred was already called!" message in the logs.
See KUDU-3277 for details.
Fixed crash in Kudu server due to lack of getrandom(2)
system call in Linux kernel version
earlier than 3.17 by instead using /dev/random
for uuid generation in the Boost library.
Crash includes the following message in the logs
"terminate called after throwing an instance of 'boost::wrapexcept<boost::uuids::entropy_error>'".
See the fix for a sample stack trace.
Kudu 1.15.0 is wire-compatible with previous versions of Kudu:
Kudu 1.15 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Rolling upgrade between Kudu 1.14 and Kudu 1.15 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
Kudu 1.0 clients may connect to servers running Kudu 1.15 with the exception of the below-mentioned restrictions regarding secure clusters.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.15 and versions earlier than 1.3:
If a Kudu 1.15 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.15 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
Updated hash computation for empty strings in the FastHash implementation to conform with the handling in Apache Impala. For Bloom filter predicate pushdown feature that uses FastHash, this makes the Kudu client older than version 1.15.0 incompatible with Kudu server version 1.15.0 and Kudu client version at or newer than 1.15.0 incompatible with Kudu server version earlier than 1.15.0. Both client library and Kudu server need to be updated to version 1.15.0 or above if using the Bloom filter predicate feature. One manifestation of this incompatibility is following message in the logs, "Not implemented: call requires unsupported application feature flags: 4". See KUDU-3286 for details.
The Kudu 1.15 Java client library is API- and ABI-compatible with Kudu 1.14. Applications written against Kudu 1.14 will compile and run against the Kudu 1.15 client library and vice-versa.
The Kudu 1.15 C++ client is API- and ABI-forward-compatible with Kudu 1.14. Applications written and compiled against the Kudu 1.14 client library will run without modification against the Kudu 1.15 client library. Applications written and compiled against the Kudu 1.15 client library will run without modification against the Kudu 1.14 client library.
The Kudu 1.15 Python client is API-compatible with Kudu 1.14. Applications written against Kudu 1.14 will continue to run against the Kudu 1.15 client and vice-versa.
Please refer to the Known Issues and Limitations section of the documentation.
Kudu 1.15.0 includes contributions from 12 people, including 2 first-time contributors:
Abhishek Chennaka
shenxingwuying
Thank you for your contributions!
Support for CentOS 6/RHEL 6, Ubuntu 14, Ubuntu 16, and Debian 8 platforms has been dropped given they are at or near end-of-life. We will no longer validate these platforms as a part of the release process, though patches will still be accepted going forward.
Developer support for OS X 10.10 Yosemite, OS X 10.11 El Capitan, and OS X 10.12 Sierra has been dropped. We will no longer validate these versions as a part of the release process, though patches will still be accepted going forward.
Support for Python 2.x and Python 3.4 and earlier is deprecated and may be removed in the next minor release.
The kudu-mapreduce
integration has been deprecated and may be removed in the
next minor release. Similar functionality and capabilities now exist via the
Apache Spark, Apache Hive, Apache Impala, and Apache NiFi integrations.
Full support for INSERT_IGNORE
, UPDATE_IGNORE
, and DELETE_IGNORE
operations
was added. The INSERT_IGNORE
operation will insert a row if one matching the key
does not exist and ignore the operation if one already exists. The UPDATE_IGNORE
operation will update the row if one matching the key exists and ignore the operation
if one does not exist. The DELETE_IGNORE
operation will delete the row if one matching
the key exists and ignore the operation if one does not exist. These operations are
particularly useful in situations where retries or duplicate operations could occur and
you do not want to handle the errors that could result manually or you do not want to cause
unnecessary writes and compaction work as a result of using the UPSERT
operation.
The Java client can check if the cluster it is communicating with supports these operations
by calling the supportsIgnoreOperations()
method on the KuduClient. See
KUDU-1563 for more details.
Spark 3 compatible JARs compiled for Scala 2.12 are now published for the Kudu Spark integration. See KUDU-3202 for more details.
Every Kudu cluster now has an automatically generated cluster Id that can be used to uniquely
identify a cluster. The cluster Id is shown in the masters web-UI, the kudu master list
tool,
and in master server logs. See KUDU-2574
for more details.
It is now possible to enforce that OpenSSL is initialized in FIPS approved mode in the servers and the C++ client by setting the KUDU_REQUIRE_FIPS_MODE environment variable to “1”, “yes” or “true”. See KUDU-3210 for more details.
Downloading the WAL data and data blocks when copying tablets to another tablet server is now parallelized, resulting in much faster tablet copy operations. These operations occur when recovering from a down tablet server or when running the cluster rebalancer. See KUDU-1728 and KUDU-3214 for more details.
The HMS integration now supports multiple Kudu clusters associated with a single HMS including Kudu clusters that do not have HMS synchronization enabled. This is possible, because the Kudu master will now leverage the cluster Id to ignore notifications from tables in a different cluster. Additionally, the HMS plugin will check if the Kudu cluster associated with a table has HMS synchronization enabled. See KUDU-3192 and KUDU-3187 for more details.
The HMS integration now supports gzipped HMS notifications. This is important in order to support Hive 4 where the default encoder was changed to be the GzipJSONMessageEncoder. See KUDU-3201 for more details.
Kudu will now fail tablet replicas that have been corrupted due to KUDU-2233 instead of crashing the tablet server. If a healthy majority still exists, a new replica will be created and the failed replica will be evicted and deleted. See KUDU-3191 and KUDU-2233 for more details.
DeltaMemStores will now be flushed as long as any DMS in a tablet is older than the point
defined by --flush_threshold_secs
, rather than flushing once every --flush_threshold_secs
period. This can reduce memory pressure under update- or delete-heavy workloads, and lower tablet
server restart times following such workloads. See
KUDU-3195 for more details.
The kudu perf loadgen
CLI tool now supports UPSERT
for storing the generated data into
the table. To switch to UPSERT
for row operations (instead of default INSERT
), add the
--use_upsert
command-line flag.
Users can now specify the level of parallelization when copying a tablet using the
kudu local_replica copy_from_remote
CLI tool by passing the
--tablet_copy_download_threads_nums_per_session
argument.
The Kudu Masters now discriminate between overlapped and exact duplicate key ranges when adding
new partitions, returning Status::AlreadyPresent()
for exact range duplicates and
Status::InvalidArgument()
for otherwise overlapped ones. In prior releases, the master
returned Status::InvalidArgument()
both in case of duplicate and otherwise overlapped ranges.
The handling of an empty list of master addresses in Kudu C++ client has improved. In prior
releases, KuduClientBuilder::Build()
would hang in ConnectToCluster()
if no master addresses
were provided. Now, KuduClientBuilder::Build()
instantly returns Status::InvalidArgument()
in such a case.
The connection negotiation timeout for Kudu C client is now programmatically configurable. To customize the connection negotiation timeout, use the newly introduced `KuduClientBuilder::connection_negotiation_timeout()` method in the Kudu C client API.
All RPC-related kudu
CLI tools now have --negotiation_timeout_ms
command line flag to
control the client-side connection negotiation timeout. The default value for the new flag is
set to 3000 milliseconds for backward compatibility. Keep in mind that the total RPC timeout
includes the connection negotiation time, so in general it makes sense to bump --timeout_ms
along with --negotiation_timeout_ms
by the same delta.
Kudu now reports on slow SASL calls (i.e. calls taking more than 250 milliseconds to complete) when connecting to a server. This is to help diagnose issues like described in KUDU-3217.
MaintenanceManager now has a new histogram-based maintenance_op_find_best_candidate_duration
metric to capture the stats on how long it takes (in microseconds) to find the best maintenance
operation among available candidates. The newly introduced metric can help in diagnosing
conditions where MaintenanceManager seems lagging behind the rate of write operations in a busy
Kudu cluster with many replicas per tablet server.
The KuduScanToken Java API has been extended with a deserializeIntoScannerBuilder()
method that
can be used to further customize generated tokens.
Logging of the error message produced when applying an op while a Java KuduSession is closed has been throttled. See KUDU-3012 for more details.
Added a new uptime
metric for a Kudu server. The metric’s value is reported as the length of
the time interval passed from the start of the server, in microseconds. Knowing the server’s
uptime, it’s easier to interpret and compare metrics reported by different Kudu servers.
Documentation for Kudu’s metrics are now automatically generated for each release and can be seen here.
Fixed lock contention between MaintenanceManager op registration and the scheduling of new maintenance ops. On particularly dense tablet servers, this contention was previously shown to significantly slow down startup times. See KUDU-3149 for more details.
Fixed lock contention between MaintenanceManager’s threads performing already scheduled operations and the scheduler thread itself. This benefits clusters with heavy ingest/update workloads that have many replicas per tablet server. See KUDU-1954 for more details.
Fixed a bug in the merge iterator that could result in a crash. This could surface as a crash when performing ordered or differential scans, particularly when the underlying data contained deletes and reinserts. See KUDU-3108 for more details.
Fixed a heap-use-after-free bug in Kudu C++ client that might manifest itself when altering a table to update the partitioning schema. See KUDU-3238 for more details.
Fixed a bug where building scan tokens would result in a NullPointerException if a tablet not found error occurred before generating the token. See KUDU-3205 for more details.
Fixed a bug where a delete operation would fail if the row being deleted contained exactly 64 columns and all values were set on the row. See KUDU-3198 for more details.
Fixed a bug where Slf4j classes were shaded into the Spark integration JARs. See KUDU-3157 for more details.
Fixed a bug where the 'kudu hms fix' tool mistakenly reports non-matching master addresses when the addresses are in-fact canonically the same. See KUDU-2884 for more details.
Kudu 1.14.0 is wire-compatible with previous versions of Kudu:
Kudu 1.14 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Rolling upgrade between Kudu 1.13 and Kudu 1.14 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
Kudu 1.0 clients may connect to servers running Kudu 1.14 with the exception of the below-mentioned restrictions regarding secure clusters.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.14 and versions earlier than 1.3:
If a Kudu 1.14 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.14 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
The Kudu 1.14 Java client library is API- and ABI-compatible with Kudu 1.13. Applications written against Kudu 1.13 will compile and run against the Kudu 1.14 client library and vice-versa.
The Kudu 1.14 C++ client is API- and ABI-forward-compatible with Kudu 1.13. Applications written and compiled against the Kudu 1.13 client library will run without modification against the Kudu 1.14 client library. Applications written and compiled against the Kudu 1.14 client library will run without modification against the Kudu 1.13 client library.
The Kudu 1.14 Python client is API-compatible with Kudu 1.13. Applications written against Kudu 1.13 will continue to run against the Kudu 1.14 client and vice-versa.
Please refer to the Known Issues and Limitations section of the documentation.
Kudu 1.14.0 includes contributions from 12 people, including 1 first-time contributors:
liguohao
Thank you for your contributions!
The Sentry integration has been removed and the Ranger integration should now be used in its place for fine-grained authorization.
Support for Python 2.x and Python 3.4 and earlier is deprecated and may be removed in the next minor release.
The kudu-mapreduce
integration has been deprecated and may be removed in the
next minor release. Similar functionality and capabilities now exist via the
Apache Spark, Apache Hive, Apache Impala, and Apache NiFi integrations.
Added table ownership support. All newly created tables are automatically owned by the user creating them. It is also possible to change the owner by altering the table. You can also assign privileges to table owners via Apache Ranger (see KUDU-3090).
An experimental feature is added to Kudu that allows it to automatically
rebalance tablet replicas among tablet servers. The background task can be
enabled by setting the --auto_rebalancing_enabled
flag on the Kudu masters.
Before starting auto-rebalancing on an existing cluster, the CLI rebalancer
tool should be run first (see
KUDU-2780).
Bloom filter column predicate pushdown has been added to allow optimized execution of filters which match on a set of column values with a false-positive rate. Support for Impala queries utilizing Bloom filter predicate is available yielding performance improvements of 19% to 30% in TPC-H benchmarks and around 41% improvement for distributed joins across large tables. Support for Spark is not yet available. (see KUDU-2483).
AArch64-based (ARM) architectures are now supported including published Docker images.
The Java client now supports the columnar row format returned from the server transparently. Using this format can reduce the server CPU and size of the request over the network for scans. The columnar format can be enabled via the setRowDataFormat() method on the KuduScanner.
An experimental feature that can be enabled by setting the
--enable_workload_score_for_perf_improvement_ops
prioritizes flushing and
compacting hot tablets.
Hive metastore synchronization now supports Hive 3 and later.
The Spark KuduContext accumulator metrics now track operation counts per table instead of cumulatively for all tables.
The kudu local_replica delete
CLI tool now accepts multiple tablet
identifiers. Along with the newly added --ignore_nonexistent
flag, this
helps with scripting scenarios when removing multiple tablet replicas from a
particular Tablet Server.
Both Master’s and Tablet Server’s web UI now displays the name for a service
thread pool group at the /threadz
page
Introduced queue_overflow_rejections_
metrics for both Masters and Tablet
Servers: number of RPC requests of a particular type dropped due to RPC
service queue overflow.
Introduced a CoDel-like queue control mechanism for the apply queue. This
helps to avoid accumulating too many write requests and timing them out in
case of seek-bound workloads (e.g., uniform random inserts). The newly
introduced queue control mechanism is disabled by default. To enable it, set
the --tablet_apply_pool_overload_threshold_ms
Tablet Server’s flag to
appropriate value, e.g. 250 (see
KUDU-1587).
Java client’s error collector can be resized (see KUDU-1422).
Calls to the Kudu master server are now drastically reduced when using scan tokens. Previously deserializing a scan token would result in a GetTableSchema request and potentially a GetTableLocations request. Now the table schema and location information is serialized into the scan token itself avoiding the need for any requests to the master when processing them.
The default size of Master’s RPC queue is now 100 (it was 50 in earlier releases). This is to optimize for use cases where a Kudu cluster has many clients working concurrently.
Masters now have an option to cache table location responses. This is
targeted for Kudu clusters which have many clients working concurrently. By
default, the caching of table location responses is disabled. To enable table
location caching, set the proper capacity of the table location cache using
Master’s --table_locations_cache_capacity_mb
flag (setting to 0 disables the
caching). Up to 17% of improvement is observed in GetTableLocations request
rate when enabling the caching.
Removed lock contention on Raft consensus lock in Tablet Servers while processing a write request. This helps to avoid RPC queue overflows when handling concurrent write requests to the same tablet from multiple clients (see KUDU-2727).
Master’s performance for handling concurrent GetTableSchema requests has been improved. End-to-end tests indicated up to 15% improvement in sustained request rate for high concurrency scenarios.
Kudu servers now use protobuf Arena objects to perform all RPC request/response-related memory allocations. This gives a boost for overall RPC performance, and with further optimization the result request rate was increased significantly for certain methods. For example, the result request rate increased up to 25% for Master’s GetTabletLocations() RPC in case of highly concurrent scenarios (see KUDU-636).
Tablet Servers now use protobuf Arena for allocating Raft-related runtime structures. This results in substantial reduction of CPU cycles used and increases write throughput (see KUDU-636).
Tablet Servers now use protobuf Arena for allocating EncodedKeys to reduce allocator contention and improve memory locality (see KUDU-636).
Bloom filter predicate evaluation for scans can be computationally expensive. A heuristic has been added that verifies rejection rate of the supplied Bloom filter predicate below which the Bloom filter predicate is automatically disabled. This helped reduce regression observed with Bloom filter predicate in TPC-H benchmark query #9 (see KUDU-3140).
Improved scan performance of dictionary and plain-encoded string columns by avoiding copying them (see KUDU-2844).
Improved maintenance manager’s heuristics to prioritize larger memstores (see KUDU-3180).
Spark client’s KuduReadOptions now supports setting a snapshot timestamp for repeatable reads with READ_AT_SNAPSHOT consistency mode (see KUDU-3177).
Kudu scans now honor location assignments when multiple tablet servers are co-located with the client.
Fixed a bug that caused IllegalArgumentException to be thrown when trying to create a predicate for a DATE column in Kudu Java client (see KUDU-3152).
Fixed a potential race when multiple RPCs work on the same scanner object.
Kudu 1.13.0 is wire-compatible with previous versions of Kudu:
Kudu 1.13 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Rolling upgrade between Kudu 1.12 and Kudu 1.13 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
Kudu 1.0 clients may connect to servers running Kudu 1.13 with the exception of the below-mentioned restrictions regarding secure clusters.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.13 and versions earlier than 1.3:
If a Kudu 1.13 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.13 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
The Kudu 1.13 Java client library is API- and ABI-compatible with Kudu 1.12. Applications written against Kudu 1.12 will compile and run against the Kudu 1.13 client library and vice-versa.
The Kudu 1.13 C++ client is API- and ABI-forward-compatible with Kudu 1.12. Applications written and compiled against the Kudu 1.12 client library will run without modification against the Kudu 1.13 client library. Applications written and compiled against the Kudu 1.13 client library will run without modification against the Kudu 1.12 client library.
The Kudu 1.13 Python client is API-compatible with Kudu 1.12. Applications written against Kudu 1.12 will continue to run against the Kudu 1.13 client and vice-versa.
Please refer to the Known Issues and Limitations section of the documentation.
Kudu 1.13.0 includes contributions from 22 people, including 9 first-time contributors:
Jim Apple
Kevin J McCarthy
Li Zhiming
Mahesh Reddy
Romain Rigaux
RuiChen
Shuping Zhou
ningw
wenjie
The Flume sink has been migrated to the Apache Flume project and removed
from Kudu. Users depending on the Flume integration can use the old
kudu-flume
jars or migrate to the Flume jars containing the Kudu sink.
Support for Apache Sentry authorization has been deprecated and may be removed in the next minor release. Users depending on the Sentry integration should migrate to the Apache Ranger integration for fine-grained authorization.
Support for Python 2 has been deprecated and may be removed in the next minor release.
Support for CentOS/RHEL 6, Debian 8, Ubuntu 14 has been deprecated and may be removed in the next minor release.
Kudu now supports native fine-grained authorization via integration with Apache Ranger. Kudu may now enforce access control policies defined for Kudu tables and columns stored in Ranger. See the authorization documentation for more details.
Kudu’s web UI now supports proxying via Apache Knox. Kudu may be deployed in a firewalled state behind a Knox Gateway which will forward HTTP requests and responses between clients and the Kudu web UI.
Kudu’s web UI now supports HTTP keep-alive. Operations that access multiple URLs will now reuse a single HTTP connection, improving their performance.
The kudu tserver quiesce
tool is added to quiesce tablet servers. While a
tablet server is quiescing, it will stop hosting tablet leaders and stop
serving new scan requests. This can be used to orchestrate a rolling restart
without stopping on-going Kudu workloads.
Introduced auto
time source for HybridClock timestamps. With
--time_source=auto
in AWS and GCE cloud environments, Kudu masters and
tablet servers use the built-in NTP client synchronized with dedicated NTP
servers available via host-only networks. With --time_source=auto
in
environments other than AWS/GCE, Kudu masters and tablet servers rely on
their local machine’s clock synchronized by NTP. The default setting for
the HybridClock time source (--time_source=system
) is backward-compatible,
requiring the local machine’s clock to be synchronized by the kernel’s NTP
discipline.
The kudu cluster rebalance
tool now supports moving replicas away from
specific tablet servers by supplying the --ignored_tservers
and
--move_replicas_from_ignored_tservers
arguments (see
KUDU-2914 for more
details).
The kudu table create
tool is added to allow users to specify table
creation options using JSON.
Kudu now supports DATE and VARCHAR data types. See the schema design documentation for more details.
Write Ahead Log file segments and index chunks are now managed by Kudu’s file cache. With that, all long-lived file descriptors used by Kudu are managed by the file cache, and there’s no longer a need for capacity planning of file descriptor usage.
Kudu no longer requires the running of kudu fs update_dirs
to change a
directory configuration or recover from a disk failure
(see KUDU-2993).
Kudu tablet servers and masters now expose a tablet-level metric
num_raft_leaders
for the number of Raft leaders hosted on the server.
Kudu’s maintenance operation scheduling has been updated to prioritize reducing WAL retention under memory pressure. Kudu would previously prioritize operations that yielded high-memory reduction, which could result in high WAL disk usage in workloads that contained updates (see KUDU-3002).
A new maintenance operation is introduced to remove rowsets that have had all of their rows deleted and whose newest delete operations are considered ancient (see KUDU-1625).
The built-in NTP client is now fully supported as the time source for Kudu’s
HybridTime clock, i.e. it’s no longer marked as experimental. To switch the
time source from the existing system
time source (which is the default)
to the built-in NTP client, use --time_source=builtin
.
Introduced additional metrics for the built-in NTP client (see KUDU-3048).
Updated /config
page of masters' and tablet servers' web UI to display
configured and effective time source. In addition, the effective list of
reference servers for the built-in NTP client is shown there as well,
if applicable.
chronyd
(version 3.4 and newer) is now supported as NTP server for
synchronizing the local machine’s clock in a Kudu cluster. It’s important to
have the rtcsync
option enabled in the configuration of the chronyd
NTP
daemon (see KUDU-2573).
Kudu now supports building and running on RHEL/CentOS 8. This has been tested with CentOS 8.1.
The processing of Raft consensus vote requests has been improved to be more robust during high contention scenarios like election storms.
Added a validator to enforce consistency between the maximum size of an RPC
and the maximum size of tablet transaction memory, controlled by
--rpc_max_message_size
and --tablet_transaction_memory
flags
correspondingly. In prior releases, if the limit on the size of RPC requests
is increased and the limit on tablet transaction memory size is kept with the
default setting, certain Raft transactions could be committed but not applied
(see KUDU-3023).
The metrics endpoint now supports filtering metrics by a metric severity level. See the documentation for more details.
Many kudu local_replica
tools are updated to not open the block manager,
which significantly reduces the amount of IO done when running them (see
KUDU-3070 for more
details).
The Kudu Java client now exposes a way to get the resource metrics associated with a given scanner (see KUDU-2162 for more details).
Scan predicates are pushed down to RLE decoders, improving predicate-evaluation-efficiency in some workloads (see KUDU-2852 for more details).
The log block manager will now attempt to use multiple threads to open blocks in each data directory, in some tests reducing startup time by up to 20% (see KUDU-2977 and KUDU-3001 for more details).
Kudu’s tablet server web UI scans page is updated to show the number of round trips per scanner.
Kudu’s master and tablet server web UIs are updated to show critical partition information, including tablet count and on-disk size.
Kudu servers now expose the last_read_elapsed_seconds
and
last_write_elapsed_seconds
tablet-level metrics that indicate how long ago
the most recent read and write operations to a given tablet were.
Kudu servers now expose the transaction_memory_limit_rejections
tablet-level metric that tracks the number of transactions rejected because a
given tablet’s transactional memory limit was reached (see
KUDU-3021 for more
details).
Fixed a bug in which Kudu would not schedule compactions if a server were under memory pressure (see KUDU-2929).
Fixed a bug where DDL operations like ALTER TABLE on tables with huge number of partitions might result in a DoS situation for Kudu masters (see KUDU-3036).
Fixed a bug where Kudu Java client cannot negotiate a secure connection with Kudu masters and tablet servers if using BouncyCastle JCE provider (see KUDU-3106).
Kudu masters will now crash immediately upon hitting a disk failure (see KUDU-2904 for more details).
Fixed an issue in the Kudu master in which delays in receiving tablet server heartbeats could result in an excess amount of RPC traffic between the masters and tablet servers (see KUDU-2992 for more details).
Fixed an issue with Kudu’s location placement policy that would place all replicas in one location when two locations were available (see KUDU-3008 for more details).
The Java client will now correctly propagate timestamps when sending write batches (see KUDU-3035 for more detail).
Fixed an issue with the Kudu backup Spark jobs in which Kudu would return with a non-zero exit if the job succeeded but backed up no rows (see KUDU-3099 for more details).
The raft_term
and time_since_last_leader_heartbeat
aggregated table
metrics will now return the maximum metric reported instead of the sum.
Kudu 1.12.0 is wire-compatible with previous versions of Kudu:
Kudu 1.12 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Rolling upgrade between Kudu 1.11 and Kudu 1.12 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
Kudu 1.0 clients may connect to servers running Kudu 1.12 with the exception of the below-mentioned restrictions regarding secure clusters.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.12 and versions earlier than 1.3:
If a Kudu 1.12 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.12 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
The Kudu 1.12 Java client library is API- and ABI-compatible with Kudu 1.11. Applications written against Kudu 1.11 will compile and run against the Kudu 1.12 client library and vice-versa.
The Kudu 1.12 C++ client is API- and ABI-forward-compatible with Kudu 1.11. Applications written and compiled against the Kudu 1.11 client library will run without modification against the Kudu 1.12 client library. Applications written and compiled against the Kudu 1.12 client library will run without modification against the Kudu 1.11 client library.
The Kudu 1.12 Python client is API-compatible with Kudu 1.11. Applications written against Kudu 1.11 will continue to run against the Kudu 1.12 client and vice-versa.
Please refer to the Known Issues and Limitations section of the documentation.
Kudu 1.12 includes contributions from 33 people, including 8 first-time contributors:
Andy Singer
Michele Milesi
Ning Wang
Renhai Zhao
Sheng Liu
Thomas D’Silva
Tianhua Huang
Waleed Fateem
Thank you for your help in making Kudu even better!
Apache Kudu 1.11.1 is a bug-fix release which fixes one critical licensing issue in Kudu 1.11.0.
When upgrading from earlier versions of Kudu, if support for Kudu’s NVM
(non-volatile memory) block cache is desired, install the memkind
library
of version 1.8.0 or newer as documented in Kudu
Installation for corresponding platform. This is a mandatory step for existing
users of the NVM block cache (i.e. those who set --block_cache_type=NVM
for kudu-master
and kudu-tserver
): they must install memkind
, otherwise
their Kudu processes will crash at startup.
Fixed an issue with distributing libnuma
dynamic library with kudu-binary
JAR artifact. Also, fixed the issue of statically compiling in libnuma.a
into kudu-master
and kudu-tserver
binaries when building Kudu
from source in release mode. The fix removes both numactl
and memkind
projects from Kudu’s thirdparty dependencies and makes the dependency on the
libmemkind
library optional, opening the library using dlopen()
and
resolving required symbols via dlsym()
(see KUDU-2990).
Fixed an issue with kudu cluster rebalancer
CLI tool crashing when running
against a location-aware cluster if a tablet server in one location doesn’t
contain a single tablet replica
(see KUDU-2987).
Fixed an issue with connection negotiation using SASL mechanism when server FQDN is longer than 64 characters (see KUDU-2989).
Fixed an issue in the test harness of the kudu-binary JAR artifact. With this
fix, kudu-master
and kudu-tserver
processes of the mini-cluster’s test
harness no longer rely on the test NTP server to synchronize their built-in
NTP client. Instead, the test harness relies on the local machine clock
synchronized by the system NTP daemon
(see KUDU-2994).
Since KUDU-2625 is addressed, tablet servers now reject individual write operations which violate schema constraints in a batch of write operations. In prior versions the behavior was to reject the whole batch of write operations if a violation of the schema constraints is detected even for a single row. It’s recommended to revise applications which relied on the behavior mentioned above upon upgrading to Kudu 1.11.0.
The Kudu Flume integration is deprecated and may be removed in the next minor release. The integration will be moved to the Apache Flume project going forward (see FLUME-3345).
Kudu now supports putting tablet servers into maintenance. While in this
mode, the tablet server’s replicas will not be re-replicated if it fails.
Only upon exiting maintenance will re-replication be triggered for any
remaining under-replicated tablets. The kudu tserver state enter_maintenance
and kudu tserver state exit_maintenance
tools are added to orchestrate
tablet server maintenance, and the kudu tserver list
tool is amended with
a "state" column option to display current state of each tablet server
(see KUDU-2069).
Kudu now has a built-in NTP client which maintains the internal wallclock
time used for generation of HybridTime timestamps. When enabled, system clock
synchronization for nodes running Kudu is no longer necessary. This is useful
for containerized deployments and in other cases when it’s troublesome
to maintain properly configured system NTP service at each node of a Kudu
cluster. The list of NTP servers to synchronize against is specified with the
--builtin_ntp_servers
flag. By default, Kudu masters and tablet servers use
public servers hosted by the NTP Pool project. To use the built-in NTP
client, set --time_source=builtin
and reconfigure --builtin_ntp_servers
if necessary
(see KUDU-2935).
Aggregated table statistics are now available to Kudu clients via
KuduClient.getTableStatistics()
and KuduTable.getTableStatistics()
methods in the Kudu Java client and KuduClient.GetTableStatistics()
in the Kudu C++ client. This allows for various query optimizations.
For example, Spark now uses it to perform join optimizations.
The statistics are available via the API of both C++ and Java Kudu clients.
In addition, per-table statistics are available via kudu table statistics
CLI tool. The statistics are also available via master’s Web UI at
master:8051/metrics
and master:8051/table?id=<uuid>
URIs
(see KUDU-2797 and
KUDU-2921).
The kudu
CLI tool now supports altering table columns. Use the newly
introduced sub-commands such as kudu table column_set_default
,
kudu table column_remove_default
, kudu table column_set_compression
,
kudu table column_set_encoding
, and kudu table column_set_block_size
to alter a column of the specified table.
The kudu
CLI tool now supports dropping table columns. Use the newly
introduced kudu table delete_column
sub-command to drop a column of the
specified table.
The kudu
CLI tool now supports getting and setting extra
configuration properties for a table. Use kudu table get_extra_configs
and kudu table set_extra_config
sub-commands to perform the corresponding
operations
(see KUDU-2514).
The kudu
CLI tool now supports creating and dropping range partitions
for a table. Use kudu table add_range_partition
and
kudu table drop_range_partition
sub-commands to perform the corresponding
operations
(see KUDU-2881).
The kudu fs dump uuid
CLI tool is now significantly faster and consumes
significantly less IO.
The memory consumed by CFileReaders and BloomFileReaders is factored out and accounted separately by the tablet server memory tracking. The stats are available via Web UI as "CFileReaders" and "BloomFileReaders" entries.
KuduScanBatch::const_iterator
in Kudu C++ client now supports
operator→()
(see KUDU-1561).
Master server Web UI now supports sorting the list of tables by the columns of "Table Name", "Create Time", and "Last Alter Time".
Tablet servers now expand a tablet’s data directory group with available healthy directories when all directories of the group are full (see KUDU-2907).
For scan operations run with CLOSEST_REPLICA
selection mode, the Kudu Java
client now picks a random available replica in case no replica is located at
the same node with the client that initiated the scan operation. This helps
to spread the load generated by multiple scan requests to the same tablet
among all available replicas. In prior releases, all such scan requests might
end up fetching data from the same tablet replica
(see KUDU-2348).
The serialization of in-memory rows to Kudu’s wire format has been optimized to be more CPU efficient (see KUDU-2847).
Tablet servers and masters can now aggregate metrics by the same attribute.
For example, it’s now possible to fetch aggregated metrics from a tablet
server by retrieving data from URLs of form
http://<host>:<port>/metrics?merge_rules=tablet|table|table_name
Introduced Docker image for Python Kudu client (see KUDU-2849).
Tablet servers now consider available disk space when choosing a set of data directories for a tablet’s data directory group, and when deciding in which data directory a new block should be written (see KUDU-2901).
Added a quick-start example of using Apache Spark to load, query, and modify a real data set stored in Kudu.
Added a quick-start example of using Apache Nifi to ingest data into Kudu.
Tablet servers now reject individual write operations which violate schema constraints in a batch of write operations received from a client. The previous behavior was to reject the whole batch of write operations if a violation of the schema constraints is detected even for a single row (see KUDU-2625).
Tablet replicas can now be optionally placed in accordance with a
dimension-based placement policy. To specify a dimension label for a table,
use the KuduTableCreator::dimension_label()
and
CreateTableOptions.setDimensionLabel()
methods of the C++ and Java Kudu
clients. To add a partition with a dimension label, use the
KuduTableAlterer::AddRangePartitionWithDimension()
and
AlterTableOptions.addRangePartition()
methods of the C++ and Java Kudu
clients
(see KUDU-2823).
Kudu RPC now enables TCP keepalive for all outbound connections for faster detection of no-longer-reachable nodes (see KUDU-2192).
The kudu table scan
and kudu table copy
CLI tools now fail gracefully
rather than crashing upon hitting an error
(see KUDU-2851).
Optimized decoding of deltas' timestamps (see KUDU-2867).
Optimized the initialization of DeltaMemStore for the case when no matching deltas are present (see KUDU-2381).
Improved the rehydration of scan tokens. Now a scan token created before renaming a column can be used even after the column has been renamed.
The memory reserved by tcmalloc is now released to OS periodically to avoid potential OOM issues in the case of read-only workloads (see KUDU-2836).
Optimized evaluation of predicates on columns of primitive types and
NULL
/NOT NULL
predicates to leverage SIMD instructions
(see KUDU-2846).
Fixed an issue of fault-tolerant scan operation failing for a projection with key columns specified in other than the table schema’s order (see KUDU-2980).
Fixed an issue that would cause frequent leader elections in case when persisting Raft transactions to the WAL took longer than the leader election timeout. The issue was contributing to election storms (see KUDU-2947).
Fixed a tablet server crash in cases where blocks were not removed due to IO error. This issue may have surfaced after recovering from a disk failure (see KUDU-2635).
Fixed a crash in master and tablet server by validating the size of default
values when de-serializing ColumnSchemaPB
(see KUDU-2622).
Fixed RPC negotiation failure in the case when TLS v1.3 is supported at both the client and the server side. This is a temporary workaround before the connection negotiation code is properly updated to support 1.5-RTT handshake used in TLS v1.3. The issue affected Linux distributions shipped or updated with OpenSSL version 1.0.2 and newer (see KUDU-2871).
Fixed a race between GetTabletLocations()
and tablet report processing.
The race could crash the Kudu master
(see KUDU-2842).
Fixed a bug in AlterSchemaTransactionState::ToString()
that led to a crash
of tablet server when removing a tablet replica with a pending AlterSchema
transaction.
Kudu 1.11.0 is wire-compatible with previous versions of Kudu:
Kudu 1.11 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Rolling upgrade between Kudu 1.10 and Kudu 1.11 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
Kudu 1.0 clients may connect to servers running Kudu 1.11 with the exception of the below-mentioned restrictions regarding secure clusters.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.11 and versions earlier than 1.3:
If a Kudu 1.11 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.11 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
The Kudu 1.11 Java client library is API- and ABI-compatible with Kudu 1.10. Applications written against Kudu 1.10 will compile and run against the Kudu 1.11 client library and vice-versa.
The Kudu 1.11 C++ client is API- and ABI-forward-compatible with Kudu 1.10. Applications written and compiled against the Kudu 1.10 client library will run without modification against the Kudu 1.11 client library. Applications written and compiled against the Kudu 1.11 client library will run without modification against the Kudu 1.10 client library.
The Kudu 1.11 Python client is API-compatible with Kudu 1.10. Applications written against Kudu 1.10 will continue to run against the Kudu 1.11 client and vice-versa.
Please refer to the Known Issues and Limitations section of the documentation.
Kudu 1.11 includes contributions from 24 people, including 8 first-time contributors:
Hannah Nguyen
lingbin
Ritwik Yadav
Scott Reynolds
Volodymyr Verovkin
Xiaokai Wang
Xin He
Yao Wang
Thank you for your help in making Kudu even better!
Apache Kudu 1.10.1 is a bug-fix release which fixes one critical licensing issue in Kudu 1.10.0.
When upgrading from earlier versions of Kudu, if support for Kudu’s NVM
(non-volatile memory) block cache is desired, install the memkind
library
of version 1.6.0 or newer as documented in Kudu
Installation for corresponding platform. This is a mandatory step for existing
users of the NVM block cache (i.e. those who set --block_cache_type=NVM
for kudu-master
and kudu-tserver
): they must install memkind
, otherwise
their Kudu processes will crash at startup.
Fixed an issue with distributing libnuma
dynamic library with kudu-binary
JAR artifact. Also, fixed the issue of statically compiling in libnuma.a
into kudu-master
and kudu-tserver
binaries when building Kudu
from source in release mode. The fix removes both numactl
and memkind
projects from Kudu’s thirdparty dependencies and makes the dependency on the
libmemkind
library optional, opening the library using dlopen()
and
resolving required symbols via dlsym()
(see KUDU-2990).
The default tablet history retention time has been raised from 15 minutes to 7 days to better support touchless incremental backups (see KUDU-2677).
Kudu now supports both full and incremental table backups via a job implemented using Apache Spark. Additionally it supports restoring tables from full and incremental backups via a restore job implemented using Apache Spark. See the backup documentation for more details.
Kudu can now synchronize its internal catalog with the Apache Hive Metastore, automatically updating Hive Metastore table entries upon table creation, deletion, and alterations in Kudu. See the HMS synchronization documentation for more details.
Kudu now supports native fine-grained authorization via integration with Apache Sentry. Kudu may now enforce access control policies defined for Kudu tables and columns, as well as policies defined on Hive servers and databases that may store Kudu tables. See the authorization documentation for more details.
Kudu’s web UI now supports SPNEGO, a protocol for securing HTTP requests with
Kerberos by passing negotiation through HTTP headers. To enable, set the
--webserver_require_spnego
command line flag.
Column comments can now be stored in Kudu tables, and can be updated using the AlterTable API (see KUDU-1711).
The Java scan token builder can now create multiple tokens per tablet.
To use this functionality, call setSplitSizeBytes()
to specify how many bytes
of data each token should scan. The same API is also available in Kudu’s
Spark integration, where it can be used to spawn multiple Spark tasks per
scanned tablet
(see KUDU-2670).
Experimental Kudu Docker images are now published on Docker Hub.
Kudu now has an experimental Kubernetes StatefulSet manifest and Helm chart, which can be used to define and provision Kudu clusters using Kubernetes (see KUDU-2398).
The Kudu CLI now has rudimentary YAML-based configuration file support, which can be used to provide cluster connection information via cluster name instead of keying in comma-separated lists of master addresses. See the cluster name documentation for more details.
kudu perf table_scan
scans a table and displays a table’s row count as well
as the time it took to run the scan.
kudu table copy
copies data from one table to another, within the same
cluster or across clusters. Note, this implementation leverages a single client,
therefore it may not be suitable for large tables.
Tablet history retention time can now be configured on a table-by-table basis. (see KUDU-2514).
The performance of mutations (i.e. UPDATE, DELETE, and re-INSERT) to not-yet-flushed Kudu data has been significantly optimized (see KUDU-2826 and f9f9526d3).
Predicate performance for primitive columns has been optimized (see KUDU-2846).
IS NULL and IS NOT NULL predicate performance has been optimized (see KUDU-2846).
Optimized the performance of fetching tablet locations from the master for tables with large numbers of partitions. This can improve the performance of short-running Spark or Impala queries as well as user applications which make use of short-lived client instances (see KUDU-2711).
The tableExists()
(Java) and TableExists()
(C++) APIs are now more performant
(see KUDU-2802).
Fault tolerant scans are now much more performant and consume far less memory (see KUDU-2466).
kudu cluster ksck
now sends more requests in parallel, which should result
in a speed-up when running against clusters with many tables or when there’s
high latency between the node running the CLI and the cluster nodes.
Kudu’s block manager now deletes spent block containers when needed instead of just at server startup. This should reduce server startup times somewhat (see KUDU-2636).
DNS resolutions are now cached by Kudu masters, tablet servers, and Kudu C++ clients. The TTL for a resolved DNS entry in the cache is 15 seconds by default (see KUDU-2791).
Tables created in Kudu 1.10.0 or later will show their creation time as well as their last alteration time in the web UI (see KUDU-2750).
The Kudu CLI and C++ client now support overriding the local username using the ‘KUDU_USER_NAME’ environment variable. This allows operating against a Kudu cluster using an identity which differs from the local Unix user on the client. Note that this has no effect on secure clusters, where client identity is determined by Kerberos authentication (see KUDU-2717).
Kudu C++ client now performs stricter verification on the input data of INSERT and UPSERT operations w.r.t. table schema constraints. This helps spotting schema violations before sending the data to a tablet server.
The KuduScanner
in the Java client is now iterable. Additionally the
KuduScannerIterator
will automatically make scanner keep alive calls to
ensure scanners do not time out while iterating.
A KuduPartitioner
API was added to the Java client. The KuduPartitioner
API allows a client to determine which partition a row falls into without
actually writing that row. For example, the KuduPartitioner
is used in the
Spark integration to optionally repartition and pre-sort the data before
writing to Kudu
(see KUDU-2674 and
KUDU-2672).
The PartialRow
and RowResult
Java API have new methods that accept and return
Java Objects. These methods are useful when you don’t care about autoboxing
and your existing type handling logic is based on Java types. See the javadoc
for more details.
The Kudu Java client now logs RPC trace summaries instead of full RPC traces when
the log level is INFO
or higher. This reduces log noise and makes RPC issues
more visible in a more compact format
(see KUDU-2830).
Kudu servers now display the time at which they were started in their web UIs.
Kudu tablet servers now display a table’s total column count in the web UI.
The /metrics
web UI endpoint now supports filtering on entity types,
entity IDs, entity attributes, and metric names. This can be used to more
efficiently collect important metrics when there is a large number of tablets
on a tablet server.
The Kudu rebalancer now accepts the --ignored_tservers
command line
argument, which can be used to ignore the health status of specific tablet
servers (i.e. if they are down) when deciding whether or not it’s safe to
rebalance the cluster.
kudu master list
now displays the Raft consensus role of each master in the
cluster (i.e. LEADER or FOLLOWER)
(see KUDU-2825).
kudu table scan
no longer interleaves its output, and now projects all
columns without having to manually list the column names.
kudu perf loadgen
now supports creating empty tables. The semantics of the
special value of 0 for --num_rows_per_thread
flag has changed. A value of 0
now indicates that no rows should be generated, and a value of -1 indicates
there should be no limit to the number of rows generated.
Running make install
after building Kudu from source will now install the
Kudu binaries into appropriate locations.
(see KUDU-1344).
Fixed an issue where the Java client would fail scans that took a very long time to return a single block of rows, such as highly selective scans over a large amount of data (see KUDU-1868).
Fixed the handling of SERVICE_UNAVAILABLE errors that caused the Java client to do unnecessary master lookups.
Kudu scan tokens now work correctly when the target table is renamed between when the scan token is created and when it is rehydrated into a scanner.
Kudu’s “NTP synchronization wait” behavior at startup now works properly when Kudu is run in a containerized environment.
Fixed a crash when a flush or compaction overlapped with another compaction (see KUDU-2807).
Fixed a rare race at startup where the leader master would fruitlessly try to tablet copy to a healthy follower master, causing the cluster to operate as if it had two masters until master leadership changed (see KUDU-2748).
Under rare circumstances, it was possible for Kudu to crash in libkrb5 when negotiating multiple TLS connections concurrently. This crash has been fixed (see KUDU-2706).
Kudu no longer crashes at startup on machines with disabled CPUs (see KUDU-2721).
Kudu 1.10.0 is wire-compatible with previous versions of Kudu:
Kudu 1.10 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Rolling upgrade between Kudu 1.9 and Kudu 1.10 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
Kudu 1.0 clients may connect to servers running Kudu 1.10 with the exception of the below-mentioned restrictions regarding secure clusters.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.10 and versions earlier than 1.3:
If a Kudu 1.10 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.10 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
Support for building and running with Java 7 has been dropped in this release. It had been deprecated since Kudu 1.5.0. (see KUDU-2099).
The Kudu 1.10 Java client library is API- and ABI-compatible with Kudu 1.9. Applications written against Kudu 1.9 will compile and run against the Kudu 1.10 client library and vice-versa.
The Kudu 1.10 C++ client is API- and ABI-forward-compatible with Kudu 1.9. Applications written and compiled against the Kudu 1.9 client library will run without modification against the Kudu 1.10 client library. Applications written and compiled against the Kudu 1.10 client library will run without modification against the Kudu 1.9 client library.
The Kudu 1.10 Python client is API-compatible with Kudu 1.9. Applications written against Kudu 1.9 will continue to run against the Kudu 1.10 client and vice-versa.
Please refer to the Known Issues and Limitations section of the documentation.
Kudu 1.10 includes contributions from 27 people, including 6 first-time contributors:
Csaba Fulop
Florentino Sainz
Guangchao Deng
Jia Hongchao
Ye Yuqiang
Yifan Zhang
Thank you for your help in making Kudu even better!
Flume 1.8+ requires Java 8 at runtime even though the Kudu Flume integration is Java 7 compatible. Flume 1.9 is the default dependency version as of Kudu 1.9.0.
Hadoop 3.0+ requires Java 8 at runtime even though the Kudu Hadoop integration is Java 7 compatible. Hadoop 3.2 is the default dependency version as of Kudu 1.9.0.
Support for Java 7 has been deprecated since Kudu 1.5.0 and may be removed in the next major release.
Kudu now supports location awareness. When configured, Kudu will make a best
effort to avoid placing a majority of replicas for a given tablet at the same
location. The kudu cluster rebalance
tool has been updated to act in
accordance with the placement policy of a location-aware Kudu. The
administrative
documentation has been updated to detail the usage of this feature.
Docker scripts have been introduced to build and run Kudu on various operating
systems. See the /docker
subdirectory of the source repository for more
details. An official repository has
been created for Apache Kudu Docker artifacts.
Developers integrating with Kudu can now write Java tests that start a Kudu mini cluster without having to first locally build and install Kudu. This is made possible by the Kudu team providing platform-specific binaries available to Gradle or Maven for download and install at test time. More information on this feature can be found here. This binary test artifact is currently considered to be experimental.
When creating a table, the master now enforces a restriction on the total
number of replicas rather than the total number of partitions. If manually
overriding --max_create_tablets_per_ts
, the maximum size of a new table
has effectively been cut by a factor of its replication factor. Note that
partitions can still be added after table creation.
The compaction policy has been updated to favor reducing the number of rowsets. This can lead to faster scans and lower bootup times, particularly in the face of a “trickling inserts” workload, where rows are inserted slowly in primary key order (see KUDU-1400).
A tablet-level metric average_diskrowset_height
has been added to indicate
how much a replica needs to be compacted, as indicated by the average number
of rowsets per unit of keyspace.
Scans which read multiple columns of tables undergoing a heavy UPDATE
workload are now more CPU efficient. In some cases, scan performance of such
tables may be several times faster upon upgrading to this release.
Kudu-Spark users can now provide the short “kudu” format alias to Spark. This
enables using .format(“kudu”)
in places where you would have needed to
provide the fully qualified name like .format(“org.apache.kudu.spark.kudu")
or imported org.apache.kudu.spark.kudu._
and used the implicit .kudu
functions. The
Spark
integration documentation has been updated to reflect this improvement.
The KuduSink
class has been added to the Spark integration as a
StreamSinkProvider
, allowing structured streaming writes into Kudu (see
KUDU-2640).
The amount of server-side logging has been greatly reduced for Kudu’s consensus implementation and background processes. This logging was determined to be not useful and unnecessarily verbose.
The web UI now more obviously depicts which columns are a part of the primary key (see KUDU-2477).
The kudu table describe
tool has been added to support describing table
attributes, including schema, partitioning, replication factor, column
encodings, compressions, and default values.
The kudu table scan
tool has been added to scan rows from a table,
supporting comparison, in-list, and is-null predicates.
The kudu locate_row
tool has been added to allow users to determine what
tablet a given primary key belongs to, and whether a row exists for that
primary key.
The kudu diagnose dump_mem_trackers
tool is added to allow users to output
the contents of the /mem-trackers
web UI page in a CSV format.
To avoid glitches and undefined behavior, the Kudu Python client now detects and reports on conflicting/incorrect initialization of the OpenSSL library.
Fixed a crash caused by a race between altering tablet schemas and deleting tablet replicas (see KUDU-1678).
Fixed an issue that would prevent the kudu fs update_dirs
tool from
removing directories in the presence of tablet tombstones (see
KUDU-2680).
The --cmeta_force_fsync
flag may be used to fsync Kudu’s consensus
metadata more aggressively. Setting this to true
may decrease Kudu’s
performance, but improve its durability in the face of power failures and
forced shutdowns (see
KUDU-2195).
Fixed an issue that would cause an excessive amount of RPC traffic from Kudu masters if the tablet servers were configured with duplicated master addresses (see KUDU-2684).
Fixed an issue that would cause the kudu cluster rebalance
tool to run
indefinitely in the case of tables with a replication factor of 2 (see
KUDU-2688).
Fixed an issue that could lead to a failure to bootstrap tablet replicas that were a part of workloads with many alter table operations (see KUDU-2690).
Fixed an issue with the Java scanner’s keepAlive
that could lead to a
permanent hang in the scanner (see
KUDU-2710).
Fixed an issue that would cause undefined behavior upon connecting to a secure cluster concurrently from multiple C++ clients (see KUDU-2706).
Kudu 1.9.0 is wire-compatible with previous versions of Kudu:
Kudu 1.9 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Rolling upgrade between Kudu 1.8 and Kudu 1.9 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
Kudu 1.0 clients may connect to servers running Kudu 1.9 with the exception of the below-mentioned restrictions regarding secure clusters.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.9 and versions earlier than 1.3:
If a Kudu 1.9 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.9 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
The Kudu 1.9 Java client library is API- and ABI-compatible with Kudu 1.8. Applications written against Kudu 1.8 will compile and run against the Kudu 1.9 client library and vice-versa.
The Kudu 1.9 C++ client is API- and ABI-forward-compatible with Kudu 1.8. Applications written and compiled against the Kudu 1.8 client library will run without modification against the Kudu 1.9 client library. Applications written and compiled against the Kudu 1.9 client library will run without modification against the Kudu 1.8 client library.
The Kudu 1.9 Python client is API-compatible with Kudu 1.8. Applications written against Kudu 1.8 will continue to run against the Kudu 1.9 client and vice-versa.
Please refer to the Known Issues and Limitations section of the documentation.
Kudu 1.9 includes contributions from 24 people, including 5 first-time contributors:
Bankim Bhavsar
Mike Parker
Mitch Barnett
Tim Armstrong
Yingchun Lai
Thank you for your help in making Kudu even better!
Upgrading directly from Kudu 1.7.0 is supported and no special upgrade steps are required. A rolling upgrade may work, however it has not been tested. When upgrading Kudu, it is recommended to first shut down all Kudu processes across the cluster, then upgrade the software on all servers, then restart the Kudu processes on all servers in the cluster.
Kudu Flume Sink released with Kudu 1.8.0 is compiled against Apache Flume 1.8 and might not function with earlier versions of Flume. Note that Flume 1.8 requires Java 1.8 or higher.
Hadoop 3.0+ requires Java 8 at runtime even though the Kudu Hadoop integration is Java 7 compatible. Hadoop 3.1 is the default dependency version as of Kudu 1.8.0, used by certain features in the Java client.
The -table_num_buckets
configuration option of the kudu perf loadgen
tool is now
removed in favor of -table_num_hash_partitions
and -table_num_range_partitions
(see KUDU-1861).
Support for Java 7 has been deprecated since Kudu 1.5.0 and may be removed in the next major release.
The producer.skipMissingColumn
, producer.skipBadColumnValue
, and
producer.warnUnmatchedRows
Kudu Flume sink configuration parameters have been
deprecated in favor of producer.missingColumnPolicy
, producer.badColumnValuePolicy
,
and producer.unmatchedRowPolicy
respectively (see
KUDU-1882).
Examples showcasing functionality in C++, Java, and Python, previously
hosted in a separate repository have been added. They can be found in the
examples/
top-level subdirectory.
Added kudu diagnose parse_stacks
, a tool to parse sampled stack traces out of a
diagnostics log (see KUDU-2353).
Added support for IS NULL
and IS NOT NULL
predicates to the Kudu Python client (see
KUDU-2399).
Introduced manual data rebalancer into the kudu
CLI tool. The rebalancer can be used to redistribute table replicas among tablet
servers. The rebalancer can be run via kudu cluster rebalance
sub-command. Using the
new tool, it’s possible to rebalance Kudu clusters of version 1.4.0 and newer.
Added kudu tserver get_flags
and kudu master get_flags
, two tools that allow
superusers to retrieve all the values of command line flags from remote Kudu processes.
The get_flags
tools support filtering the returned flags by tag, and by default will
return only flags that were explicitly set.
Added kudu tablet unsafe_replace_tablet
, a tool to replace a tablet with a new one.
This tool is meant to be used to recover a table when one of its tablets has permanently
lost all replicas. The data in the tablet that is replaced is lost, so this tool should
only be used as a last resort (see
KUDU-2290).
There is a new metric for each tablet replica tracking the number of election failures since the last successful election attempt and the time since the last heartbeat from the leader (see KUDU-2287).
Kudu now supports building and running on Ubuntu 18.04 (“Bionic Beaver”) (see KUDU-2427).
Kudu now supports building and running against OpenSSL 1.1 (see KUDU-1889).
Added Kerberos support to the Kudu Flume sink (see KUDU-2012).
The Kudu Spark connector now supports Spark Streaming DataFrames (see KUDU-2539).
Added -tables
filtering argument to kudu table list
(see
KUDU-2529).
Clients now support setting a limit on the number of returned rows in scans (see KUDU-16).
Added Pandas support to the Python client (see KUDU-1276).
Enabled configuration of mutation buffer in the Python client (see KUDU-2441).
Added a keepAlive
API call to the KuduScanner
and AsyncKuduScanner
in the Java
client. This API can be used to keep the scanners alive on the server when processing
of messages will take longer than the scanner TTL (see
KUDU-2095).
The Kudu Spark integration now uses the keepAlive API when reading data. By default it will call keepAlive on a scanner with a period of 15 seconds. This will ensure that Spark jobs with large batch sizes or slow processing times do not fail with scanner not found errors (see KUDU-2563).
Number of reactor threads in the C++ client is now configurable (see KUDU-2368).
Added an optimization to reduce CPU consumption when performing hot metadata lookups in the C++ client (see KUDU-1977).
Added an optimization to avoid bottlenecks on getpwuid_r()
in libnss during a Raft
leader election storm (see
KUDU-2395).
Improved rowset tree pruning making scans with open-ended intervals on primary key (see KUDU-2566).
The kudu perf loadgen
tool now supports generating range-partitioned tables. The
-table_num_buckets
configuration is now removed in favor of
-table_num_hash_partitions
and -table_num_range_partitions
(see
KUDU-1861).
CFile checksum failures will now cause the affected tablet replicas to be failed and re-replicated elsewhere (see KUDU-2469).
Servers are now able to start up with data directories missing on disk (see KUDU-2359).
The kudu perf loadgen
tool now creates tables with a period-separated database name,
for example default.loadgen_auto_abc123
. This new behavior does not take effect if the
--table
flag is provided. The database of the table can be changed using a new
--auto_database
flag. This change is made in anticipation of an eventual Kudu/HMS
integration (see KUDU-2191).
Introduced FAILED_UNRECOVERABLE
replica health status. This is to mark replicas which
are not able to catch up with the leader due to GC-collected segments of WAL and other
unrecoverable cases like disk failure. With that, the replica management scheme becomes
hybrid: the system evicts replicas with FAILED_UNRECOVERABLE
health status before
adding a replacement if it anticipates that it can commit the transaction, while in
other cases it first adds a non-voter replica and removes the failed one only after
promoting a newly added replica to voter role.
Two additional configuration parameters, socketReadTimeoutMs
and scanRequestTimeout
have been added to the Spark connector to allow better tuning to avoid scan timeouts
under high load.
The kudu table
tool now supports two new options to rename tables and columns,
rename_table
and rename_column
respectively.
Kudu will now wait for the clock to become synchronized at startup, controlled by a new
flag -ntp_initial_sync_wait_secs
(see
KUDU-2242).
Tablet deletions are now throttled, which will help Kudu clusters remain stable even
when many tablets are deleted at once. The number of tablets that a tablet server will
delete at once is controlled by the new flag -num_tablets_to_delete_simultaneously
(see KUDU-2289).
The kudu cluster ksck
tool has been significantly enhanced. It now checks master
health and consensus status, displays any unsafe or hidden flags set in the cluster, and
produces a summary of the Kudu versions running on the master and tablet servers. In
addition, it now supports JSON output, both in pretty-printed and compact form. The
output format is controlled by the -ksck_format
flag.
When a tablet server was wiped and recreated with the same RPC address, ksck
listed it
twice, both as healthy, even though only one of them was there. This bug is now fixed by
verifying the UUID of the server (see
KUDU-2364).
Fixed an issue preventing Kudu from starting when using Vormetric’s encrypted filesystem (secfs2) on ext4 (see KUDU-2406).
Fixed an issue where Kudu’s block cache memory tracking (as seen on the /mem-trackers
web UI page) wasn’t accounting for all of the overhead of the cache itself (see
KUDU-972).
Fixed an issue where the C++ client would fail to reopen an expired scanner; instead, the client would retry in a tight loop and eventually timeout (see KUDU-2414).
When a tablet is deleted, its write-ahead log recovery directory is also deleted, if it exists (see KUDU-1038).
Fixed a tablet server crash when a tablet is scanned with two predicates on its primary key and the predicates do not overlap (see KUDU-2447).
Fixed an issue where the Kudu MapReduce connector’s KuduTableInputFormat
may exhaust
its scan too early (see
KUDU-2525).
Fixed an issue with failed tablet copies that would cause subsequent tablet copies to crash the tablet server (see KUDU-2293).
Fixed a bug in which incorrect results would be returned in scans following a server restart (see KUDU-2463).
Fixed a bug causing a tablet server crash when a write batch request from a client failed coarse-grained authorization (see KUDU-2540).
Fixed use-after-free in case of WAL replay error (see KUDU-2509).
Fixed authentication token reacquisition in the C++ client (see KUDU-2580).
Fixed a bug where leader logged excessively when the followers fell behind (see KUDU-2322).
Fixed reporting of leader health during lifecycle transitions (see KUDU-2335).
Fixed moving single-replica tablets (see KUDU-2443).
Fixed an error that would cause the kudu CLI tool to unexpectedly exit when the connection to the master or tserver was abruptly closed.
Fixed a rare issue where system failure could leave unexpected null bytes at the end of metadata files, causing Kudu to be unable to restart (see KUDU-2260).
Fixed an issue where kudu cluster ksck
running a snapshot checksum scan would use a
single snapshot timestamp for all tablets. This caused the checksum process to fail if
the checksum process took a long time and the number of tablets was sufficiently large.
The tool should now be able to checksum tables even if the process takes many hours.
(see KUDU-2179).
Kudu 1.8.0 is wire-compatible with previous versions of Kudu:
Kudu 1.8 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Kudu 1.0 clients may connect to servers running Kudu 1.8 with the exception of the below-mentioned restrictions regarding secure clusters.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.8 and versions earlier than 1.3:
If a Kudu 1.8 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.8 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
The Kudu 1.8 Java client library is API- and ABI-compatible with Kudu 1.7. Applications written against Kudu 1.7 will compile and run against the Kudu 1.8 client library and vice-versa.
The Kudu 1.8 C++ client is API- and ABI-forward-compatible with Kudu 1.7. Applications written and compiled against the Kudu 1.7 client library will run without modification against the Kudu 1.8 client library. Applications written and compiled against the Kudu 1.8 client library will run without modification against the Kudu 1.7 client library.
The Kudu 1.8 Python client is API-compatible with Kudu 1.7. Applications written against Kudu 1.7 will continue to run against the Kudu 1.8 client and vice-versa.
Please refer to the Known Issues and Limitations section of the documentation.
Kudu 1.8 includes contributions from 40 people, including 15 first-time contributors:
Anupama Gupta
Attila Piros
Brian McDevitt
Fengling Wang
Ferenc Szabó
Greg Solovyev
Kiyoshi Mizumaru
Shriya Gupta
Thomas Tauber-Marshall
Tigerquoll
Yao Xu
ZhangYao
helifu
jinxing64
qqchang2nd
Thank you for helping to make Kudu even better!
Apache Kudu 1.7.1 is a bug-fix release which fixes critical issues in Kudu 1.7.0.
Fixed and issue where a leader replica could report a follower’s health status as FAILED instead of FAILED_UNRECOVERABLE. In configurations where the tablet replication factor equals to the total number of tablet servers in the cluster, that lead to situations where the tablet could not be automatically recovered until a new leader was elected or corresponding tablet servers were restarted. (see KUDU-2367).
Fixed an issue where Kudu would fail to start if RLIMIT_NPROC was set to -1. (see KUDU-2377).
Fixed an issue where kudu-spark
was unable to connect to secure clusters.
(see KUDU-2379).
Fixed an issue where the kudu-python
client would not compile in environments
where __int128
is not supported. This was most commonly el6 environments.
(see KUDU-2412).
Fixed an issue where unaligned loads of __int128
integers could result
in a crash.
(see KUDU-2378).
Fixed a bug in PartialRow.setMin
that could lead to incorrect partition
pruning when a decimal
column is part of the tables range partition but
not a part of the query predicate.
(see KUDU-2416).
Fixed an equality check on decimal
column predicates that could result
in pruning that is too conservative.
Upgrading directly from Kudu 1.6.0 is supported and no special upgrade steps are required. A rolling upgrade of the server side will not work because the default replica management scheme changed, and running masters and tablet servers with different replica management schemes is not supported, see Incompatible Changes in Kudu 1.7.0 for details. However, mixing client and server sides of different versions is not a problem. You can still update your clients before your servers or vice versa. When upgrading to Kudu 1.7, it is required to first shut down all Kudu processes across the cluster, then upgrade the software on all servers, then restart the Kudu processes on all servers in the cluster.
The tcmalloc_contention_time
metric, which previously tracked the amount
of time spent in memory allocator lock contention, has been removed.
Support for Java 7 has been deprecated since Kudu 1.5.0 and may be removed in the next major release.
Kudu now supports the decimal column type. The decimal type is a numeric data type with fixed scale and precision suitable for financial and other arithmetic calculations where the imprecise representation and rounding behavior of float and double make those types impractical. The decimal type is also useful for integers larger than int64 and cases with fractional values in a primary key. See Decimal Type for more details.
The strategy Kudu uses for automatically healing tablets which have lost a replica due to server or disk failures has been improved. The new re-replication strategy, or replica management scheme, first adds a replacement tablet replica before evicting the failed one. With the previous replica management scheme, the system first evicts the failed replica and then adds a replacement. The new replica management scheme allows for much faster recovery of tablets in scenarios where one tablet server goes down and then returns back shortly after 5 minutes or so. The new scheme also provides substantially better overall stability on clusters with frequent server failures. (see KUDU-1097).
The kudu fs update_dirs
tool now supports removing directories. Unless the
--force
flag is specified, Kudu will not allow the removal of a directory
across which tablets are configured to spread data. If specified, all tablet
replicas configured to use that directory will fail upon starting up and be
replicated elsewhere, provided a majority exists elsewhere.
Users can use the new --fs_metadata_dir
to specify the directory in which
to place tablet-specific metadata. It is recommended, although not
necessary, that this be placed on a high-performance drive with high
bandwidth and low latency, e.g. a solid-state drive. If not specified,
metadata will be placed in the directory specified by --fs_wal_dir
, or in
the directory specified by the first entry of --fs_data_dirs
if metadata
already exists there from a pre-Kudu 1.7 deployment. Kudu will not
automatically move existing metadata based on this configuration.
Kudu 1.7 introduces a new scan read mode READ_YOUR_WRITES. Users can specify READ_YOUR_WRITES when creating a new scanner in C++, Java and Python clients. If this mode is used, the client will perform a read such that it follows all previously known writes and reads from this client. Reads in this mode ensure read-your-writes and read-your-reads session guarantees, while minimizing latency caused by waiting for outstanding write transactions to complete. Note that this is still an experimental feature which may be stabilized in future releases.
The tablet server web UI scans dashboard (/scans) has been improved with several new features, including: showing the most recently completed scans, a pseudo-SQL scan descriptor that concisely shows the selected columns and applied predicates, and more complete and better documented scan statistics.
Kudu daemons now expose a web page /stacks
which dumps the current stack
trace of every thread running in the server. This information can be helpful
when diagnosing performance issues.
By default, each tablet replica will now stripe data blocks across 3 data directories instead of all data directories. This decreases the likelihood that any given tablet will be affected in the event of a single disk failure. No substantial performance impact is expected due to this feature based on performance testing. This change only affects new replicas created after upgrading to Kudu 1.7.
Kudu servers previously offered the ability to enable a separate metrics log which stores periodic snapshots of all metrics available on a server. This functionality is now available as part of a more general “diagnostics log” which is enabled by default. The diagnostics log includes periodic dumps of server metrics as well as collections of thread stack traces. The default configuration ensures that no more than 640MB of diagnostics logs are retained, and typically the space consumption is significantly less due to compression. The format and contents of this log file are documented in the Administration guide.
The handling of errors in the synchronous Java client has been improved so that, when an exception is thrown, the stack trace indicates the correct location where the client function was invoked rather than a call stack of an internal worker thread. The original call stack from the worker thread is available as a “suppressed exception”.
The logging of errors in the Java client has been improved to exclude exception stack traces for expected scenarios such as failure to connect to a server in a cluster. Instead, only a single line informational message will be logged in such cases to aid in debugging.
The Java client now uses a predefined prioritized list of TLS ciphers when establishing an encrypted connection to Kudu servers. This cipher list matches the list of ciphers preferred for server-to-server communication and ensures that the most efficient and secure ciphers are preferred. When the Kudu client is running on Java 8 or newer, this provides a substantial speed-up to read and write performance.
Reporting for the kudu cluster ksck
tool has been updated so tablets and
tables with on-going tablet copies are shown as "recovering". Additional
reporting changes have been made to make various common scenarios,
particularly tablet copies, less alarming.
The performance of inserting rows containing many string or binary columns has been improved, especially in the case of highly concurrent write workloads.
By default, Spark tasks that scan Kudu will now be able to scan non-leader replicas. This allows Spark to more easily schedule kudu-spark tasks local to the data. Users can disable this behavior by passing 'leader_only' to the 'kudu.scanLocality' option."
The number of OS threads used in the steady state and during bursts of
activity (such as in Raft leader elections triggered by a node failure) has
been drastically reduced and should no longer exceed the value of ulimit -u
.
As such, it should no longer be necessary to increase the value of ulimit -u
(or of /proc/sys/kernel/threads-max) in order to run a Kudu tablet server in
most cases.
(see KUDU-1913).
An issue where sparse column predicates could cause excessive data-block reads has been fixed. Previously in certain scans with sparsely matching predicates on multiple columns, Kudu would read and decode the same data blocks many times. The improvement typically results in a 5-10x performance increase for the affected scans. (see KUDU-2231).
The efficiency and on-disk size of large updated values has been improved. This will improve update-heavy workloads which overwrite large (1KiB+) values. (see KUDU-2253).
Fixed a scenario where the on-disk data of a tablet server was completely erased and and a new tablet server was started on the same host. This issue could prevent tablet replicas previously hosted on the server from being evicted and re-replicated. Tablets now immediately evict replicas that respond with a different server UUID than expected. (see KUDU-1613).
Fixed a rare race condition when connecting to masters during their startup which might cause a client to get a response without a CA certificate and/or authentication token. This would cause the client to fail to authenticate with other servers in the cluster. The leader master now always sends a CA certificate and an authentication token (when applicable) to a Kudu client with a successful ConnectToMaster response. (see KUDU-1927).
The Kudu Java client now will retry a connection if no master is discovered as a leader, and the user has a valid authentication token. This avoids failure in recoverable cases when masters are in the process of the very first leader election after starting up. (see KUDU-2262).
The Java client will now automatically attempt to re-acquire Kerberos
credentials from the ticket cache when the prior credentials are about to
expire. This allows client instances to persist longer than the expiration
time of a single Kerberos ticket so long as some other process renews the
credentials in the ticket cache. Documentation on interacting with Kerberos
authentication has been added to the Javadoc for the AsyncKuduClient
class.
(see KUDU-2264).
Follower masters are now able to verify authentication tokens even if they have never been a leader. Prior to this fix, if a follower master had never been a leader, clients would be unable to authenticate to that master, resulting in spurious error messages being logged. (see KUDU-2265).
Fixed a tablet server crash when a tablet replica is deleted during a scan. (see KUDU-2295).
The evaluation order of predicates in scans with multiple predicates has been made deterministic. Due to a bug, this was not necessarily the case previously. Predicates are applied in most to least selective order, with ties broken by column index. The evaluation order may change in the future, particularly when better column statistics are made available internally. (see KUDU-2312).
Previously, the kudu tablet change_config move_replica
tool required all
tablet servers in the cluster to be available when performing a move. This
restriction has been relaxed: only the tablet server that will receive a replica
of the tablet being moved and the hosts of the tablet’s existing replicas need to be
available for the move to occur.
(see KUDU-2331).
Fixed a bug in the Java client which prevented the client from locating the new leader master after a leader failover in the case that the previous leader either remained online or restarted quickly. This bug resulted in the client timing out operations with errors indicating that there was no leader master. (see KUDU-2343).
The Unix process username of the client is now included inside the exported
security credentials, so that the effective username of clients who import
credentials and subsequently use unauthenticated (SASL PLAIN) connections
matches the client who exported the security credentials. For example, this is
useful to let the Spark executors know which username to use if the Spark
driver has no authentication token. This change only affects clusters with
encryption disabled using --rpc_encryption=disabled
.
(see KUDU-2259).
Kudu 1.7.0 is wire-compatible with previous versions of Kudu:
Kudu 1.7 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Rolling upgrade between Kudu 1.6 and Kudu 1.7 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
Kudu 1.0 clients may connect to servers running Kudu 1.7 with the exception of the below-mentioned restrictions regarding secure clusters.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.7 and versions earlier than 1.3:
If a Kudu 1.7 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.7 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
The newly introduced replica management scheme is not compatible with the old scheme, so it’s not possible to run pre-1.7 Kudu masters with 1.7 Kudu tablet servers or vice versa. This is a server-side incompatibility only and it does not affect client compatibility. In other words, Kudu clients of prior versions are compatible with upgraded Kudu clusters.
Kudu masters of 1.7 version will not register Kudu tablet servers of 1.6 and prior versions.
Kudu tablet servers of 1.7 version will not work with Kudu masters of 1.6 and prior versions.
The format of the previously-optional metrics log has changed to include a human-readable timestamp on each line. The path of the log file has also changed with the word “diagnostics” replacing the word “metrics” in the file name. The metrics log has been optimized to only include those metrics which have changed in between successive samples, and to not include entity attributes such as tablet partition information in the log. (see KUDU-2297).
The Kudu 1.7 Java client library is API- and ABI-compatible with Kudu 1.6. Applications written against Kudu 1.6 will compile and run against the Kudu 1.7 client library and vice-versa.
The Kudu 1.7 C++ client is API- and ABI-forward-compatible with Kudu 1.6. Applications written and compiled against the Kudu 1.6 client library will run without modification against the Kudu 1.7 client library. Applications written and compiled against the Kudu 1.7 client library will run without modification against the Kudu 1.6 client library.
The Kudu 1.7 Python client is API-compatible with Kudu 1.6. Applications written against Kudu 1.6 will continue to run against the Kudu 1.7 client and vice-versa.
Kudu 1.7 clients that attempt to create a table with a decimal column on a target server running Kudu 1.6 or earlier will receive an error response. Similarly Kudu clients running Kudu 1.6 or earlier will result in an error when attempting to access any table containing containing a decimal column.
Upgrading directly from Kudu 1.5.0 is supported and no special upgrade steps are required. A rolling upgrade may work, however it has not been tested. When upgrading Kudu, it is recommended to first shut down all Kudu processes across the cluster, then upgrade the software on all servers, then restart the Kudu processes on all servers in the cluster.
Support for Spark 1 (kudu-spark_2.10) has been removed in Kudu 1.6.0 and now only Spark 2 is supported. Spark 1 support was deprecated in Kudu 1.5.0.
Support for Java 7 has been deprecated since Kudu 1.5.0 and may be removed in the next major release.
Tablet servers' tolerance of disk failures is now enabled by default and has
been extended to handle data directory failures at runtime. In the event of
a disk failure at runtime, any tablets with data on a failed disk will be
shut down and restarted on another tablet server. There is a configurable
tradeoff between a newly added tablet’s tolerance to disk failures and its
ability to parallelize reads via the experimental
--fs_target_data_dirs_per_tablet
flag. Tablets that are spread across fewer
disks are less likely to be affected by a disk failure, at the cost of
reduced parallelism. By default, tablets are striped across all available
disks. Note that the first configured data directory and the WAL directory
cannot currently tolerate disk failures. This will be further improved in
future Kudu releases.
Kudu servers can now adopt new data directories via the new
kudu fs update_dirs
tool. The new directory will be used by new tablet
replicas only. Note that removing directories is not yet supported
(see KUDU-2202).
Kudu servers have two new flags to control webui TLS/HTTPS
settings: --webserver_tls_ciphers
and --webserver_tls_min_protocol
.
These flags allow the advertised TLS ciphers and TLS protocol versions to be
configured. Additionally, the webserver now excludes insecure legacy ciphers
by default
(see KUDU-2190).
Kudu servers can now tolerate short interruptions in NTP clock synchronization. NTP synchronization is still required when any Kudu daemon starts up. If NTP synchronization is not available, diagnostic information is now logged to help pinpoint the issue (see KUDU-1578).
Tablet server startup time has been improved significantly on servers containing large numbers of blocks.
The log block manager now performs disk data deletion in batches. This optimization can significantly reduce the time taken to delete data on a tablet.
The usage of sensitive data redaction flag has been slightly changed. By
setting --redact=log
flag, redaction will be disabled in the web UI but
retained for server logs. Alternatively, --redact=none
can be used to
disable redaction completely.
The Spark DataSource integration now can take advantage of scan locality for better scan performance, the scan will take place at the closest replica instead of going to the leader.
Various optimizations were made to reduce the 99th percentile latency of writes on the tablet server. This can also improve throughput on certain write workloads, particularly on larger clusters.
Kudu may now be configured to ignore system-wide auth_to_local mappings
configured in /etc/krb5.conf by setting the configuration flag
--use_system_auth_to_local=false
(see KUDU-2198).
The performance of the compaction scheduler has been improved. In previous versions, certain types of time series workloads were found to cause compaction scheduling to take tens of seconds. These workloads now schedule compactions an order of magnitude more efficiently.
The compaction scheduler has been improved to avoid running a compaction when the benefit of that compaction is extremely small.
Tablet servers now consider the health of all replicas of a tablet before deciding to evict one. This can improve stability of the Kudu cluster after experiencing multiple simultaneous daemon failures (see KUDU-2048).
Several performance improvements have been made to the Kudu master, particularly in concurrency of clients opening tables. This should improve performance in highly concurrent workloads.
The on-disk size metric for a tablet now includes all data and metadata. Previously, it excluded WAL segments and consensus metadata (see KUDU-1755).
Added verbose mode for the 'kudu cluster ksck' command to enable output of detailed information on the cluster’s metadata, even when no errors are detected.
HybridTime timestamp propagation now works in the Java client when using scan tokens (see KUDU-1411).
Fixed an error message commonly found in tablet server logs indicating that operations were being read "from the future" (see KUDU-1078).
Tombstoned tablets no longer report metrics (see KUDU-2044).
Fixed a bug in the C++ client which could cause tablets to be erroneously pruned, or skipped, during certain scans, resulting in fewer results than expected being returned from queries. The bug only affected tables whose range partition columns are a proper prefix of the primary key (see KUDU-2173).
Published Kudu Java artifacts are now fully compatible with JRE 7 and JRE 8. There was previously a bug in the release process which made them compatible only with JRE 8 (see KUDU-2188).
Fixed a typo in the list of default TLS ciphers used by Kudu servers. As a result, two additional cipher suites are now available:
ECDHE-RSA-AES128-SHA256 TLSv1.2 Kx=ECDH Au=RSA Enc=AES(128) Mac=SHA256
AES256-GCM-SHA384 TLSv1.2 Kx=RSA Au=RSA Enc=AESGCM(256) Mac=AEAD
Kudu 1.6.0 is wire-compatible with previous versions of Kudu:
Kudu 1.6 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Rolling upgrade between Kudu 1.5 and Kudu 1.6 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
Kudu 1.0 clients may connect to servers running Kudu 1.6 with the exception of the below-mentioned restrictions regarding secure clusters.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.6 and versions earlier than 1.3:
If a Kudu 1.6 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.6 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
The Kudu 1.6 Java client library is API- and ABI-compatible with Kudu 1.5. Applications written against Kudu 1.5 will compile and run against the Kudu 1.6 client library and vice-versa.
The Kudu 1.6 C++ client is API- and ABI-forward-compatible with Kudu 1.5. Applications written and compiled against the Kudu 1.5 client library will run without modification against the Kudu 1.6 client library. Applications written and compiled against the Kudu 1.6 client library will run without modification against the Kudu 1.5 client library.
The Kudu 1.6 Python client is API-compatible with Kudu 1.5. Applications written against Kudu 1.5 will continue to run against the Kudu 1.6 client and vice-versa.
Kudu 1.5 now enables the optional ability to compute, store, and verify checksums on all pieces of data stored on a server by default. Due to storage format changes, downgrading to versions 1.3 or earlier is not supported and will result in an error.
Spark 2.2+ requires Java 8 at runtime even though Kudu Spark 2.x integration is Java 7 compatible. Spark 2.2 is the default dependency version as of Kudu 1.5.0.
The kudu-spark-tools module has been renamed to kudu-spark2-tools_2.11 in order to include the Spark and Scala base versions. This matches the pattern used in the kudu-spark module and artifacts.
To improve security, world-readable Kerberos keytab files are no longer
accepted by default. Set --allow_world_readable_credentials=true
to override
this behavior. See
KUDU-1955 for additional
details.
Support for Java 7 is deprecated as of Kudu 1.5.0 and may be removed in the next major release.
Support for Spark 1 (kudu-spark_2.10) is deprecated as of Kudu 1.5.0 and may be removed in the next minor release.
Tablet servers are now optionally able to tolerate disk failures at
startup. This feature is experimental; by default, Kudu will crash if it
experiences a disk failure. When enabled, tablets with any data on the failed
disk will not be opened and will be replicated as needed. To enable this, set
the --crash_on_eio
flag to false
. Additionally, there is a configurable
tradeoff between a newly added tablet’s tolerance to disk failures and its
parallelization of I/O via the --fs_target_data_dirs_per_tablet
flag.
Tablets that are spread across fewer disks are less likely to be affected by a
disk failure, at the cost of reduced parallelism. Note that the first
configured data directory and the WAL directory cannot currently tolerate disk
failures, and disk failures during run-time are still fatal.
Kudu server web UIs have a new configuration dashboard (/config) which provides a high level summary of important security configuration values, such as whether RPC authentication is required, or web server HTTPS encryption is enabled. Other types of configuration will be added in future releases.
The kudu
command line tool has two new features: kudu tablet change_config
move_replica
and kudu local_replica data_size
. The 'tablet change_config
move_replica' tool moves a tablet replica from one tablet server to another,
under the condition that the tablet is healthy. An operator can use this tool to
rebalance tablet replicas between tablet servers. The 'local_replica data size'
tool summarizes the space usage of a tablet, breaking it down by type of file,
column, and rowset.
kudu-client-tools now supports exporting CSV files and importing Apache Parquet files. This feature is unstable and may change APIs and functionality in future releases.
kudu-spark-tools now supports importing and exporting CSV, Apache Avro and Apache Parquet files. This feature is unstable and may change APIs and functionality in future releases.
The log block manager now performs disk synchronization in batches. This optimization can significantly reduce the time taken to copy tablet data from one server to another; in one case tablet copy time is reduced by 35%. It also improves the general performance of flushes and compactions.
A new feature referred to as "tombstoned voting" is added to the Raft
consensus subsystem to allow tablet replicas in the TABLET_DATA_TOMBSTONED
state to vote in tablet leader elections. This feature increases Kudu’s
stability and availability by improving the likelihood that Kudu will be able
to self-heal in more edge-case scenarios, such as when tablet copy operations
fail. See KUDU-871 for
details.
The tablet on-disk size metric has been made more accurate. Previously, the metric included only REDO deltas; it now counts all deltas. Additionally, the metric includes the size of bloomfiles, ad hoc indexes, and the tablet superblock. WAL segments and consensus metadata are still not counted. The latter is very small compared to the size of data, but the former may be significant depending on the workload (this will be resolved in a future release).
The number of threads used by the Kudu tablet server has been further reduced. Previously, each follower tablet replica used a dedicated thread to detect leader tablet replica failures, and each leader replica used one dedicated thread per follower to send Raft heartbeats to that follower. The work performed by these dedicated threads has been reassigned to other threads. Other improvements were made to facilitate better thread sharing by tablets. For the purpose of capacity planning, expect the Kudu tablet server to create one thread for every five "cold" (i.e. those not servicing writes) tablets, and an additional three threads for every "hot" tablet. This will be further improved upon in future Kudu releases.
The Java Kudu client now automatically requests new authentication tokens after expiration. As a result, long-lived Java clients are now supported. See KUDU-2013 for more details.
Multiple Kerberos compatibility bugs have been fixed, including support for environments with disabled reverse DNS, FreeIPA compatibility, principal names including uppercase characters, and hosts without a FQDN.
A bug in the binary prefix decoder which could cause a tablet server 'check' assertion crash has been fixed. The crash could only be triggered in very specific scenarios; see KUDU-2085 for additional details.
Kudu 1.5.0 is wire-compatible with previous versions of Kudu:
Kudu 1.5 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Rolling upgrade between Kudu 1.4 and Kudu 1.5 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
Kudu 1.0 clients may connect to servers running Kudu 1.5 with the exception of the below-mentioned restrictions regarding secure clusters.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.5 and versions earlier than 1.3:
If a Kudu 1.5 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.5 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
The Kudu 1.5 Java client library is API- and ABI-compatible with Kudu 1.4. Applications written against Kudu 1.4 will compile and run against the Kudu 1.5 client library and vice-versa, unless one of the following newly added APIs is used:
The Kudu 1.5 C++ client is API- and ABI-forward-compatible with Kudu 1.4. Applications written and compiled against the Kudu 1.4 client library will run without modification against the Kudu 1.5 client library. Applications written and compiled against the Kudu 1.5 client library will run without modification against the Kudu 1.4 client library.
The Kudu 1.5 Python client is API-compatible with Kudu 1.4. Applications written against Kudu 1.4 will continue to run against the Kudu 1.5 client and vice-versa.
The Maintenance Manager now fully uses the threads it’s given (see the improvements described further below), so it’s now able to generate a lot more IO by flushing and compacting more often. Generally, the recommended ratio of MM threads to data directories is 1:3; operators of clusters above that ratio should be mindful of this when upgrading.
The C++ and Java client libraries now support the ability to alter the storage attributes (e.g. encoding and compression) and default value of existing columns. Additionally, it is now possible to rename a column which is part of a table’s primary key.
The C++ client library now includes an experimental KuduPartitioner
API which may
be used to efficiently map rows to their associated partitions and hosts.
This may be used to achieve better locality or distribution of writes
in client applications.
The Java client library now supports enabling fault tolerance on scanners. Fault tolerant scanners are able to transparently recover from concurrent server crashes at the cost of some performance overhead. See the Java API documentation for more details on usage.
The kudu
command line tool now includes a new advanced administrative
command kudu remote_replica unsafe_change_config
. This command may be used
to force a tablet to perform an unsafe change of its Raft replication
configuration. This can be used to recover from scenarios such as a loss
of a majority of replicas, at the risk of losing edits.
The kudu
command line tool now includes the kudu fs check
command
which performs various offline consistency checks on the local on-disk
storage of a Kudu Tablet Server or Master. In addition to detecting
various inconsistencies or corruptions, it can also detect and remove
data blocks that are no longer referenced by any tablet but were not
fully removed from disk due to a crash or a bug in prior versions of Kudu.
The kudu
command line tool can now be used to list the addresses and
identifiers of the servers in the cluster using either kudu master list
or kudu tserver list
.
Kudu 1.4 now includes the optional ability to compute, store, and verify checksums on all pieces of data stored on a server. Prior versions only performed checksums on certain portions of the stored data. This feature is not enabled by default since it makes a backward-incompatible change to the on-disk formats and thus prevent downgrades. Kudu 1.5 will enable the feature by default.
kudu cluster ksck
now detects and reports new classes of
inconsistencies and issues. In particular, it is better able to
detect cases where a configuration change such as a replica eviction
or addition is pending but is unable to be committed. It also now
properly detects and reports cases where a tablet has no elected
leader.
The default size for Write Ahead Log (WAL) segments has been reduced from 64MB to 8MB. Additionally, in the case that all replicas of a tablet are fully up to date and data has been flushed from memory, servers will now retain only a single WAL segment rather than two. These changes are expected to reduce the average consumption of disk space on the configured WAL disk by 16x, as well as improve the startup speed of tablet servers by reducing the number and size of WAL segments that need to be re-read.
The default on-disk storage system used by Kudu servers (Log Block Manager) has been improved to compact its metadata and remove dead containers. This compaction and garbage collection occurs only at startup. Thus, the first startup after upgrade is expected to be longer than usual, and subsequent restarts should be shorter.
The usability of the Kudu web interfaces has been improved, particularly for the case where a server hosts many tablets or a table has many partitions. Pages that list tablets now include a top-level summary of tablet status and show the complete list under a toggleable section.
The Maintenance Manager has been improved to improve utilization of the configured maintenance threads. Previously, maintenance work would only be scheduled a maximum of 4 times per second, but now maintenance work will be scheduled immediately whenever any configured thread is available. This can improve the throughput of write-heavy workloads.
The Maintenance Manager will now aggressively schedule flushes of in-memory data when memory consumption crosses 60% of the configured process-wide memory limit. The backpressure mechanism which begins to throttle client writes has been accordingly adjusted to not begin throttling until reaching 80% of the configured limit. These two changes together result in improved write throughput, more consistent latency, and fewer timeouts due to memory exhaustion.
Many performance improvements were made to write performance. Applications which send large batches of writes to Kudu should see substantially improved throughput in Kudu 1.4.
Several improvements were made to reduce the memory consumption of Kudu Tablet Servers which hold large volumes of data. The specific amount of memory saved varies depending on workload, but the expectation is that approximately 350MB of excess peak memory usage has been eliminated per TB of data stored.
The number of threads used by the Kudu Tablet Server has been reduced. Previously, each tablet used a dedicated thread to append to its WAL. Those threads now automatically stop running if there is no activity on a given tablet for a short period of time.
KUDU-2020 Fixed an issue where re-replication after a failure would proceed significantly slower than expected. This bug caused many tablets to be unnecessarily copied multiple times before successfully being considered re-replicated, resulting in significantly more network and IO bandwidth usage than expected. Mean time to recovery on clusters with large amounts of data is improved by up to 10x by this fix.
KUDU-1982
Fixed an issue where the Java client would call NetworkInterface.getByInetAddress
very often, causing performance problems particularly on Windows
where this function can be quite slow.
KUDU-1755
Improved the accuracy of the on_disk_size
replica metrics to
include the size consumed by bloom filters, primary key indexes,
and superblock metadata, and delta files. Note that, because the size
metric is now more accurate, the reported values are expected to
increase after upgrading to Kudu 1.4. This does not indicate that
replicas are using more space after the upgrade; rather, it is
now accurately reporting the amount of space that has always been
used.
KUDU-1192
Kudu servers will now periodically flush their log messages to disk
even if no WARNING
-level messages have been logged. This makes it
easier to tail the logs to see progress output during normal startup.
KUDU-1999 Fixed the ability to run Spark jobs in "cluster" mode against Kudu clusters secured by Kerberos.
Kudu 1.4.0 is wire-compatible with previous versions of Kudu:
Kudu 1.4 clients may connect to servers running Kudu 1.0 or later. If the client uses features that are not available on the target server, an error will be returned.
Kudu 1.0 clients may connect to servers running Kudu 1.4 with the exception of the below-mentioned restrictions regarding secure clusters.
Rolling upgrade between Kudu 1.3 and Kudu 1.4 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
The authentication features introduced in Kudu 1.3 place the following limitations on wire compatibility between Kudu 1.4 and versions earlier than 1.3:
If a Kudu 1.4 cluster is configured with authentication or encryption set to "required", clients older than Kudu 1.3 will be unable to connect.
If a Kudu 1.4 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
Kudu servers, by default, will now only allow unencrypted or unauthenticated connections from trusted subnets, which are private networks (127.0.0.0/8,10.0.0.0/8,172.16.0.0/12, 192.168.0.0/16,169.254.0.0/16) and local subnets of all local network interfaces. Unencrypted or unauthenticated connections from publicly routable IPs will be rejected, even if encryption and authentication are not configured.
The trusted subnets can be configured using the --trusted_subnets
flag, which can be set
to IP blocks represented in CIDR notation separated by comma. Set it to '0.0.0.0/0' to
allow unauthenticated connections from all remote IP addresses. However, if network access
is not otherwise restricted by a firewall, malicious users may be able to gain unauthorized
access. This can be mitigated if authentication and encryption are configured to be
required.
The Kudu 1.4 Java client library is API- and ABI-compatible with Kudu 1.3. Applications written against Kudu 1.3 will compile and run against the Kudu 1.4 client library and vice-versa, unless one of the following newly added APIs is used:
[Async]KuduScannerBuilder.setFaultTolerant(…)
New methods in AlterTableOptions
: removeDefault
, changeDefault
, changeDesiredBlockSize
,
changeEncoding
, changeCompressionAlgorithm
KuduClient.updateLastPropagatedTimestamp
KuduClient.getLastPropagatedTimestamp
New getters in PartialRow
: getBoolean
, getByte
, getShort
, getInt
, getLong
,
getFloat
, getDouble
, getString
, getBinaryCopy
, getBinary
, isNull
,
isSet
.
The Kudu 1.4 C++ client is API- and ABI-forward-compatible with Kudu 1.3. Applications written and compiled against the Kudu 1.3 client library will run without modification against the Kudu 1.4 client library. Applications written and compiled against the Kudu 1.4 client library will run without modification against the Kudu 1.3 client library unless they use one of the following new APIs:
KuduPartitionerBuilder
`KuduPartitioner
KuduScanner::SetRowFormatFlags
(unstable API)
KuduScanBatch::direct_data
, KuduScanBatch::indirect_data
(unstable API)
The Kudu 1.4 Python client is API-compatible with Kudu 1.3. Applications written against Kudu 1.3 will continue to run against the Kudu 1.4 client and vice-versa.
Kudu 1.3 adds support for strong authentication based on Kerberos. This optional feature allows users to authenticate themselves using Kerberos tickets, and also provides mutual authentication of servers using Kerberos credentials stored in keytabs. This feature is optional, but recommended for deployments requiring security.
Kudu 1.3 adds support for encryption of data on the network using Transport Layer Security (TLS). Kudu will now use TLS to encrypt all network traffic between clients and servers as well as any internal traffic among servers, with the exception of traffic determined to be within a localhost network connection. Encryption is enabled by default whenever it can be determined that both the client and server support the feature.
Kudu 1.3 adds coarse-grained service-level authorization of access to the cluster. The operator may set up lists of permitted users who may act as administrators and as clients of the cluster. Combined with the strong authentication feature described above, this can enable a secure environment for some use cases. Note that fine-grained access control (e.g. table-level or column-level) is not yet supported.
Kudu 1.3 adds a background task to tablet servers which removes historical versions of data which have fallen behind the configured data retention time. This reduces disk space usage in all workloads, but particularly in those with a higher volume of updates or upserts.
Kudu now incorporates Google Breakpad, a library which writes crash reports in the case of a server crash. These reports can be found within the configured log directory, and can be useful during bug diagnosis.
Kudu servers will now change the file permissions of data directories and contained
data files based on a new configuration flag --umask
. As a result, after upgrading,
permissions on disk may be more restrictive than in previous versions. The new default
configuration improves data security.
Kudu’s web UI will now redact strings which may include sensitive user data. For example, the monitoring page which shows in-progress scans no longer includes the scanner predicate values. The tracing and RPC diagnostics endpoints no longer include contents of RPCs which may include table data.
By default, Kudu now reserves 1% of each configured data volume as free space. If a volume is seen to have less than 1% of disk space free, Kudu will stop writing to that volume to avoid completely filling up the disk.
The default encoding for numeric columns (int, float, and double) has been changed
to BIT_SHUFFLE
. The default encoding for binary and string columns has been
changed to DICT_ENCODING
. Dictionary encoding automatically falls back to the old
default (PLAIN
) when cardinality is too high to be effectively encoded.
These new defaults match the default behavior of other storage mechanisms such as Apache Parquet and are likely to perform better out of the box.
Kudu now uses LZ4
compression when writing its Write Ahead Log (WAL). This improves
write performance and stability for many use cases.
Kudu now uses LZ4
compression when writing delta files. This can improve both
read and write performance as well as save substantial disk usage, especially
for workloads involving a high number of updates or upserts containing compressible
data.
The Kudu API now supports the ability to express IS NULL
and IS NOT NULL
predicates
on scanners. The Spark DataSource integration will take advantage of these new
predicates when possible.
Both C++ and Java clients have been optimized to prune partitions more effectively
when performing scans using the IN (…)
predicate.
The exception messages produced by the Java client are now truncated to a maximum length of 32KB.
KUDU-1893
Fixed a critical bug in which wrong results would be returned when evaluating
predicates applied to columns added using the ALTER TABLE
operation.
KUDU-1905 Fixed a crash after inserting a row sharing a primary key with a recently-deleted row in tables where the primary key is comprised of all of the columns.
KUDU-1899 Fixed a crash after inserting a row with an empty string as the single-column primary key.
KUDU-1904 Fixed a potential crash when performing random reads against a column using RLE encoding and containing long runs of NULL values.
KUDU-1853 Fixed an issue where disk space could be leaked on servers which experienced an error during the process of copying tablet data from another server.
KUDU-1856 Fixed an issue in which disk space could be leaked by Kudu servers storing data on partitions using the XFS file system. Any leaked disk space will be automatically recovered upon upgrade.
KUDU-1888, KUDU-1906 Fixed multiple issues in the Java client where operation callbacks would never be triggered, causing the client to hang.
Kudu 1.3.0 is wire-compatible with previous versions of Kudu:
Kudu 1.3 clients may connect to servers running Kudu 1.0. If the client uses features that are not available on the target server, an error will be returned.
Kudu 1.0 clients may connect to servers running Kudu 1.3 with the exception of the below-mentioned restrictions regarding secure clusters.
Rolling upgrade between Kudu 1.2 and Kudu 1.3 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
The authentication features newly introduced in Kudu 1.3 place the following limitations on wire compatibility with older versions:
If a Kudu 1.3 cluster is configured with authentication or encryption set to "required", older clients will be unable to connect.
If a Kudu 1.3 cluster is configured with authentication and encryption set to "optional" or "disabled", older clients will still be able to connect.
Due to storage format changes in Kudu 1.3, downgrade from Kudu 1.3 to earlier versions is not supported. After upgrading to Kudu 1.3, attempting to restart with an earlier version will result in an error.
In order to support running MapReduce and Spark jobs on secure clusters, these frameworks now connect to the cluster at job submission time to retrieve authentication credentials which can later be used by the tasks to be spawned. This means that the process submitting jobs to Kudu clusters must have direct access to that cluster.
The embedded web servers in Kudu processes now specify the X-Frame-Options: DENY
HTTP
header which prevents embedding Kudu web pages in HTML iframe
elements.
The Kudu 1.3 Java client library is API- and ABI-compatible with Kudu 1.2. Applications written against Kudu 1.2 will compile and run against the Kudu 1.3 client library and vice-versa, unless one of the following newly added APIs is used:
[Async]KuduClient.exportAuthenticationCredentials(…)
(unstable API)
[Async]KuduClient.importAuthenticationCredentials(…)
(unstable API)
[Async]KuduClient.getMasterAddressesAsString()
KuduPredicate.newIsNotNullPredicate()
KuduPredicate.newIsNullPredicate()
The Kudu 1.3 C++ client is API- and ABI-forward-compatible with Kudu 1.2. Applications written and compiled against the Kudu 1.2 client library will run without modification against the Kudu 1.3 client library. Applications written and compiled against the Kudu 1.3 client library will run without modification against the Kudu 1.2 client library unless they use one of the following new APIs:
kudu::DisableOpenSSLInitialization()
KuduClientBuilder::import_authentication_credentials(…)
KuduClient::ExportAuthenticationCredentials(…)
KuduClient::NewIsNotNullPredicate(…)
KuduClient::NewIsNullPredicate(…)
The Kudu 1.3 Python client is API-compatible with Kudu 1.2. Applications written against Kudu 1.2 will continue to run against the Kudu 1.3 client and vice-versa.
Kudu clients and servers now redact user data such as cell values
from log messages, Java exception messages, and Status
strings.
User metadata such as table names, column names, and partition
bounds are not redacted.
Redaction is enabled by default, but may be disabled by setting the new
log_redact_user_data
flag to false
.
Kudu’s ability to provide consistency guarantees has been substantially improved:
Replicas now correctly track their "safe timestamp". This timestamp is the maximum timestamp at which reads are guaranteed to be repeatable.
A scan created using the SCAN_AT_SNAPSHOT
mode will now
either wait for the requested snapshot to be "safe" at the replica
being scanned, or be re-routed to a replica where the requested
snapshot is "safe". This ensures that all such scans are repeatable.
Kudu Tablet Servers now properly retain historical data when a row with a given primary key is inserted and deleted, followed by the insertion of a new row with the same key. Previous versions of Kudu would not retain history in such situations. This allows the server to return correct results for snapshot scans with a timestamp in the past, even in the presence of such "reinsertion" scenarios.
The Kudu clients now automatically retain the timestamp of their latest
successful read or write operation. Scans using the READ_AT_SNAPSHOT
mode
without a client-provided timestamp automatically assign a timestamp
higher than the timestamp of their most recent write. Writes also propagate
the timestamp, ensuring that sequences of operations with causal dependencies
between them are assigned increasing timestamps. Together, these changes
allow clients to achieve read-your-writes consistency, and also ensure
that snapshot scans performed by other clients return causally-consistent
results.
Kudu servers now automatically limit the number of log files.
The number of log files retained can be configured using the
max_log_files
flag. By default, 10 log files will be retained
at each severity level.
The logging in the Java and C++ clients has been substantially quieted. Clients no longer log messages in normal operation unless there is some kind of error.
The C++ client now includes a KuduSession::SetErrorBufferSpace
API which can limit the amount of memory used to buffer
errors from asynchronous operations.
The Java client now fetches tablet locations from the Kudu Master in batches of 1000, increased from batches of 10 in prior versions. This can substantially improve the performance of Spark and Impala queries running against Kudu tables with large numbers of tablets.
Table metadata lock contention in the Kudu Master was substantially reduced. This improves the performance of tablet location lookups on large clusters with a high degree of concurrency.
Lock contention in the Kudu Tablet Server during high-concurrency write workloads was also reduced. This can reduce CPU consumption and improve performance when a large number of concurrent clients are writing to a smaller number of a servers.
Lock contention when writing log messages has been substantially reduced. This source of contention could cause high tail latencies on requests, and when under high load could contribute to cluster instability such as election storms and request timeouts.
The BITSHUFFLE
column encoding has been optimized to use the AVX2
instruction set present on processors including Intel® Sandy Bridge
and later. Scans on BITSHUFFLE
-encoded columns are now up to 30% faster.
The kudu
tool now accepts hyphens as an alternative to underscores
when specifying actions. For example, kudu local-replica copy-from-remote
may be used as an alternative to kudu local_replica copy_from_remote
.
KUDU-1508
Fixed a long-standing issue in which running Kudu on ext4
file systems
could cause file system corruption.
KUDU-1399
Implemented an LRU cache for open files, which prevents running out of
file descriptors on long-lived Kudu clusters. By default, Kudu will
limit its file descriptor usage to half of its configured ulimit
.
Gerrit #5192
Fixed an issue which caused data corruption and crashes in the case that
a table had a non-composite (single-column) primary key, and that column
was specified to use DICT_ENCODING
or BITSHUFFLE
encodings. If a
table with an affected schema was written in previous versions of Kudu,
the corruption will not be automatically repaired; users are encouraged
to re-insert such tables after upgrading to Kudu 1.2 or later.
Gerrit #5541
Fixed a bug in the Spark KuduRDD
implementation which could cause
rows in the result set to be silently skipped in some cases.
KUDU-1551 Fixed an issue in which the tablet server would crash on restart in the case that it had previously crashed during the process of allocating a new WAL segment.
KUDU-1764 Fixed an issue where Kudu servers would leak approximately 16-32MB of disk space for every 10GB of data written to disk. After upgrading to Kudu 1.2 or later, any disk space leaked in previous versions will be automatically recovered on startup.
KUDU-1750 Fixed an issue where the API to drop a range partition would drop any partition with a matching lower or upper bound, rather than any partition with matching lower and upper bound.
KUDU-1766
Fixed an issue in the Java client where equality predicates which compared
an integer column to its maximum possible value (e.g. Integer.MAX_VALUE
)
would return incorrect results.
KUDU-1780
Fixed the kudu-client
Java artifact to properly shade classes in the
com.google.thirdparty
namespace. The lack of proper shading in prior
releases could cause conflicts with certain versions of Google Guava.
Gerrit #5327
Fixed shading issues in the kudu-flume-sink
Java artifact. The sink
now expects that Hadoop dependencies are provided by Flume, and properly
shades the Kudu client’s dependencies.
Fixed a few issues using the Python client library from Python 3.
Kudu 1.2.0 is wire-compatible with previous versions of Kudu:
Kudu 1.2 clients may connect to servers running Kudu 1.0. If the client uses features that are not available on the target server, an error will be returned.
Kudu 1.0 clients may connect to servers running Kudu 1.2 without limitations.
Rolling upgrade between Kudu 1.1 and Kudu 1.2 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
The replication factor of tables is now limited to a maximum of 7. In addition, it is no longer allowed to create a table with an even replication factor.
The GROUP_VARINT
encoding is now deprecated. Kudu servers have never supported
this encoding, and now the client-side constant has been deprecated to match the
server’s capabilities.
Kudu 1.2.0 introduces several new restrictions on schemas, cell size, and identifiers:
By default, Kudu will not permit the creation of tables with more than 300 columns. We recommend schema designs that use fewer columns for best performance.
No individual cell may be larger than 64KB. The cells making up a a composite key are limited to a total of 16KB after the internal composite-key encoding done by Kudu. Inserting rows not conforming to these limitations will result in errors being returned to the client.
Identifiers such as column and table names are now restricted to be valid UTF-8 strings. Additionally, a maximum length of 256 characters is enforced.
The Kudu 1.2 Java client is API- and ABI-compatible with Kudu 1.1. Applications written against Kudu 1.1 will compile and run against the Kudu 1.2 client and vice-versa.
The Kudu 1.2 C++ client is API- and ABI-forward-compatible with Kudu 1.1. Applications written and compiled against the Kudu 1.1 client will run without modification against the Kudu 1.2 client. Applications written and compiled against the Kudu 1.2 client will run without modification against the Kudu 1.1 client unless they use one of the following new APIs:
kudu::DisableSaslInitialization()
KuduSession::SetErrorBufferSpace(…)
The Kudu 1.2 Python client is API-compatible with Kudu 1.1. Applications written against Kudu 1.1 will continue to run against the Kudu 1.2 client and vice-versa.
The Python client has been brought up to feature parity with the Java and C++ clients and as such the package version will be brought to 1.1 with this release (from 0.3). A list of the highlights can be found below.
Improved Partial Row semantics
Range partition support
Scan Token API
Enhanced predicate support
Support for all Kudu data types (including a mapping of Python’s datetime.datetime
to
UNIXTIME_MICROS
)
Alter table support
Enabled Read at Snapshot for Scanners
Enabled Scanner Replica Selection
A few bug fixes for Python 3 in addition to various other improvements.
IN LIST predicate pushdown support was added to allow optimized execution of filters which match on a set of column values. Support for Spark, Map Reduce and Impala queries utilizing IN LIST pushdown is not yet complete.
The Java client now features client-side request tracing in order to help troubleshoot timeouts. Error messages are now augmented with traces that show which servers were contacted before the timeout occurred instead of just the last error. The traces also contain RPCs that were required to fulfill the client’s request, such as contacting the master to discover a tablet’s location. Note that the traces are not available for successful requests and are not programmatically queryable.
Kudu now publishes JAR files for Spark 2.0 compiled with Scala 2.11 along with the existing Spark 1.6 JAR compiled with Scala 2.10.
The Java client now allows configuring scanners to read from the closest replica instead of
the known leader replica. The default remains the latter. Use the relevant ReplicaSelection
enum with the scanner’s builder to change this behavior.
Tablet servers use a new policy for retaining write-ahead log (WAL) segments. Previously, servers used the 'log_min_segments_to_retain' flag to prioritize any flushes which were retaining log segments past the configured value (default 2). This policy caused servers to flush in-memory data more frequently than necessary, limiting write performance.
The new policy introduces a new flag 'log_target_replay_size_mb' which determines the threshold at which write-ahead log retention will prioritize flushes. The new flag is considered experimental and users should not need to modify its value.
The improved policy has been seen to improve write performance in some use cases by a factor of 2x relative to the old policy.
Kudu’s implementation of the Raft consensus algorithm has been improved to include a "pre-election" phase. This can improve the stability of tablet leader election in high-load scenarios, especially if each server hosts a high number of tablets.
Tablet server start-up time has been substantially improved in the case that the server contains a high number of tombstoned tablet replicas.
The tool kudu tablet leader_step_down
has been added to manually force a leader to step down.
The tool kudu remote_replica copy
has been added to manually copy a replica from
one running tablet server to another.
The tool kudu local_replica delete
has been added to delete a replica of a tablet.
The kudu test loadgen
tool has been added to replace the obsoleted
insert-generated-rows
standalone binary. The new tool is enriched with
additional functionality and can be used to run load generation tests against
a Kudu cluster.
Kudu 1.1.0 is wire-compatible with previous versions of Kudu:
Kudu 1.1 clients may connect to servers running Kudu 1.0. If the client uses the new 'IN LIST' predicate type, an error will be returned.
Kudu 1.0 clients may connect to servers running Kudu 1.1 without limitations.
Rolling upgrade between Kudu 1.0 and Kudu 1.1 servers is believed to be possible though has not been sufficiently tested. Users are encouraged to shut down all nodes in the cluster, upgrade the software, and then restart the daemons on the new version.
The C++ client no longer requires the old gcc5 ABI. Which ABI is actually used depends on the compiler configuration. Some new distros (e.g. Ubuntu 16.04) will use the new ABI. Your application must use the same ABI as is used by the client library; an easy way to guarantee this is to use the same compiler to build both.
The C++ client’s KuduSession::CountBufferedOperations()
method is
deprecated. Its behavior is inconsistent unless the session runs in the
MANUAL_FLUSH
mode. Instead, to get number of buffered operations, count
invocations of the KuduSession::Apply()
method since last
KuduSession::Flush()
call or, if using asynchronous flushing, since last
invocation of the callback passed into KuduSession::FlushAsync()
.
The Java client’s OperationResponse.getWriteTimestamp
method was renamed to getWriteTimestampRaw
to emphasize that it doesn’t return milliseconds, unlike what its Javadoc indicated. The renamed
method was also hidden from the public APIs and should not be used.
The Java client’s sync API (KuduClient
, KuduSession
, KuduScanner
) used to throw either
a NonRecoverableException
or a TimeoutException
for a timeout, and now it’s only possible for the
client to throw the former.
The Java client’s handling of errors in KuduSession
was modified so that subclasses of
KuduException
are converted into RowErrors instead of being thrown.
Apache Kudu 1.0.1 is a bug fix release, with no new features or backwards incompatible changes.
KUDU-1681 Fixed a bug in the tablet server which could cause a crash when the DNS lookup during master heartbeat failed.
KUDU-1660: Fixed a bug which would cause the Kudu master and tablet server to fail to start on single CPU systems.
KUDU-1652: Fixed a bug
that would cause the C++ client, tablet server, and Java client to crash or
throw an exception when attempting to scan a table with a predicate which
simplifies to IS NOT NULL
on a non-nullable column. For instance, setting a
⇐ 127
predicate on an INT8
column could trigger this bug, since the
predicate only filters null values.
KUDU-1651: Fixed a bug that would cause the tablet server to crash when evaluating a scan with predicates over a dictionary encoded column containing an entire block of null values.
KUDU-1623: Fixed a bug that would cause the tablet server to crash when handling UPSERT operations that only set values for the primary key columns.
Gerrit #4488 Fixed a bug in the Java client’s KuduException class which could cause an unexpected NullPointerException to be thrown when the exception did not have an associated message.
KUDU-1090 Fixed a bug in the memory tracker which could cause a rare crash during tablet server startup.
After approximately a year of beta releases, Apache Kudu has reached version 1.0. This version number signifies that the development team feels that Kudu is stable enough for usage in production environments.
If you are new to Kudu, check out its list of features and benefits.
Kudu 1.0.0 delivers a number of new features, bug fixes, and optimizations.
Removal of multiversion concurrency control (MVCC) history is now supported. This is known as tablet history GC. This allows Kudu to reclaim disk space, where previously Kudu would keep a full history of all changes made to a given table since the beginning of time. Previously, the only way to reclaim disk space was to drop a table.
Kudu will still keep historical data, and the amount of history retained is
controlled by setting the configuration flag --tablet_history_max_age_sec
,
which defaults to 15 minutes (expressed in seconds). The timestamp
represented by the current time minus tablet_history_max_age_sec
is known
as the ancient history mark (AHM). When a compaction or flush occurs, Kudu
will remove the history of changes made prior to the ancient history mark.
This only affects historical data; currently-visible data will not be
removed. A specialized maintenance manager background task to remove existing
"cold" historical data that is not in a row affected by the normal compaction
process will be added in a future release.
Most of Kudu’s command line tools have been consolidated under a new
top-level kudu
tool. This reduces the number of large binaries distributed
with Kudu and also includes much-improved help output.
The Kudu Flume Sink now supports processing events containing Avro-encoded
records, using the new AvroKuduOperationsProducer
.
Administrative tools including kudu cluster ksck
now support running
against multi-master Kudu clusters.
The output of the ksck
tool is now colorized and much easier to read.
The C++ client API now supports writing data in AUTO_FLUSH_BACKGROUND
mode.
This can provide higher throughput for ingest workloads.
The performance of comparison predicates on dictionary-encoded columns has been substantially optimized. Users are encouraged to use dictionary encoding on any string or binary columns with low cardinality, especially if these columns will be filtered with predicates.
The Java client is now able to prune partitions from scanners based on the provided predicates. For example, an equality predicate on a hash-partitioned column will now only access those tablets that could possibly contain matching data. This is expected to improve performance for the Spark integration as well as applications using the Java client API.
The performance of compaction selection in the tablet server has been substantially improved. This can increase the efficiency of the background maintenance threads and improve overall throughput of heavy write workloads.
The policy by which the tablet server retains write-ahead log (WAL) files has been improved so that it takes into account other replicas of the tablet. This should help mitigate the spurious eviction of tablet replicas on machines that temporarily lag behind the other replicas.
Kudu 1.0.0 maintains client-server wire-compatibility with previous releases. Applications using the Kudu client libraries may be upgraded either before, at the same time, or after the Kudu servers.
Kudu 1.0.0 does not maintain server-server wire compatibility with previous releases. Therefore, rolling upgrades between earlier versions of Kudu and Kudu 1.0.0 are not supported.
The kudu-pbc-dump
tool has been removed. The same functionality is now
implemented as kudu pbc dump
.
The kudu-ksck
tool has been removed. The same functionality is now
implemented as kudu cluster ksck
.
The cfile-dump
tool has been removed. The same functionality is now
implemented as kudu fs cfile dump
.
The log-dump
tool has been removed. The same functionality is now
implemented as kudu wal dump
and kudu local_replica dump wals
.
The kudu-admin
tool has been removed. The same functionality is now
implemented within kudu table
and kudu tablet
.
The kudu-fs_dump
tool has been removed. The same functionality is now
implemented as kudu fs dump
.
The kudu-ts-cli
tool has been removed. The same functionality is now
implemented within kudu master
, kudu remote_replica
, and kudu tserver
.
The kudu-fs_list
tool has been removed and some similar useful
functionality has been moved under 'kudu local_replica'.
Some configuration flags are now marked as 'unsafe' and 'experimental'. Such flags
are disallowed by default. Users may access these flags by enabling the additional
flags --unlock_unsafe_flags
and --unlock_experimental_flags
. Usage of such flags
is not recommended, as the flags may be removed or modified with no deprecation period
and without notice in future Kudu releases.
The TIMESTAMP
column type has been renamed to UNIXTIME_MICROS
in order to
reduce confusion between Kudu’s timestamp support and the timestamps supported
by other systems such as Apache Hive and Apache Impala (incubating). Existing
tables will automatically be updated to use the new name for the type.
Clients upgrading to the new client libraries must move to the new name for the type. Clients using old client libraries will continue to operate using the old type name, even when connected to clusters that have been upgraded. Similarly, if clients are upgraded before servers, existing timestamp columns will be available using the new type name.
KuduSession
methods in the C++ library are no longer advertised as thread-safe
to have one set of semantics for both C++ and Java Kudu client libraries.
The KuduScanToken::TabletServers
method in the C++ library has been removed.
The same information can now be found in the KuduScanToken::tablet method.
The KuduEventProducer
interface used to process Flume events into Kudu operations
for the Kudu Flume Sink has changed, and has been renamed KuduOperationsProducer
.
The existing `KuduEventProducer`s have been updated for the new interface, and have
been renamed similarly.
Kudu 0.10.0 delivers a number of new features, bug fixes, and optimizations, detailed below.
Kudu 0.10.0 maintains wire-compatibility with previous releases, meaning that applications using the Kudu client libraries may be upgraded either before, at the same time, or after the Kudu servers. However, if you begin using new features of Kudu 0.10.0 such as manually range-partitioned tables, you must first upgrade all clients to this release.
This release does not maintain full Java API or ABI compatibility with Kudu 0.9.x due to a package rename and some other small changes. See below for details.
Gerrit #3737 The Java client has been repackaged
under org.apache.kudu
instead of org.kududb
. Import statements for Kudu classes must
be modified in order to compile against 0.10.0. Wire compatibility is maintained.
Gerrit #3055 The Java client’s
synchronous API methods now throw KuduException
instead of Exception
.
Existing code that catches Exception
should still compile, but introspection of an
exception’s message may be impacted. This change was made to allow thrown exceptions to be
queried more easily using KuduException.getStatus
and calling one of Status’s methods.
For example, an operation that tries to delete a table that doesn’t exist would return a
`Status
that returns true when queried on isNotFound()
.
The Java client’s KuduTable.getTabletsLocations
set of methods is now
deprecated. Additionally, they now take an exclusive end partition key instead
of an inclusive key. Applications are encouraged to use the scan tokens API
instead of these methods in the future.
The C++ API for specifying split points on range-partitioned tables has been improved to make it easier for callers to properly manage the ownership of the provided rows.
The TableCreator::split_rows
API took a vector<const KuduPartialRow*>
, which
made it very difficult for the calling application to do proper error handling with
cleanup when setting the fields of the KuduPartialRow
. This API has been now been
deprecated and replaced by a new method TableCreator::add_range_split
which allows
easier use of smart pointers for safe memory management.
The Java client’s internal buffering has been reworked. Previously, the number of buffered write operations was constrained on a per-tablet-server basis. Now, the configured maximum buffer size constrains the total number of buffered operations across all tablet servers in the cluster. This provides a more consistent bound on the memory usage of the client regardless of the size of the cluster to which it is writing.
This change can negatively affect the write performance of Java clients which rely on
buffered writes. Consider using the setMutationBufferSpace
API to increase a
session’s maximum buffer size if write performance seems to be degraded after upgrading
to Kudu 0.10.0.
The "remote bootstrap" process used to copy a tablet replica from one host to another has been renamed to "Tablet Copy". This resulted in the renaming of several RPC metrics. Any users previously explicitly fetching or monitoring metrics related to Remote Bootstrap should update their scripts to reflect the new names.
The SparkSQL datasource for Kudu no longer supports mode Overwrite
. Users should
use the new KuduContext.upsertRows
method instead. Additionally, inserts using the
datasource are now upserts by default. The older behavior can be restored by setting
the operation
parameter to insert
.
Users may now manually manage the partitioning of a range-partitioned table. When a table is created, the user may specify a set of range partitions that do not cover the entire available key space. A user may add or drop range partitions to existing tables.
This feature can be particularly helpful with time series workloads in which new partitions can be created on an hourly or daily basis. Old partitions may be efficiently dropped if the application does not need to retain historical data past a certain point.
This feature is considered experimental for the 0.10 release. More details of the new feature can be found in the accompanying blog post.
Support for running Kudu clusters with multiple masters has been stabilized. Users may start a cluster with three or five masters to provide fault tolerance despite a failure of one or two masters, respectively.
Note that certain tools (e.g. ksck
) are still lacking complete support for
multiple masters. These deficiencies will be addressed in a following release.
Kudu now supports the ability to reserve a certain amount of free disk space in each of its configured data directories. If a directory’s free disk space drops to less than the configured minimum, Kudu will stop writing to that directory until space becomes available. If no space is available in any configured directory, Kudu will abort.
This feature may be configured using the fs_data_dirs_reserved_bytes
and
fs_wal_dir_reserved_bytes
flags.
The Spark integration’s KuduContext
now supports four new methods for writing to
Kudu tables: insertRows
, upsertRows
, updateRows
, and deleteRows
. These are
now the preferred way to write to Kudu tables from Spark.
KUDU-1516 The kudu-ksck
tool
has been improved and now detects problems such as when a tablet does not have
a majority of replicas on live tablet servers, or if those replicas aren’t in a
good state. Users who currently depend on the tool to detect inconsistencies may now see
failures when before they wouldn’t see any.
Gerrit #3477 The way operations are buffered in the Java client has been reworked. Previously, the session’s buffer size was set per tablet, meaning that a buffer size of 1,000 for 10 tablets being written to allowed for 10,000 operations to be buffered at the same time. With this change, all the tablets share one buffer, so users might need to set a bigger buffer size in order to reach the same level of performance as before.
Gerrit #3674 Added LESS and GREATER options for column predicates.
KUDU-1444 added support for passing
back basic per-scan metrics (e.g cache hit rate) from the server to the C++ client. See the
KuduScanner::GetResourceMetrics()
API for detailed usage. This feature will be supported
in the Java client API in a future release.
KUDU-1446 improved the order in which the tablet server evaluates predicates, so that predicates on smaller columns are evaluated first. This may improve performance on queries which apply predicates on multiple columns of different sizes.
KUDU-1398 improved the storage efficiency of Kudu’s internal primary key indexes. This optimization should decrease space usage and improve random access performance, particularly for workloads with lengthy primary keys.
Gerrit #3541 Fixed a problem in the Java client
whereby an RPC could be dropped when a connection to a tablet server or master was forcefully
closed on the server-side while RPCs to that server were in the process of being encoded.
The effect was that the RPC would not be sent, and users of the synchronous API would receive
a TimeoutException
. Several other Java client bugs which could cause similar spurious timeouts
were also fixed in this release.
Gerrit #3724 Fixed a problem in the Java client whereby an RPC could be dropped when a socket timeout was fired while that RPC was being sent to a tablet server or master. This would manifest itself in the same way Gerrit #3541.
KUDU-1538 fixed a bug in which recycled block identifiers could cause the tablet server to lose data. Following this bug fix, block identifiers will no longer be reused.
This is the first release of Apache Kudu as a top-level (non-incubating) project!
The default false positive rate for Bloom filters has been changed from 1% to 0.01%. This will increase the space consumption of Bloom filters by a factor of two (from approximately 10 bits per row to approximately 20 bits per row). This is expected to substantially improve the performance of random-write workloads at the cost of an incremental increase in disk space usage.
The Kudu C++ client library now has Doxygen-based API documentation available online.
Kudu now uses the Raft consensus algorithm even for unreplicated tables. This change simplifies code and will also allow administrators to enable replication on a previously-unreplicated table. This change is internal and should not be visible to users.
Kudu 0.9.1 delivers incremental bug fixes over Kudu 0.9.0. It is fully compatible with Kudu 0.9.0.
KUDU-1469 fixed a bug in our Raft consensus implementation that could cause a tablet to stop making progress after a leader election.
Gerrit #3456 fixed a bug in which servers under high load could store metric information in incorrect memory locations, causing crashes or data corruption.
Gerrit #3457 fixed a bug in which errors from the Java client would carry an incorrect error message.
Several other small bug fixes were backported to improve stability.
Kudu 0.9.0 delivers incremental features, improvements, and bug fixes over the previous versions.
The KuduTableInputFormat
command has changed the way in which it handles
scan predicates, including how it serializes predicates to the job configuration
object. The new configuration key is kudu.mapreduce.encoded.predicate
. Clients
using the TableInputFormatConfigurator
are not affected.
The kudu-spark
sub-project has been renamed to follow naming conventions for
Scala. The new name is kudu-spark_2.10
.
Default table partitioning has been removed. All tables must now be created with explicit partitioning. Existing tables are unaffected. See the schema design guide for more details.
KUDU-1002 Added support for
UPSERT
operations, whereby a row is inserted if it does not already exist, but
updated if it does. Support for UPSERT
is included in Java, C++, and Python APIs,
but not in Impala.
KUDU-1306 Scan token API for creating partition-aware scan descriptors. This API simplifies executing parallel scans for clients and query engines.
Gerrit 2848 Added a kudu datasource
for Spark. This datasource uses the Kudu client directly instead of
using the MapReduce API. Predicate pushdowns for spark-sql
and Spark filters are
included, as well as parallel retrieval for multiple tablets and column projections.
See an example of Kudu integration with Spark.
Gerrit 2992 Added the ability to update and insert from Spark using a Kudu datasource.
KUDU-678 Fixed a leak that happened during DiskRowSet compactions where tiny blocks were still written to disk even if there were no REDO records. With the default block manager, it usually resulted in block containers with thousands of tiny blocks.
KUDU-1437 Fixed a data corruption issue that occured after compacting sequences of negative INT32 values in a column that was configured with RLE encoding.
All Kudu clients have longer default timeout values, as listed below.
The default operation timeout and the default admin operation timeout are now set to 30 seconds instead of 10.
The default socket read timeout is now 10 seconds instead of 5.
The default admin timeout is now 30 seconds instead of 10.
The default RPC timeout is now 10 seconds instead of 5.
The default scan timeout is now 30 seconds instead of 15.
Some default settings related to I/O behavior during flushes and compactions have been changed:
The default for flush_threshold_mb
has been increased from 64MB to 1000MB. The default
cfile_do_on_finish
has been changed from close
to flush
.
Experiments using YCSB indicate that these
values will provide better throughput for write-heavy applications on typical server hardware.
Kudu 0.8.0 delivers incremental features, improvements, and bug fixes over the previous versions.
To upgrade to Kudu 0.8.0, see Upgrade from 0.7.1 to 0.8.0.
0.8.0 clients are not fully compatible with servers running Kudu 0.7.1 or lower. In particular, scans that specify column predicates will fail. To work around this issue, upgrade all Kudu servers before upgrading clients.
KUDU-431 A simple Flume sink has been implemented.
KUDU-839 Java RowError now uses an enum error code.
Gerrit 2138 The handling of column predicates has been re-implemented in the server and clients.
KUDU-1379 Partition pruning has been implemented for C++ clients (but not yet for the Java client). This feature allows you to avoid reading a tablet if you know it does not serve the row keys you are querying.
Gerrit 2641 Kudu now uses
earliest-deadline-first
RPC scheduling and rejection. This changes the behavior
of the RPC service queue to prevent unfairness when processing a backlog of RPC
threads and to increase the likelihood that an RPC will be processed before it
can time out.
KUDU-1337 Tablets from tables that were deleted might be unnecessarily re-bootstrapped when the leader gets the notification to delete itself after the replicas do.
KUDU-969 If a tablet server shuts down while compacting a rowset and receiving updates for it, it might immediately crash upon restart while bootstrapping that rowset’s tablet.
KUDU-1354 Due to a bug in Kudu’s MVCC implementation where row locks were released before the MVCC commit happened, flushed data would include out-of-order transactions, triggering a crash on the next compaction.
KUDU-1322 The C++ client now retries write operations if the tablet it is trying to reach has already been deleted.
Gerrit 2571 Due to a bug in the
Java client, users were unable to close the kudu-spark
shell because of
lingering non-daemon threads.
Gerrit 2239 The concept of "feature flags" was introduced in order to manage compatibility between different Kudu versions. One case where this is helpful is if a newer client attempts to use a feature unsupported by the currently-running tablet server. Rather than receiving a cryptic error, the user gets an error message that is easier to interpret. This is an internal change for Kudu system developers and requires no action by users of the clients or API.
Kudu 0.7.1 is a bug fix release for 0.7.0.
KUDU-1325 fixes a tablet server crash that could occur during table deletion. In some cases, while a table was being deleted, other replicas would attempt to re-replicate tablets to servers that had already processed the deletion. This could trigger a race condition that caused a crash.
KUDU-1341 fixes a potential data corruption and crash that could happen shortly after tablet server restarts in workloads that repeatedly delete and re-insert rows with the same primary key. In most cases, this corruption affected only a single replica and could be repaired by re-replicating from another.
KUDU-1343 fixes a bug in the Java client that occurs when a scanner has to scan multiple batches from one tablet and then start scanning from another. In particular, this would affect any scans using the Java client that read large numbers of rows from multi-tablet tables.
KUDU-1345 fixes a bug where in some cases the hybrid clock could jump backwards, resulting in a crash followed by an inability to restart the affected tablet server.
KUDU-1360 fixes a bug in the kudu-spark module
which prevented reading rows with NULL
values.
Kudu 0.7.0 is the first release done as part of the Apache Incubator and includes a number of changes, new features, improvements, and fixes.
The upgrade instructions can be found at Upgrade from 0.6.0 to 0.7.0.
The C++ client includes a new API, KuduScanBatch
, which performs better when a
large number of small rows are returned in a batch. The old API of vector<KuduRowResult>
is deprecated.
This change is API-compatible but not ABI-compatible. |
The default replication factor has been changed from 1 to 3. Existing tables will
continue to use the replication factor they were created with. Applications that create
tables may not work properly if they assume a replication factor of 1 and fewer than
3 replicas are available. To use the previous default replication factor, start the
master with the configuration flag --default_num_replicas=1
.
The Python client has been completely rewritten, with a focus on improving code quality and testing. The read path (scanners) has been improved by adding many of the features already supported by the C++ and Java clients. The Python client is no longer considered experimental.
With the goal of Spark integration in mind, a new kuduRDD
API has been added,
which wraps newAPIHadoopRDD
and includes a default source for Spark SQL.
The Java client includes new methods countPendingErrors()
and
getPendingErrors()
on KuduSession
. These methods allow you to count and
retrieve outstanding row errors when configuring sessions with AUTO_FLUSH_BACKGROUND
.
New server-level metrics allow you to monitor CPU usage and context switching.
Kudu now builds on RHEL 7, CentOS 7, and SLES 12. Extra instructions are included for SLES 12.
The file block manager’s performance was improved, but it is still not recommended for real-world use.
The master now attempts to spread tablets more evenly across the cluster during table creation. This has no impact on existing tables, but will improve the speed at which under-replicated tablets are re-replicated after a tablet server failure.
All licensing documents have been modified to adhere to ASF guidelines.
Kudu now requires an out-of-tree build directory. Review the build instructions for additional information.
The C` client library is now explicitly built against the
link:https://gcc.gnu.org/onlinedocs/libstdc/manual/using_dual_abi.html[old gcc5 ABI].
If you use gcc5 to build a Kudu application, your application must use the old ABI
as well. This is typically achieved by defining the `_GLIBCXX_USE_CXX11_ABI
macro
at compile-time when building your application. For more information, see the
previous link and link:http://developerblog.redhat.com/2015/02/05/gcc5-and-the-c11-abi/.
The Python client is no longer considered experimental.
The 0.6.0 release contains incremental improvements and bug fixes. The most notable changes are:
The Java client’s CreateTableBuilder and AlterTableBuilder classes have been renamed
to CreateTableOptions and AlterTableOptions. Their methods now also return this
objects,
allowing them to be used as builders.
The Java client’s AbstractKuduScannerBuilder#maxNumBytes() setter is now called batchSizeBytes as is the corresponding property in AsyncKuduScanner. This makes it consistent with the C++ client.
The "kudu-admin" tool can now list and delete tables via its new subcommands "list_tables" and "delete_table <table_name>".
OSX is now supported for single-host development. Please consult its specific installation instructions in OS X.
Kudu 0.5.0 was the first public release. As such, no improvements or changes were noted in its release notes.