* Mixed cluster tests with string NodeInfo version
- Move version based feature comparison to a common, deprecated method (to be replaced with real features)
- Use string comparison against old cluster version to partition new/old cluster nodes
Just found that we have a lot of inconsistency and needless verbosity
here in tests. We can just use `assertAcked` in a couple spots
to save `.get`, `.actionGet` etc., especially with the signature
change I added here.
Similar to the TransportVersions holder class, IndexVersions is the new
place to contain all constants for IndexVersion. This commit moves all
existing constants to the new class. It is purely mechanical.
- This hasnt been maintained in a while and the vagrant gradle plugin also broke compatibiliy for gradle
--configuration-cache. Also this removes a lot maintenance burden.
- Rework DistroTestPlugin and simplify task dependencies
- Rename :qa:os to :qa:packaging
- Update testing doc recommending buildkite tools for debugging packaging tests
When the test was trying to test recovering translog ops,
since we flush on close/shutdown, it failed because it never
recovered any translog ops.
The code for translog recovery is irrelevant due to that and
this PR proposes to remove it.
Alternatively, we could simulate killing nodes forcibly before
upgrading, but (a) that seems out of the ordinary for upgrades,
and (b) in trying that, it did not consistently pass the test
because sometimes the flush on close still happened.
Fixes#52031
Got asked why we have these classes again last night ... figured I'd
remove some of them to get us going here.
None of these classes are necessary, we can just inline all of these
away and make `ActionType` itself final or a record now that the action
type consists only in name and reader. See #97721 that made these things
redundant I think.
* Don't print synthetic source in mapping for bwc tests
* Move comment.
* Don't print synthetic source in mapping for bwc tests #2
* Don't print synthetic source in mapping for bwc tests #2
* Revert "Don't print synthetic source in mapping for bwc tests #2"
This reverts commit 034262c5d2.
* Revert "Don't print synthetic source in mapping for bwc tests #2"
This reverts commit 44e815635e.
* Revert "Don't print synthetic source in mapping for bwc tests (#100572)"
This reverts commit 9322ab9b91.
* Exclude synthetic source test from mixedClusterTests
* Update comment.
* Mute all tsdb tests in mixedClusterTests
This is an interim step to stop sporadic test failures, while we try to
fix version skip for mixed cluster tests.
* Remove old exclusion
* Add aggregations too
* Mute tests for versions between 8.7-8.10
* Remove mute
This commit moves back to using a `"major.minor.patch"` string for the
version field in snapshots stored in `RepositoryData`, using the marker
string `"8.11.0"` to allow older versions to filter out newer snapshots
and adding a new `index_version` field alongside.
This format is fully backwardly-compatible, except that it trips an
assertion in the versions of 8.10.x released today. When running without
assertions enabled, things work correctly in all versions.
Relates #98454
- Reinstates the few working parts of MultiVersionRepositoryAccessIT
- Fixes the version bound for compatibility with v8.9.x
- Removes an assertion that blocks fixing BwC properly in 8.11.0
Relates #98454
This is one of the few uses of Version#minimumCompatibilityVersion that isn't covered by other planned work. Since the minimum compatibility version for 7.14.0 is always going to be 6.8.0, we can just hardcode the value.
The StatusToXContent interface extends ToXContent with a REST status
method. It is used by RestStatusToXContentListener, which uses the extra
method to set the status of the RestResponse. However, given the
generics parametrization of RestToXContentListener, we can already get
the status of a response object without having these extra interfaces,
which mix serialization and status concerns.
This commit removes StatusToXContent and RestStatusToXContentListener,
instead adding some extra optional parameters to RestToXContentListener to
handle retrieving status and location information from the passed in Response
object.
To help the user know what the possible cluster states are and to
provide an accurate accounting, we added counters summarising
`running`, `partial` and `failed` clusters to the `_clusters` section.
Changes:
- Now in the response is present the number of `running` clusters.
- We split up `partial` and `successful` (before was summed up in the
`successful` counter).
- We now have a counter for `failed` clusters.
- Now `total` is always equal to `running` + `skipped` + `failed` +
`partial` + `successful`.
* Replaced some occurrences of Version.CURRENT toString
* Added unit tests for the 2 different warn header patterns in HeaderWarnings
* Comment to clarify non-semantic versions + revert change on Version.major message
Adds @SuppressWarnings("this-escape") to all necessary places to that
Elasticsearch can compile with -Werror on JDK21
No investigation has been done to determine whether any of the cases
are a potential source of errors - we have simply suppressed all
existing occurrences.
Resolves: #99845
Re-applies the changes from #99572 to move some bwc tests to a junit-based build infrastructure. Some tests that did not handle the move well have been kept in rolling-upgrade-legacy using the old gradle-based infrastructure
Two test suites did not react well to the junit-based bwc infrastructure, so those have been separated into a legacy module using the old gradle-based system until they can be looked at properly.
This unblocks the 8.11 release.
This commit renames the tracing to telemetry.tracing in both xpack/APM and elasticserach's org.elasticsearch.tracing.Tracer (the api)
the xpack/APM is renamed as follows:
org.elasticsearch.telemetry.apm - the only exported package
org.elasticsearch.telemetry.apm.settings - APMSettings
org.elasticsearch.telemetry.apm.tracing - APMTracer
org.elasticsearch.tracing.Tracer is moved to org.elasticsearch.telemetry.tracing.Tracer (responsible for majority of the changes in this PR)
The index version is needed to check the output against an upgraded cluster. Now the junit tests share a JVM between the old & upgraded clusters, we can store the old index version in the class to be read when needed
* Remove setting the number of replicas in aggregation yaml tests.
* Stop waiting for yellow/green. The create index request by default return s when at least the primary shard has been assigned.
* Adjust assertions that are sensitive to the number of shard copies.
* Muted 4 aggregation yaml tests in mixed cluster qa ([link](https://github.com/elastic/elasticsearch/pull/99413/files#diff-033fa895a9866248015cbe7871deab0b3c88fa66a5e1bc9e789079d0d1a39f24R31)). These tests assert cache hits and with replicas we currently can't reliable assert this in a multi node test cluster. Like the mentioned comment suggest if the node selector can reliable select the same node, then we can unmute. I think muting is ok for now. These yaml tests are run in aggregation module and will soon be run in serverless. I think enforcing that many yaml tests run tests against indices with no replicas is worse for test coverage than muting 4 yaml tests in this qa module. Additionally I think that testing that the request cache work should not be tested in full blown integration tests, but more targeted unit/semi integration tests.
Remove temporary TransportService#registerRequestHandler functions
that take String instead of Executor and update the callers. This
is part of a larger change to pass Executor into request handlers,
rather than creating and stashing Strings that are only used to
eventually fetch appropriate Executors anyway.
Extensive testing changes were required to make TransportService and
ThreadPool mocks capable of looking up Executors from Strings for
TransportAction constructors. Added a new MockUtils class to
modularize the additional test setup.
Part of the changes for #97879
Remove temporary TransportService#registerRequestHandler functions
that take String instead of Executor and update the callers. This
is part of a larger change to pass Executor into request handlers,
rather than creating and stashing Strings that are only used to
eventually fetch appropriate Executors anyway.
Extensive testing changes were required to make TransportService and
ThreadPool mocks capable of looking up Executors from Strings for
TransportAction constructors. Added a new MockUtils class to
modularize the additional test setup.
Part of the changes for #97879
A composable index template with no template defined in the body is mistakingly always assumed to not be a time series template. Even if it refers to a component template that has the index.mode setting set to time_series and the component template defines mappings with dimension fields or routing paths.
Closes#98834
This commit tracks progress for each shard search by cluster alias
using a new SearchProgressListener (CCSSingleCoordinatorSearchProgressListener).
Both sync and async CCS searches use this new progress listener when
minimize_roundtrips=false.
Two of the SearchProgressListener method had to be extended to allow tracking
per-cluster took values (TransportSearchAction.SearchTimeProvider) and
whether searches timed out (by passing in QuerySearchResult to the onQueryResult
listener method).
This commit brings parity between minimize_roundtrips=true and false to have
the same _cluster/details sections in CCS search responses.
Note that there are still a few differences between minimize_roundtrips=true and false.
1. The per-cluster took value for minimize_roundtrips=true is accurate, but the
for 'false' it is only measured at the granualarity of each partial reduce,
so the per cluster took time is overestimated in basically all cases.
2. For minimize_roundtrips=true, a skip_unavailable=false cluster that disconnects
during the search or has all searches on all shards fail, will cause the entire
search to fail. This is (still) not true for minimize_roundtrips=false. The search
is only failed if the skip_unavailable=false cluster cannot be connected to at the
start of the search. (This will likely be changed in a follow up ticket that implements
fail-fast logic for in-progress searches that should fail due to a skip_unavailable=true
cluster failing.)
3. The shard accounting for minimize_roundtrips=false is always accurate (total shard counts
are known at the start of the search). For minimize_roundtrips=true, the shard accounting
is only accurate per cluster unless all clusters have successful (or partially successful)
searches. For clusters that have failures we do not have shard count info.
This bumps to an IndexVersion that is not associated with any specific release version. From this point, index metadata/data versioning will be handled in the same way as TransportVersion - a new constant for every change
By default in 8.11, vectors will be indexed by default. However, various
tests that relied on the previous behavior were not updated.
This PR updates those tests.
Related: https://github.com/elastic/elasticsearch/pull/98268
Drying up the logic for looking up the customs and falling back to the empty default. Hopefully, this improves readability somewhat by taking away that complexity and some null checks.
I like it at least, saves a lot of clutter.
Re-revert of #98896, only fixing a single null check in tests.
Lorenzo Dematté
Armin Braun
Ryan Ernst
Rene Groeschke
Iraklis Psaroudakis
Kostas Krikellas
Yang Wang
David Turner
Keith Massey
Mark Vieira
William Brafford
Alan Woodward
Matteo Piergiovanni
Tim Vernum
Simon Cooper
ning pan
Przemyslaw Gomulka
Jake Landis
Martijn van Groningen
Dianna Hohensee
Brian Seeders
Ed Savage
Michael Peterson
Benjamin Trent