* Increase the max waiting time before checking the FS health after restoring the FS status in the test
Signed-off-by: Tianli Feng <ftianli@amazon.com>
* Remove another duplicate assetion statement
Signed-off-by: Tianli Feng <ftianli@amazon.com>
This commit removes LegacyESVersion.V_6_2_x constants including all
pre-release versions and bug fixes.
Signed-off-by: Nicholas Walter Knize <nknize@apache.org>
This commit removes LegacyESVersion.V_6_1_x constants including all
pre-release versions and bug fixes.
Signed-off-by: Nicholas Walter Knize <nknize@apache.org>
This PR removes LegacyESVersion.V_6_0_* constants including all pre-release
versions and bug fixes.
Signed-off-by: Nicholas Walter Knize <nknize@apache.org>
In preparation for removing all LegacyESVersion support by 3.0; this commit
largely refactors the LegacyESVersion test logic from the OpenSearch Version
test logic into an independent test class. This PR also updates Version.fromString
to ensure a proper legacy version is returned when major is > 3 (to support
legacy yaml test and build scripts).
Note that bwc w/ legacy versions are still supported so some cross compatibility
testing is retained in the Version test class.
Signed-off-by: Nicholas Walter Knize <nknize@apache.org>
* Added more detailed logging for SSLHandshakeException
Signed-off-by: Ryan Bogan <rbogan@amazon.com>
* Fixed spotless errors for logging change
Signed-off-by: Ryan Bogan <rbogan@amazon.com>
* Changed message to include original error and limited to specific instance
Signed-off-by: Ryan Bogan <rbogan@amazon.com>
* Applied troubleshooting message to all instances of SSlHandshakeException
Signed-off-by: Ryan Bogan <rbogan@amazon.com>
Zen1 discovery was deprecated in Legacy 7.x for eventual removal. OpenSearch 1.x
carries this deprecation. This commit completely removes all support for Zen1
discovery in favor of Zen2.
Signed-off-by: Nicholas Walter Knize <nknize@apache.org>
Lucene 9 removes support for SimpleFS File System format. This PR completely
removes SimpleFS support which was deprecated in a previous PR.
Signed-off-by: Nicholas Walter Knize <nknize@apache.org>
* Signed-off-by: Megha Sai Kavikondala <kavmegha@amazon.com>
Changes related to Informative error messages.
* Signed-off-by: Megha Sai Kavikondala <kavmegha@amazon.com>
Spotless changes
* Signed-off-by: Megha Sai Kavikondala <kavmegha@amazon.com>
naming and message changes.
Signed-off-by: Megha Sai Kavikondala <kavmegha@amazon.com>
* Signed-off-by: Megha Sai Kavikondala <kavmegha@amazon.com>
Deleting the log message.
The 1.1 release notes are missing a set of commits. This updates the release notes
to the commit specified in our relase 15e9f13.
Signed-off-by: Marc Handalian <handalm@amazon.com>
I observed a test failure with the message
'Attempted to append to non-started appender mock' from an assertion in
`OpenSearchTestCase::after`. I believe this indicates that a
MockLogAppender (which is named "mock") was added as an appender to the
static logging context and some other test in the same JVM happened to
cause a logging statement to hit that appender and cause an error, which
then caused an unrelated test to fail (because they share static state
with the logger). Almost all usages of MockLogAppender start it
immediately after creation. I found a few that did not and fixed those.
I also made a static helper in MockLogAppender to start it upon
creation.
Signed-off-by: Andrew Ross <andrross@amazon.com>
* Fix documentation reference from Java 14 to 17
Signed-off-by: Andrew Ross <andrross@amazon.com>
* Update developer guide reference to download JDK 14
Signed-off-by: Andrew Ross <andrross@amazon.com>
* initial commit of CODEOWNERS
Signed-off-by: CEHENKLE <henkle@amazon.com>
* put the @ sign in the right spot
Signed-off-by: CEHENKLE <henkle@amazon.com>
* updating wild card
Signed-off-by: CEHENKLE <henkle@amazon.com>
* removed extra comment
Signed-off-by: CEHENKLE <henkle@amazon.com>
* you know what? I didn't like that other wildcard. f++ would nnot wildcard again
Signed-off-by: CEHENKLE <henkle@amazon.com>
* Made changes.
Signed-off-by: Megha Sai Kavikondala <kavmegha@amazon.com>
* Signed-off-by: Megha Sai Kavikondala <kavmegha@amazon.com>
Changes made by deleting the TestSettingsIT file and adding new lines in FullClusterRestartSettingsUpgradeIT.java
I observed a [test failure][1] that I believe was caused by the fact
that an OpenSearch server happened to be running at localhost:9200 when
these unit tests were executed. The code under test has logic to try to
connect to localhost:9200 and then fall back to defaults if that port is
not open. The tests expect these defaults and will fail if different
data is returned. The change here is to use a non-default port to make
it very unlikely that a real instance will be running at the non-default
port. I don't know why an OpenSearch service happened to be running
during the linked test failure, but I think making these unit tests more
independent and isolated is helpful no matter the underlying cause in
this case.
[1]: https://fork-jenkins.searchservices.aws.dev/job/OpenSearch_CI/job/PR_Checks/job/Gradle_Check/975/artifact/gradle_check_975.log/*view*/
Signed-off-by: Andrew Ross <andrross@amazon.com>
* Added logic to allow {dot} files on startup
Signed-off-by: Ryan Bogan <rbogan@amazon.com>
* Ensures that only plugin directories are returned by findPluginDirs()
Signed-off-by: Ryan Bogan <rbogan@amazon.com>
* Prevents . files from being returned as plugins
Signed-off-by: Ryan Bogan <rbogan@amazon.com>
This change updates the logic in RunTask to use deterministic port values for all nodes rather than just the first node in a cluster. The local address of the first node is also used as the seed_hosts value to allow the cluster to form correctly. The seed_hosts setting is also allow-listed as an overrideable value.
Signed-off-by: Kartik Ganesh <85275476+kartg@users.noreply.github.com>