* `test.cluster.node.seed` was only used in one place where it wasn't adding value and was now replaced with a constant
* `tests.portsfile` is now an official setting and has been renamed to `node.portsfile`
this commit also convert `search.default_keep_alive` and `search.keep_alive_interval` to the new settings infrastrucutre
This is consistent with what happens in elasticsearch.sh/.bat, and it means
tests will work even if there is a crazy "system" JNA installed on the machine.
The rest test framework, because it used to be tightly integrated with
ESIntegTestCase, currently expects the addresses for the test cluster to
be passed using the transport protocol port. However, it only uses this
to then find the http address.
This change makes ESRestTestCase extend from ESTestCase instead of
ESIntegTestCase, and changes the sysprop used to tests.rest.cluster,
which now takes the http address.
closes#15459
Site plugins used to be used for things like kibana and marvel, but
there is no longer a need since kibana (and marvel as a kibana plugin)
uses node.js. This change removes site plugins, as well as the flag for
jvm plugins. Now all plugins are jvm plugins.
With this commit we do not check only if an endpoint is up but
we also check that the cluster status is green. Previously,
builds sporadically failed to pass this condition.
1. Uses forbidden patterns to prevent things from referencing
java.io.Serializable or from mentioning serialVersionUID.
2. Uses -Xlint:-serial so we don't have to hear from javac that we aren't
declaring serialVersionUID on any classes that we make that happen to extend
Serializable.
3. Remove Serializable and serialVersionUID declarations.
I didn't use forbidden apis because it doesn't look like it has a way to ban
explicitly implementing Serializable. If you try to ban Serializable with
forbidden apis you end up banning all Exceptions and all Strings.
Closes#15847
This commit removes and now forbids use of
org.apache.lucene.index.IndexWriter#isLocked as this method was
deprecated in LUCENE-6508. The deprecation is due to the fact that
checking if a lock is held before acquiring that lock is subject to a
time-of-check-to-time-of-use race condition. There were three uses of
IndexWriter#isLocked in the code base:
- a logging statement in o.e.i.e.InternalEngine where we are already in
an exceptional condition that the lock was held; in this case,
logging whether or not the directory is locked is superfluous
- in o.e.c.l.u.VersionsTests where we were verifying that a write lock
is released upon closing an IndexWriter; in this case, the check is
not needed as successfully closing an IndexWriter releases its
write lock
- in o.e.t.s.MockFSDirectoryService where we were verifying that a
directory is not write-locked before (implicitly) trying to obtain
such a write lock in org.apache.lucene.index.CheckIndex#<init> (this
is the exact type of a situation that is subject to a race
condition); in this case we can proceed by just (implicitly) trying
to obtain the write lock and failing if we encounter a
LockObtainFailedException
This commit removes and now forbids all uses of
java.util.concurrent.ThreadLocalRandom across the codebase. The
underlying issue with ThreadLocalRandom is that it can not be
seeded. This means that if ThreadLocalRandom is used in production code,
then tests that cover any code path containing ThreadLocalRandom will be
prevented from being reproducible by use of ThreadLocalRandom. Instead,
using org.elasticsearch.common.random.Randomness#get will give
reproducible sources of random when running under tests and otherwise
still give an instance of ThreadLocalRandom when running as production
code.
This was originally intended to be general purpose in #15555, but
that still had problems. Instead, this change fixes the issue explicitly
for slf4j-api, since that is the problematic dep that is not actually
included in the distributions.
This change adds a Fixture class for use by gradle. A Fixture is an
external process that integration tests will use. It can be added as a
dependsOn for integTest, and will automatically be shutdown upon success
or failure, as well as relevant information dumped on failure. There is
also an example fixture in this change.
This new task allows setting code, similar to a doLast or doFirst,
except it is specifically geared at running ant (and thus called doAnt).
It adjusts the ant logging while running the ant so that the log
level/behavior can be tweaked, and automatically buffers based on gradle
logging level, and dumps the ant output upon failure.
This fixes the `lenient` parameter to be `missingClasses`. I will remove this boolean and we can handle them via the normal whitelist.
It also adds a check for sheisty classes (jar hell with the jdk).
This is inspired by the lucene "sheisty" classes check, but it has false positives. This check is more evil, it validates every class file against the extension classloader as a resource, to see if it exists there. If so: jar hell.
This jar hell is a problem for several reasons:
1. causes insanely-hard-to-debug problems (like bugs in forbidden-apis)
2. hides problems (like internal api access)
3. the code you think is executing, is not really executing
4. security permissions are not what you think they are
5. brings in unnecessary dependencies
6. its jar hell
The more difficult problems are stuff like jython, where these classes are simply 'uberjared' directly in, so you cant just fix them by removing a bogus dependency. And there is a legit reason for them to do that, they want to support java 1.4.
This change removes hardcoded ports from cluster formation. It passes
port 0 for http and transport, and then uses a special property to have
the node log the ports used for http and transport (just for tests).
This does not yet work for multi node tests. This brings us one step
closer to working with --parallel.
This commit removes and now forbids all uses of
Collections#shuffle(List) and Random#<init>() across the codebase. The
rationale for removing and forbidding these methods is to increase test
reproducibility. As these methods use non-reproducible seeds, production
code and tests that rely on these methods contribute to
non-reproducbility of tests.
Instead of Collections#shuffle(List) the method
Collections#shuffle(List, Random) can be used. All that is required then
is a reproducible source of randomness. Consequently, the utility class
Randomness has been added to assist in creating reproducible sources of
randomness.
Instead of Random#<init>(), Random#<init>(long) with a reproducible seed
or the aforementioned Randomess class can be used.
Closes#15287
Wildcard imports are terrible, they cause ambiguity in the code,
make it not compile with the future versions of java in many cases.
We should simply fail the build on this, it is messiness, caused by
messy Intellij configuration
We have some tests which have crazy dependencies, like on other plugins.
This change adds a "messy-test" gradle plugin which can be used for qa
projects that these types of tests can run in. What this adds over
regular standalone tests is the plugin properties and metadata on the
classpath, so that the plugins are properly initialized.
We have eclipse settings added to all projects when running gradle
eclipse, but buildSrc is its own special project that is not
encapsulated by allprojects blocks. This adds eclipse settings to
buildSrc.
This is a relic from shading where it was trickier to implement.
Third party signatures are already in e.g. the test list, there
is no reason to separate them out.
Instead, we could have a third party signatures that does
something different... like keep tabs on third party libraries.
This commit removes and now forbids all uses of the type-unsafe empty
Collections fields Collections#EMPTY_LIST, Collections#EMPTY_MAP, and
Collections#EMPTY_SET. The type-safe methods Collections#emptyList,
Collections#emptyMap, and Collections#emptySet should be used instead.
Typical failure:
```
:test-framework:dependencyLicenses (Thread[main,5,main]) started.
:test-framework:dependencyLicenses
Executing task ':test-framework:dependencyLicenses' (up-to-date check took 0.0 secs) due to:
Task has not declared any outputs.
:test-framework:dependencyLicenses FAILED
:test-framework:dependencyLicenses (Thread[main,5,main]) completed. Took 0.023 secs.
FAILURE: Build failed with an exception.
* What went wrong:
Execution failed for task ':test-framework:dependencyLicenses'.
> Licences dir /mnt/jenkins/workspace/es_core_master_strong/test-framework/licenses does not exist, but there are dependencies
```
Related to #15168
This change attempts to simplify the gradle tasks for precommit. One
major part of that is using a "less groovy style", as well as being more
consistent about how tasks are created and where they are configured. It
also allows the things creating the tasks to set up inter task
dependencies, instead of assuming them (ie decoupling from tasks
eleswhere in the build).
We had increased this in maven, but it was lost in the transition to
gradle. This change adds it as a configurable setting the the logger for
randomized testing and bumps it to 25.
The current delay waits until later after normal configuration, but
still just after resolution (ie when paths would be known for
dependencies but not actual execution). This delays the checks further
to be done right before we actually execute the copy task.
closes#15068
Currently if running with --info, the command line for ES, along with
env vars, are logged before they may be ammended to add debug options.
This moves the adding JAVA_OPTS to before we print the command.
This adds the standalone tests so they will compile (and thus can be
modified with import completion) within IntelliJ. It also explicitly
sets up buildSrc as a module.
Note that this does *not* mean eg evil-tests can be run from intellij.
These are special tests that require special settings (eg disabling
security manager). They need to be run from the command line.
closes#15075
The Exec task outputs stdout/stderr to the standard streams by default.
However, to keep output short, we currently capture this, and only
output if the task failed. This change makes a small wrapper around Exec
to facilitate this behavior anywhere we use Exec.
This change delays the lookup for whatever is passed to extra config as
the source file to happen at execution time. This allows using eg a task
which generates a file, but maintains the checks that the file is not a
dir and that it exists at runtime.
This change allows copy extra files into the integ test cluster before
it runs. However, it explicitly forbids overwriting elasticsearch.yml,
since that is generated.
The current mechanism for adding plugins to the integTest cluster is to
have a FileCollection. This works well for the integTests for a single
plugin, which automatically adds itself to be installed. However, for qa
tests where many plugins may be installed, and from other projects, it
is cumbersome to add configurations, dependencies and dependsOn
statements over and over. This simplifies installing a plugin from
another project by moving this common setup into the cluster
configuration code.
The current wait condition for an integ test cluster being up is a
simple http get on the root path for elasticsearch. However, it is
useful to allow having arbitrary wait conditions. This change reworks
the wait task to first check that each node process started successfully
and has a socket up, followed by an arbitrary wait condition which
defaults to the current http get.
Also, cluster settings are allowed to be added, and overriden. Finally,
custom setup commands are made relative to the elasticsearch home dir
for each node.
Simon Willnauer
Ryan Ernst
Robert Muir
Nik Everett
Daniel Mitterdorfer
Jason Tedor
Yannick Welsch
Adrien Grand
Robert Muir
David Pilato