This commit propagates the preference and routing of the original SearchRequest in the ShardSearchRequest.
This information is then use to fix a bug in sliced scrolls when executed with a preference (or a routing).
Instead of computing the slice query from the total number of shards in the index, this commit computes this number from the number of shards per index that participates in the request.
Fixes#27550
In order to have less qa/ projects, this commit removes the
watcher-smoke-test-mustache and the watcher-smoke-test-painless projects
and moves the tests of both into the watcher-smoke-test projects.
This results in less projects to parse when calling gradle and a shorter test
time due to being able to run everything within one elasticsearch
instance.
Today when forking setup commands we do not set JAVA_HOME. This means
that we might not use a version of Java compatible with the version of
Java the command is expecting to run on (for example, 5.6 nodes would
expect JDK 8, and this is true even for their setup commands). This
commit sets JAVA_HOME when configuring setup command tasks.
The low-level REST client targets JDK 7. To avoid compiling against JDK
functionality not available in JDK 7, we use animal sniffer. However,
when we switched to using the JDK 9 and now the JDK 10 compiler which
has built-in support for targeting previous JDKs, we no longer need to
use animal sniffer. This is because the JDK is now packaged with the
signatures needed to ensure that when we target JDK 7 at compile-time it
is detected that we are only using JDK 7 functionality. This commit
removes the use of animal sniffer from the low-level REST client build.
Adds the `.vscode` directory to `.gitignore`. This didn't come up when
we had x-pack in a different repository than Elasticsearch because I'd
open the directory that contained both of them in VSCode. Now that they
are merged I tend to just open the Elasticsearch repository which causes
VSCode to create this directory that we don't need to commit.
Also removes leftover maven files from `.gitignore`.
Today we always add no-ops to translog regardless of its origin, thus a
noop may appear in the translog multiple times. This is not a big deal
as noops are small and rare to appear.
This commit ensures to add a noop to translog only if its origin is not
from local translog. This restriction has been applied for index and
delete.
Adds tasks that check that the all jars that we build have LICENSE.txt
and NOTICE.txt files and that the files are correct. Sets check to
depend on these task.
This is mostly there for extra parnoia because we automatically
configure all Jar tasks to include the LICENSE.txt and NOTICE.txt
files anyway. But it is quite possible to add configuration to those
tasks that would override either file.
This causes check to depend on several more things than it used to.
Take, for example, javadoc:
check depends on the new verifyJavadocJarNotice which depends on
extractJavadocJar which depends on javadocJar which depends on
javadoc, this check now depends on javadoc.
The packaging tests for Debian based distro is loooking
for docs in /usr/share/elasticsearch, but it should be
/usr/share/elasticsearch-oss for the oss package.
The bundled configuration isn't recognised by eclipse so these
dependencies are missed when it imports the `x-pack:plugin:sql:jdbc`
project. This change makes these dependencies compile dependencies if
the build is running for Eclipse.
This metric previously existed for backwards compatibility reasons
although the suggest stats were folded into search stats. This metric
was deprecated in 6.3.0 and this commit removes them for 7.0.0.
This commit fixes two issues with the byte size value equals/hash code
test.
The first problem is due to a test failure when the original instance is
zero bytes and we pick the mutation branch where we preserve the size
but change the unit. The mutation should result in a different byte size
value but changing the unit on zero bytes still leaves us with zero
bytes.
During the course of fixing this test I discovered another problem. When
we need to randomize size, we could randomly select a size that would
lead to an overflow of Long.MAX_VALUE.
This commit fixes both of these issues.
This commit gives us Gradle 4.7 which brings with it official JDK 10
support (which we are already using as our minimum compiler version) and
support for JDK 11 early-access releases (previous versions of Gradle
will not even start on JDK 11).
Tests need to wait for changes to the job's established memory usage to
propagate and an over enthusiastic optimisation meant jobs were updated
from stale state causing recent change to be lost.
This commit adds some build time checks that the archive distributions
and package distributions contain the appropriate license and notice
files, and the package distributions contain the appropriate license
metadata.
This commit uses the customFields setting of the Deb task in ospackage
to work around the fact it does not know anything about the License
attribute natively.
THe deb distribution has a special copyright file instead of
LICENSE.txt, but the distributions were including the template file
instead of the rendered file (which includes the license name and text).
This commit fixes the classpath for the SQL CLI tool on Windows. As the
x-pack bin folder was collapsed into the distribution bin folder, the
location of the classpath here needed to no longer contain the old
plugins directory.
This commit adds the distribution type to the startup scripts so that we
can discern from log output and the main response the type of the
distribution (deb/rpm/tar/zip).
With the move of X-Pack to a module, the classpath for the scripts needs
to be adjusted. This was done on Unix, but not for Windows. This commit
addresses Windows.
This commit moves the apache and elastic license files into a new
root level `licenses` directory and rewrites the top level LICENSE.txt
to clarify the repository has a mix of apache and elastic licensed code.
This commit adds license metadata to rpm and deb packages. Additionally,
it makes the copyright file for deb files follow the machine readable
specification, and sets the correct license text based on the oss vs
default deb packages.
This commit adapts the license headers check to no longer look for the
ealsticsearch confidential license, but instead to look for the new
Elastic License header.
With the switch to X-Pack as a module, we lost production of POMs for
the JARs that we publish, and did not have a license/notice file in the
zip archives nor the exploded module. This commit ensures that we
generate these POMs, and license/notice files.
X-Pack can no longer be installed as a plugin. This commit adds special
handling for when a user attempts to install X-Pack. This special
handling informs the user of the oss distribution that they should
download the default distribution and the user of the default
distribution that X-Pack does not require installation as it is included
by default.
This commit adds the distribution flavor (default versus oss) to the
build process which is passed through the startup scripts to
Elasticsearch. This change will be used to customize the message on
attempting to install/remove x-pack based on the distribution flavor.
This commit makes x-pack a module and adds it to the default
distrubtion. It also creates distributions for zip, tar, deb and rpm
which contain only oss code.
This change updates the WatcherLifecycleService to have its own single
thread executor that is used for lifecycle changes in order to have a
guarantee for the order that the changes are executed in.
Previously, a runnable would be submitted to the generic threadpool for
each lifecycle change that is needed. There was no guarantee of
ordering for these changes and no checks to see if a state change was
already in flight.
Relates elastic/x-pack-elasticsearch#4429
Original commit: elastic/x-pack-elasticsearch@14b73381db
The PKI realm has never been a caching realm as the need had not
presented itself until now. The PKI realm relies on role mappings to
map the DN from a certificate to roles so that the users have the
appropriate access permissions. Without caching, this role mapping will
happen on every request. For file based role mappings, this is not an
issue as the mappings are based on equality checks for the DN.
However, the design of the API based role mappings allows for more
complex matches. These matches are implemented using automata, which
are built on every request that needs role mappings. Building automata
is an expensive operation and in combination with the PKI realm's lack
of caching leads to a significant performance impact.
The change in this commit makes the PkiRealm a caching realm using the
same pattern as other caching realms. The cache provided by
elasticsearch core is used to map the fingerprint of a certificate to
the user that was resolved from this certificate. The semantics of
modifications to this cache during iteration requires that we use a
read-write lock to protect access. There can be multiple concurrent
modifications and retrievals but iteration must be protected from any
attempts to modify the cache.
Additionally, some PKI tests were converted to single node tests as
part of this change. One test only used a single node and the other did
not require multiple nodes.
relates elastic/x-pack-elasticsearch#4406
Original commit: elastic/x-pack-elasticsearch@214772e1c1
Jim Ferenczi
Alexander Reelsen
Julie Tibshirani
Jason Tedor
Nik Everett
Nhat Nguyen
Tanguy Leroux
Jack Conradson
Colin Goodheart-Smithe
Adrien Grand
ZarHenry96
Michael Basnight
David Kyle
Ryan Ernst
Jay Modi