cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sylvain Lebresne (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-10148) batchlog dtests regression
Date Fri, 21 Aug 2015 09:29:45 GMT


Sylvain Lebresne commented on CASSANDRA-10148:

It's probably worth looking into why that test has started failing consistently from that
commit as maybe there is something fishy, but I'll note that there is in theory nothing wrong
with the test getting an UnavailableException instead of a TimedOutException: it just means
that the failure detector has been fast at detecting the node failures. That is, when a test
kills a node so that a particular consistency cannot be achieved, then said test should always
assume it can get either a timeout or an unavailable since getting one or the other depends
on timing internal to the server.

> batchlog dtests regression
> --------------------------
>                 Key: CASSANDRA-10148
>                 URL:
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Jim Witschey
>            Assignee: Ariel Weisberg
> {{acknowledged_by_batchlog_not_set_when_batchlog_write_fails_test}} and {{acknowledged_by_batchlog_set_when_batchlog_write_succeeds_test}}
started failing at the same commit on CassCI:
> Looking at the commit where the errors started it's not clear what could have caused
the errors. The errors look something like:
> {code}
> Expecting TimedOutException, got:code=1000 [Unavailable exception] message="Cannot achieve
consistency level THREE" info={'required_replicas': 3, 'alive_replicas': 2, 'consistency':
> {code}

This message was sent by Atlassian JIRA

View raw message