This PR introduces two new fields in to `RepositoryData` (index-N) to track the blob name of `IndexMetaData` blobs and their content via setting generations and uuids. This is used to deduplicate the `IndexMetaData` blobs (`meta-{uuid}.dat` in the indices folders under `/indices` so that new metadata for an index is only written to the repository during a snapshot if that same metadata can't be found in another snapshot.
This saves one write per index in the common case of unchanged metadata thus saving cost and making snapshot finalization drastically faster if many indices are being snapshotted at the same time.
The implementation is mostly analogous to that for shard generations in #46250 and piggy backs on the BwC mechanism introduced in that PR (which means this PR needs adjustments if it doesn't go into `7.6`).
Relates to #45736 as it improves the efficiency of snapshotting unchanged indices
Relates to #49800 as it has the potential of loading the index metadata for multiple snapshots of the same index concurrently much more efficient speeding up future concurrent snapshot delete
Currently we combine coordinating and primary bytes into a single bucket
for indexing pressure stats. This makes sense for rejection logic.
However, for metrics it would be useful to separate them.
We have recently added internal metrics to monitor the amount of
indexing occurring on a node. These metrics introduce back pressure to
indexing when memory utilization is too high. This commit exposes these
stats through the node stats API.
Backport of #59293 to 7.x branch.
* Create new data-stream xpack module.
* Move TimestampFieldMapper to the new module,
this results in storing a composable index template
with data stream definition only to work with default
distribution. This way data streams can only be used
with default distribution, since a data stream can
currently only be created if a matching composable index
template exists with a data stream definition.
* Renamed `_timestamp` meta field mapper
to `_data_stream_timestamp` meta field mapper.
* Add logic to put composable index template api
to fail if `_data_stream_timestamp` meta field mapper
isn't registered. So that a more understandable
error is returned when attempting to store a template
with data stream definition via the oss distribution.
In a follow up the data stream transport and
rest actions can be moved to the xpack data-stream module.
In 7.x, we have java 8 as minimum jdk version. In the past, for
packaging tests, we relied on the system to provide an alternative jdk
to be used by the no-jdk distributions. Master switched to using a build
provided jdk, but 7.x was stuck relying on the system because it needed
a java 8 jdk. The jdk download plugin was updated a while ago to support
jdk 8, and so this PR converts the distro tests to use the build
provided jdk just as master branch does.
Note also this fixes a failure that would sometimes occur on older jdks
in windows where the expected gc filename can be different.
These tests sometimes install a template so they can be compatible with older versions, but they run
amok of the occasionally installed "global" template which changes the default number of shards.
This commit adds `allowedWarnings` and allows these warnings to be present, but doesn't fail if they
are not (since the global template is only randomly installed).
Resolves#58807Resolves#58258
Backport of #59076 to 7.x branch.
The commit makes the following changes:
* The timestamp field of a data stream definition in a composable
index template can only be set to '@timestamp'.
* Removed custom data stream timestamp field validation and reuse the validation from `TimestampFieldMapper` and
instead only check that the _timestamp field mapping has been defined on a backing index of a data stream.
* Moved code that injects _timestamp meta field mapping from `MetadataCreateIndexService#applyCreateIndexRequestWithV2Template58956(...)` method
to `MetadataIndexTemplateService#collectMappings(...)` method.
* Fixed a bug (#58956) that cases timestamp field validation to be performed
for each template and instead of the final mappings that is created.
* only apply _timestamp meta field if index is created as part of a data stream or data stream rollover,
this fixes a docs test, where a regular index creation matches (logs-*) with a template with a data stream definition.
Relates to #58642
Relates to #53100Closes#58956Closes#58583
If the recovery source is on an old node (before 7.2), then the recovery
target won't have the safe commit after phase1 because the recovery
source does not send the global checkpoint in the clean_files step. And
if the recovery fails and retries, then the recovery stage won't
transition properly. If a sync_id is used in peer recovery, then the
clean_files step won't be executed to move the stage to TRANSLOG.
Relates ##7187
Closes#57708
This commit creates a new Gradle plugin to provide a separate task name
and source set for running YAML based REST tests. The only project
converted to use the new plugin in this PR is distribution/archives/integ-test-zip.
For which the testing has been moved to :rest-api-spec since it makes the most
sense and it avoids a small but awkward change to the distribution plugin.
The remaining cases in modules, plugins, and x-pack will be handled in followups.
This plugin is distinctly different from the plugin introduced in #55896 since
the YAML REST tests are intended to be black box tests over HTTP. As such they
should not (by default) have access to the classpath for that which they are testing.
The YAML based REST tests will be moved to separate source sets (yamlRestTest).
The which source is the target for the test resources is dependent on if this
new plugin is applied. If it is not applied, it will default to the test source
set.
Further, this introduces a breaking change for plugin developers that
use the YAML testing framework. They will now need to either use the new source set
and matching task, or configure the rest resources to use the old "test" source set that
matches the old integTest task. (The former should be preferred).
As part of this change (which is also breaking for plugin developers) the
rest resources plugin has been removed from the build plugin and now requires
either explicit application or application via the new YAML REST test plugin.
Plugin developers should be able to fix the breaking changes to the YAML tests
by adding apply plugin: 'elasticsearch.yaml-rest-test' and moving the YAML tests
under a yamlRestTest folder (instead of test)
* Replace compile configuration usage with api (#58451)
- Use java-library instead of plugin to allow api configuration usage
- Remove explicit references to runtime configurations in dependency declarations
- Make test runtime classpath input for testing convention
- required as java library will by default not have build jar file
- jar file is now explicit input of the task and gradle will ensure its properly build
* Fix compile usages in 7.x branch
When a packaging test completes, we move the elasticsearch logfile to
another location so that if the next test fails, we have just the
contents of the test that ran. However, once the test fails, we continue
trying to move this file in the subsequent calls to teardown when the
rest of the tests are skipped. This commit skips the teardown step once
we have marked the suite as failed.
This commit converts the bats tests for the plugin cli into the java
packaging test framework. The new tests only use the example plugin to
test the plugin cli. The tests for each individual plugin's contents
after being installed are handled by a new unit test for the plugin
installer added in #58287.
Today we have individual settings for configuring node roles such as
node.data and node.master. Additionally, roles are pluggable and we have
used this to introduce roles such as node.ml and node.voting_only. As
the number of roles is growing, managing these becomes harder for the
user. For example, to create a master-only node, today a user has to
configure:
- node.data: false
- node.ingest: false
- node.remote_cluster_client: false
- node.ml: false
at a minimum if they are relying on defaults, but also add:
- node.master: true
- node.transform: false
- node.voting_only: false
If they want to be explicit. This is also challenging in cases where a
user wants to have configure a coordinating-only node which requires
disabling all roles, a list which we are adding to, requiring the user
to keep checking whether a node has acquired any of these roles.
This commit addresses this by adding a list setting node.roles for which
a user has explicit control over the list of roles that a node has. If
the setting is configured, the node has exactly the roles in the list,
and not any additional roles. This means to configure a master-only
node, the setting is merely 'node.roles: [master]', and to configure a
coordinating-only node, the setting is merely: 'node.roles: []'.
With this change we deprecate the existing 'node.*' settings such as
'node.data'.
* Enable TTY password OS tests, plus refactoring (#57759)
Two keystore tests were unintentionally ignored when the
password-protected keystore work was merged. I've reënabled those tests
here.
I've also refactored the test methods a little bit to reduce the API
surface: instead of having a "startElasticsearchTtyPassword" method and
a "startElasticsearchStandardInputPassword" method, I've made a single
"startElasticsearch" method with a "useTty" boolean argument.
* Separate daemonization and non-daemonization case for tty
Centos 6 uses a version of expect that kills the elasticsearch process
when it tries to daemonize. I will fix this in future work but for now
I'm replacing it with a todo.
This commit creates a shared withCustomConfig method that may be used by
any packaging test. The method will copy the config directory and
override the conf path appropriately depending on the distribution type.
The packaging tests currently have a couple different ways of deciding
where temp files should be placed, and then sometimes used fixed file
or directory names within that dir. This commit conslidates some of that
temp dir handling by making it more compatible with the handling that
exists within the bats tests, where /tmp is not always appropriate due
to how systemd interacts with it. This commit also adds a utility
methhod for creating temp dirs, so as to ensure the new directory is
created as if a umask of 022 were used, which is not the case when using
Files.createTempDirectory without a set of permissions (it assumes 077).
This change allows to use an `index_filter` in the
field capabilities API. Indices are filtered from
the response if the provided query rewrites to `match_none`
on every shard:
````
GET metrics-*
{
"index_filter": {
"bool": {
"must": [
"range": {
"@timestamp": {
"gt": "2019"
}
}
}
}
}
````
The filtering is done on a best-effort basis, it uses the can match phase
to rewrite queries to `match_none` instead of fully executing the request.
The first shard that can match the filter is used to create the field
capabilities response for the entire index.
Closes#56195
Backport of #50920. Part of #48366. Implement an API for listing,
importing and deleting dangling indices.
Co-authored-by: David Turner <david.turner@elastic.co>
This template was added for 7.0 for what I am guessing is a BWC issue related to deprecation
warnings. It unfortunately seems to cause failures because templates for these tests are not cleared
after the test (because these are upgrade tests).
Resolves#56363
* Remove usage of deprecated testCompile configuration
* Replace testCompile usage by testImplementation
* Make testImplementation non transitive by default (as we did for testCompile)
* Update CONTRIBUTING about using testImplementation for test dependencies
* Fail on testCompile configuration usage
This commit re-enables windows testing for archives packaging tests.
These were disabled previously because of constant failure due to
windows file locks, but the failure does not occur outside of CI, so
they are being re-enabled to further investigate the failure.
relates #50825
We can't use the high level create snapshot request any longer
since we changed some of its default parameters in `8` and those
are not understood by older versions like `7.4`.
Closes#57650
Almost every outbound message is serialized to buffers of 16k pagesize.
We were serializing these messages off the IO loop (and retaining the concrete message
instance as well) and would then enqueue it on the IO loop to be dealt with as soon as the
channel is ready.
1. This would cause buffers to be held onto for longer than necessary, causing less reuse on average.
2. If a channel was slow for some reason, not only would concrete message instances queue up for it, but also 16k of buffers would be reserved for each message until it would be written+flushed physically.
With this change, the serialization happens on the event loop which effectively limits the number of buffers that `N` IO-threads will ever use so long as messages are small and channels writable.
Also, this change dereferences the reference to the concrete outbound message as soon as it has been serialized to save some more on GC.
This reduces the GC time for a default PMC run by about 50% in experiments (3 nodes, 2G heap each, loopback ... obvious caveat is that GC isn't that heavy in the first place with recent changes but still a measurable gain).
I also expect it to be helpful for master node stability by causing less of a spike if master is e.g. hit by a large number of requests that are processed batched (e.g. shard snapshot status updates) and responded to in a short time frame all at once.
Obviously, the downside to this change is that it introduces more latency on the IO loop for the serialization. But since we read all of these messages on the IO loop as well I don't see it as much of a qualitative change really and the more predictable buffer use seems much more valuable relatively.
As the datastream information is stored in the `ClusterState.Metadata` we exposed
the `Metadata` to the `AsyncWaitStep#evaluateCondition` method in order for
the steps to be able to identify when a managed index is part of a DataStream.
If a managed index is part of a DataStream the rollover target is the DataStream
name and the highest generation index is the write index (ie. the rolled index).
(cherry picked from commit 6b410dfb78f3676fce1b7401f1628c1ca6fbd45a)
Signed-off-by: Andrei Dan <andrei.dan@elastic.co>
The new translog bwc test checks a corruption case before 6.3.0.
However, it needs to restart the old node to reproduce, which does not
currently work given how testclusters works when plugins are installed.
As a workaround, this commit omits creating bwc tests before 6.3.0 only
when the default distribution is used.
fixes#57252
If an upgraded node is restarted multiple times without flushing a new
index commit, then we will wrongly exclude all commits from the starting
commits. This bug is reproducible with these minimal steps: (1) create
an empty index on 6.1.4 with translog retention disabled, (2) upgrade
the cluster to 7.7.0, (3) restart the upgraded the cluster. The problem
is that with the new translog policy can trim translog without having a
new index commit, while the existing commit still refers to the previous
translog generation.
Closes#57091
The packaging tests start elasticsearch in various ways. All of these
currently expect it is started asynchronously, yet some tests expect it
will fail to start and want to check the error output. This commit adds
a daemonize flag to the utility methods to start elasticsearch for such
cases, so that when the start method returns, all the error output
should already be available since the process will have exited.
relates #51716
Added support for decompression at LLRC and added integration test
(cherry picked from commit 2621452473e0c236aa28db749f782a24eca6c974)
Signed-off-by: Andrei Dan <andrei.dan@elastic.co>
Co-authored-by: Hakky54 <hakangoudberg@hotmail.com>
In normal operation native controllers are not expected to write
anything to stdout or stderr. However, if due to an error or
something unexpected with the environment a native controller
does write something to stdout or stderr then it will block if
nothing is reading that output.
This change makes the stdout and stderr of native controllers
reuse the same stdout and stderr as the Elasticsearch JVM (which
are by default redirected to es.stdout.log and es.stderr.log) so
that if something unexpected is written to native controller
output then:
1. The native controller process does not block, waiting for
something to read the output
2. We can see what the output was, making it easier to debug
obscure environmental problems
Backport of #56491
Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com>
Martijn van Groningen
Armin Braun
Tim Brooks
Przemko Robakowski
Ryan Ernst
Lee Hinman
Nhat Nguyen
Jake Landis
Rene Groeschke
Alan Woodward
Nik Everett
Jason Tedor
William Brafford
Jim Ferenczi
Stuart Tettemer
Rory Hunter
Dan Hermann
Mark Vieira
Andrei Dan
Julie Tibshirani
Gordon Brown
David Roberts
Ignacio Vera