[
https://issues.apache.org/jira/browse/CASSANDRA-5009?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Brandon Williams updated CASSANDRA-5009:
----------------------------------------
Attachment: 5009.txt
Patch to maintain a set of completed ranges, then check at the end that every
range was successful or raise an exception.
This patch is against 1.1, but I'll note that when merged RangeStreamer should
look exactly the same. That is to say, all the FD business will be removed, as
I can't figure out how it makes any sense to check the FD at all since we're
going to block on the streaming latch first. If any ranges (not necessarily
streams) failed, then it's time to bail regardless of the FD, and if the FD
does convict before that it won't break us out of the latch wait.
> RangeStreamer has no way to report failures, allowing bootstrap/move etc to
> complete without data
> -------------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-5009
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5009
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 1.1.0
> Reporter: Brandon Williams
> Assignee: Brandon Williams
> Priority: Critical
> Fix For: 1.1.8
>
> Attachments: 5009.txt
>
>
> It looks like we fixed this for 1.2 by having RS.fetch() throw, but in 1.1 it
> does not and there doesn't appear to be a way to detect an RS failure, which
> among other things will cause bootstrap to succeed even though it failed to
> fetch any data.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
