This commit adds special handling for the `java.io.IOException: An established connection was aborted by the software in your host machine` in `Netty4HttpClient#exceptionCaught ` method. This exception only occurs when running tests on Windows hosts
Resolves: #108193
(cherry picked from commit 93ec9d6142)
`ExpandSearchPhase` was leaking `SearchHits` when a pooled `SearchHits`
that was read from the wire was added to an unpooled `SearchHit`.
This commit makes the relevant `SearchHit` instances that need to be
pooled so they released nested hits, pooled. This requires a couple of
smaller adjustments in the codebase, mainly around error handling.
Offsets in memory segments should be computed as longs to avoid integer overflow on large segments.
---------
Co-authored-by: ChrisHegarty <chegar999@gmail.com>
Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com>
* Handle must_not clauses when disabling the weight matches highlighting mode (#108453)
This change makes sure we check all queries, even the must_not ones, to decide if we should disable weight matches highlighting or not.
Closes#101667Closes#106693
* adapt test skip version
The word `cannot` implies Elasticsearch prevents you from doing these
things, but it doesn't have this protection today (see #73947). This
commit clarifies this by saying `must not` instead.
Closes#108450
This PR logs tasks that are running after the disruption is cleared,
allowing us to investigate why the disruption tests failed in #107347.
Relates #107347
The rate aggregation will only be used in time-series aggregations,
where ordinal grouping should never be applied. Therefore, we should
remove these tests to reduce unnecessary noise.
Closes#107568Closes#107710
Today, we do not wait for remote sinks to stop before completing the
main request. While this doesn't affect correctness, it's important that
we do not spawn child requests after the parent request is completed.
Closes#105859
One user reached out mentioning that it would be a good idea to remind
users to re-upload the license after full cluster recovery from snapshot
as one can easily miss this when trying to figure out why some features
aren't working after the restore.
(cherry picked from commit 295fba33d8)
Co-authored-by: Rafi Estrada <62263912+TheRiffRafi@users.noreply.github.com>
Linux systems with multiarch (e.g. i386 & x86_64) libraries
may have libsystemd.0 in two subdirectories of an entry in
java.library.path. For example, libsystemd.so.0 may be found
in both /usr/lib/i386-linux-gnu and /usr/lib/x86_64-linux-gnu.
Instead of attempting to load any library found, attempt all
and stop as soon as one is successfully loaded.
Fixes an error is when the normalisation step produces text that is longer
than the input text and an offset needs to be added to map back to the
original.
* [DOCS] Add local dev setup instructions
- Replace existing Run ES in Docker locally page, with simpler no-security local dev setup
- Move this file into Quickstart folder, along with existing quickstart guide
- Update self-managed instructions in Quickstart guide to use local dev approach
This adds some clarifications on the time unit strings the function
takes as arguments, noting the differences between these and the time
span literals, as well as the abbreviations' source.
(cherry picked from commit de725aef80)
ZSTD compression was added behind feature flag in #103374.
But the feature flag isn't removed yet and therefor the zstd compression will not be enabled, deflate and lz4 are still active.
The DirectEncoder currently returns the incorrect value for the
positionCount() method, which should be the number of positions ready in
the current batch. We need to keep track of whether a position is loaded
via encodeNextBatch() and consumed via the read() method. However, we
can always return 1 for positionCount(), indicating that one position is
already loaded. Our tests failed to catch this because mv_ordering
wasn't enabled when generating test blocks, effectively disabling the
DirectEncoders.
Closes#108268
This reworks the integration-test-only csv testing for `metadata` to use
the `required_feature:` syntax instead of the `-IT_tests_only`
extension. This is a little more flexible and way nicer on the eyes.
Previously we were logging all ESQL queries. That's a lot! Plus maybe
there's PII in there or something. Let's not do that unless you ask for
it. This changes the query logging to the `debug` log level you can
still get at these if you want them, but you don't have them by default.
you have to turn it on.
This takes the CIDR_MATCH out of the operators group and adds it to a
new `IP functions` group.
The change also re-aranges the groups, grouping together the
type-specific functions and ordering them alphabetically.
(cherry picked from commit b26d7d3e14)
This fixes the generation of the signatures for variadic functions,
except for those that take a list as last argument; i.e. functions with
optional arguments (like ROUND) or functions with overloading-like
signatures (like BUCKET).
Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com>
This PR moves the doPrivileged wrapper closer to the actual deletion
request to ensure the necesary security context is established at all
times. Also added a new repository setting to configure max size for s3
deleteObjects request.
Fixes: #108049
Currently, loading ordinals multiple times (after advanceExact) for
documents with values spread across multiple blocks in the TSDB codec
will fail due to the absence of re-seeking for the ordinals block.
Doc-values of a document can spread across multiple blocks in two cases:
when it has more than 128 values or when it exceeds the remaining space
in the current block.
