[
https://issues.apache.org/jira/browse/CASSANDRA-5193?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Aleksey Yeschenko updated CASSANDRA-5193:
-----------------------------------------
Attachment: 5193.txt
The patch makes DiskAwareRunnable blacklist-aware and makes best_effort policy
work properly again.
> Memtable flushwriter can pick a blacklisted directory
> -----------------------------------------------------
>
> Key: CASSANDRA-5193
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5193
> Project: Cassandra
> Issue Type: Bug
> Affects Versions: 1.2.1
> Reporter: Aleksey Yeschenko
> Assignee: Aleksey Yeschenko
> Fix For: 1.2.2
>
> Attachments: 5193.txt
>
>
> The top-level data directory will be picked by DiskAwareRunnable (directory =
> Directories.getLocationCapableOfSize(writeSize)), and the top-level data
> directory itself might not be blacklisted (most likely won't be).
> For the same reason we can't just add a blacklist-check in the middle of
> Directories#getLocationCapableOfSize - most often it's the sstable directory
> that gets blacklisted.
> The issue seems to be caused by/related to CASSANDRA-4292, which was
> committed just two days prior 2116-2118 and undid some blacklist-aware
> directory-picking logic.
> Anyway, DiskAwareRunnable should be altered to respect directory blacklist.
--
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
