cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sylvain Lebresne (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-11197) upgrade bootstrap tests flap when migration tasks fail
Date Mon, 22 Feb 2016 10:25:18 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-11197?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15156738#comment-15156738
] 

Sylvain Lebresne commented on CASSANDRA-11197:
----------------------------------------------

Actually, seems that error is logged when, on bootstrap, migration hasn't finished after {{cassandra.migration_task_wait_in_seconds}}
which is one second by default (this is new code from CASSANDRA-10731). And I can't help notice
that all the failures of this I've seen are from "node5" which means the test use at least
5 nodes, which probably make CI slow down substantially make the default of 1 seconds likely
too low.

So ideally we'd really want to bump {{cassandra.migration_task_wait_in_seconds}} when there
is more nodes. Maybe that's even something that can be done automatically by ccm [~philipthompson][~mambocab][~rhatch].

> upgrade bootstrap tests flap when migration tasks fail
> ------------------------------------------------------
>
>                 Key: CASSANDRA-11197
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-11197
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Jim Witschey
>            Assignee: DS Test Eng
>              Labels: dtest
>
> I've seen these tests flap:
> {code}
> upgrade_tests/upgrade_through_versions_test.py:ProtoV4Upgrade_3_1_UpTo_3_2_HEAD.bootstrap_test
> upgrade_tests/upgrade_through_versions_test.py:ProtoV4Upgrade_3_3_UpTo_Trunk_HEAD.bootstrap_test
> upgrade_tests/upgrade_through_versions_test.py:ProtoV3Upgrade_3_0_UpTo_3_1_HEAD.bootstrap_multidc_test
> upgrade_tests/upgrade_through_versions_test.py:ProtoV3Upgrade_3_2_UpTo_3_3_HEAD.bootstrap_multidc_test
> upgrade_tests/upgrade_through_versions_test.py:ProtoV3Upgrade_3_3_UpTo_Trunk_HEAD.bootstrap_multidc_test
> upgrade_tests/upgrade_through_versions_test.py:ProtoV4Upgrade_3_3_UpTo_Trunk_HEAD.bootstrap_multidc_test
> {code}
> There may be more upgrade paths that flap, I'm not sure. All the failures I've seen look
like this:
> {code}
> Unexpected error in node5 node log: ['ERROR [main] 2016-02-18 20:05:13,012 MigrationManager.java:164
- Migration task failed to complete\nERROR [main] 2016-02-18 20:05:14,012 MigrationManager.java:164
- Migration task failed to complete']
> {code}
> [~rhatch] Do these look familiar at all?



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message