You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Varun Thacker (JIRA)" <ji...@apache.org> on 2015/11/18 02:22:10 UTC
[jira] [Commented] (SOLR-8309) TestRandomRequestDistribution test
failures
[ https://issues.apache.org/jira/browse/SOLR-8309?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15009991#comment-15009991 ]
Varun Thacker commented on SOLR-8309:
-------------------------------------
This is an excerpt from a successful test run:
{code}
[junit4] 2> 283122 INFO (zkCallback-171-thread-1-processing-n:127.0.0.1:56703_u_rq) [n:127.0.0.1:56703_u_rq ] o.a.s.c.c.ZkStateReader Updating data for football from 6 to 7
[junit4] 2> 283122 INFO (zkCallback-175-thread-1-processing-n:127.0.0.1:56707_u_rq) [n:127.0.0.1:56707_u_rq ] o.a.s.c.c.ZkStateReader Updating data for football from 6 to 7
[junit4] 2> 283176 INFO (TEST-TestRandomRequestDistribution.test-seed#[A0F2F7B493EFD7DE]) [n:127.0.0.1:56637_u_rq c:collection1 s:shard2 r:core_node3 x:collection1] o.a.s.c.AbstractFullDistribZkTestBase Firing query against path=http://127.0.0.1:56707/u_rq/football
{code}
This is an excerpt from build #624
{code}
[junit4] 2> 555088 INFO (TEST-TestRandomRequestDistribution.test-seed#[A0F2F7B493EFD7DE]) [ ] o.a.s.c.AbstractFullDistribZkTestBase Firing query against path=http://127.0.0.1:59394/u_rq/football
[junit4] 2> 555091 INFO (zkCallback-357-thread-1-processing-n:127.0.0.1:43311_u_rq) [n:127.0.0.1:43311_u_rq ] o.a.s.c.c.ZkStateReader Updating data for football from 6 to 7
[junit4] 2> 555091 INFO (zkCallback-353-thread-1-processing-n:127.0.0.1:59394_u_rq) [n:127.0.0.1:59394_u_rq ] o.a.s.c.c.ZkStateReader Updating data for football from 6 to 7
{code}
The difference seems to be that the zkCallback happens after the query is fired in the failed test runs ( same in build 3774 ) vs before in successful test runs.
So essentially the cluster state does not know about the change when the query is fired hence it executes it against the down replica.
Telling the test to sleep for sometime before executing the query will probably solve the problem but is there a better way to solve the problem?
> TestRandomRequestDistribution test failures
> -------------------------------------------
>
> Key: SOLR-8309
> URL: https://issues.apache.org/jira/browse/SOLR-8309
> Project: Solr
> Issue Type: Bug
> Reporter: Varun Thacker
> Assignee: Varun Thacker
> Attachments: build-3774.txt, build-624.txt
>
>
> There have been a couple of Jenkins failures for TestRandomRequestDistribution .
> Creating a Jira to track it
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org