[jira] [Commented] (CASSANDRA-5223) Secondary index doesn't get repaired

["amorton (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-5223?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13571483#comment-13571483 ] 

amorton commented on CASSANDRA-5223:
------------------------------------

I tried running this but had to install the dtest package. I then got this when I ran it:

{noformat}
Traceback (most recent call last):
  File "secondary_index_repair_test.py", line 1, in <module>
    from dtest import Tester
  File "/Users/aaron/code/scratch/CASSANDRA-5223/.env/lib/python2.7/site-packages/dtest/__init__.py", line 105, in <module>
    from dtest.capture import Capturer
  File "/Users/aaron/code/scratch/CASSANDRA-5223/.env/lib/python2.7/site-packages/dtest/capture.py", line 102, in <module>
    from eventlet.corolocal import local
ImportError: No module named eventlet.corolocal
{noformat}

I'm also not sure where {{self.cluster}} is set. 

Can you remove all the unnecessary imports make it a simple script?

                
> Secondary index doesn't get repaired
> ------------------------------------
>
>                 Key: CASSANDRA-5223
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-5223
>             Project: Cassandra
>          Issue Type: Bug
>    Affects Versions: 1.2.1
>         Environment: Ubuntu dtest, laptop with 8G RAM
>            Reporter: Alexei Bakanov
>         Attachments: secondary_index_repair_test.py
>
>
> Looks like secondary indexes don't get repaired by any of the Cassandra repair mechanisms for NetworkTopologyStrategy. SimpleStrategy however works fine.
> The issue started as a mail to Cassandra userlist: http://article.gmane.org/gmane.comp.db.cassandra.user/30909
> I striped reproduction recipe down and made a d-test which inserts two rows into a cluster of two nodes, where the first node goes down after the first insert and goes back up after the second insert. Secondary index doesn't get repaired by neither hinted-handoff, read repair, nor manual 'repair' operation which the d-test triggers before reading the data.
> Moreover the second row is not visible in the index unless I do CL.ALL read or drop/create the index.

--
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