Switch more tests to zen2 #36367
Merged
Switch more tests to zen2 #36367
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
andrershov
added
the
:Distributed Coordination/Cluster Coordination
|
Pinging @elastic/es-distributed |
ywelsch
suggested changes
Dec 7, 2018
| @@ -276,11 +276,18 @@ public void testTwoNodesSingleDoc() throws Exception { | |||
| } | |||
|
|
|||
| public void testDanglingIndices() throws Exception { | |||
| /*TODO This test test does not work with Zen2, because once master node looses its cluster state during restart | |||
There was a problem hiding this comment.
thanks for the analysis on this one. It still has Zen2 disabled though? Can you just add the TODO comment, but leave the test unchanged until we have a proper fix?
| @@ -328,14 +335,14 @@ public boolean clearData(String nodeName) { | |||
| */ | |||
| public void testIndexDeletionWhenNodeRejoins() throws Exception { | |||
| final String indexName = "test-index-del-on-node-rejoin-idx"; | |||
| final int numNodes = 2; | |||
| // We need at least 3 nodes to make sure, that once one node is stopped, remaining nodes can elect a new master | |||
There was a problem hiding this comment.
I think the proper fix here is to add the down-scale / up-scale logic that's in InternalTestCluster#stopNodesAndClients to restartNode as well, just as we do with the autoManageMinMasterNodes scaling. The test should then not need any adaptation
| @@ -555,8 +546,10 @@ public Settings onNodeStopped(String nodeName) throws Exception { | |||
| } | |||
|
|
|||
| public void testResolvePath() throws Exception { | |||
| internalCluster().startMasterOnlyNode(); | |||
There was a problem hiding this comment.
I think this one can be fixed by moving the scope to ESIntegTestCase.Scope.TEST. As these tests are starting up nodes anyway, there's no need to keep the cluster per suite. Maybe that also fixes the testCorruptTranslogTruncationOfReplica
There was a problem hiding this comment.
4d9d141 this helps this test but does not help another test in this class (it has different root cause)
added 4 commits
December 7, 2018 18:39
Without changes test fails, because Zen2 retries snapshot creation as soon as network partition heals. This results into race between creating snapshot and test cleanup logic (deleting index). Zen1 on the other hand, also schedules retry, but it takes some time after network partition heals, so cleanup logic executes latter and test passes. The check that snapshot is eventually created is added to the end of the test
ywelsch
approved these changes
Dec 10, 2018
| nodeAndClient.recreateNode(newSettings, () -> rebuildUnicastHostFiles(emptyList())); | ||
| nodeAndClient.startNode(); | ||
| if (activeDisruptionScheme != null) { | ||
| activeDisruptionScheme.applyToNode(nodeAndClient.name, this); | ||
| } | ||
| if (callback.validateClusterForming() || updateMinMaster) { | ||
|
|
||
| if (callback.validateClusterForming() && excludedNodeIds.isEmpty() == false) { |
There was a problem hiding this comment.
This should be an ||, not an &&?
andrershov
force-pushed
the
zen2_more_tests
branch
from
6e467bf to
a500d20
Compare
jasontedor
added a commit
to dnhatn/elasticsearch
that referenced
this pull request
Dec 11, 2018
* elastic/master: (36 commits) Add check for minimum required Docker version (elastic#36497) Minor search controller changes (elastic#36479) Add default methods to DocValueFormat (elastic#36480) Fix the mixed cluster REST test explain/11_basic_with_types. Modify `BigArrays` to take name of circuit breaker (elastic#36461) Move LoggedExec to minimumRuntime source set (elastic#36453) added 6.5.4 version Add test logging for elastic#35644 Tests- added helper methods to ESRestTestCase for checking warnings (elastic#36443) SQL: move requests' parameters to requests JSON body (elastic#36149) [Zen2] Respect the no_master_block setting (elastic#36478) Require soft-deletes when access changes snapshot (elastic#36446) Switch more tests to zen2 (elastic#36367) [Painless] Add extensive tests for def to primitive casts (elastic#36455) Add setting to bypass Rollover action (elastic#36235) Try running CI against Zulu (elastic#36391) [DOCS] Reworked the shard allocation filtering info. (elastic#36456) Log [initial_master_nodes] on formation failure (elastic#36466) converting ForbiddenPatternsTask to .java (elastic#36194) fixed typo ...
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
testDanglingIndicesrequire changes the source code (added TODO).testIndexDeletionWhenNodeRejoinsbecause it's using just twonodes, adding the node to exclusions is needed on restart.
testCorruptTranslogTruncationOfReplicastarts dedicated masterone, because otherwise, the cluster does not form, if nodes are stopped
and one node is started back.
testResolvePathneeds TEST cluster, because all nodes are stoppedat the end of the test and it's not possible to perform checks needed
by SUITE cluster.
SnapshotDisruptionIT. Without changes, the test fails because Zen2retries snapshot creation as soon as network partition heals. This
results into the race between creating snapshot and test cleanup logic
(deleting index). Zen1 on the
other hand, also schedules retry, but it takes some time after network
partition heals, so cleanup logic executes latter and test passes. The
check that snapshot is eventually created is added to
the end of the test.