[
https://issues.apache.org/jira/browse/CASSANDRA-8669?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14481578#comment-14481578
]
Yuki Morishita commented on CASSANDRA-8669:
-------------------------------------------
bq. the snapshots should contain all of the compacting sstables, rather than be
empty? But how does the empty snapshot interact with repair (sounds like it
should stream nothing, if we operate off the snapshots..)?
Empty snapshot produces MerkleTree with root hash of 'empty'. If the other
replica has MerkleTree that is not 'empty', then the node thinks two trees are
different and streaming of MT range happens.
Worst case is that both end produces empty MT where it should produce
difference. In this case, both MT are empty and the nodes think there is no
difference and repair does not happen.
It seems that every SSTable goes through the state of EARLY when calling
{{SSTableRewriter#finish()}}. I think this issue happens during that time.
> simple_repair test failing on 2.1
> ---------------------------------
>
> Key: CASSANDRA-8669
> URL: https://issues.apache.org/jira/browse/CASSANDRA-8669
> Project: Cassandra
> Issue Type: Bug
> Reporter: Philip Thompson
> Assignee: Yuki Morishita
> Fix For: 2.1.5
>
> Attachments: 8669-debug.txt, debug-result.txt
>
>
> The dtest simple_repair_test began failing on 12/22 on 2.1 and trunk. The
> test fails intermittently both locally and on cassci.
> The test is here:
> https://github.com/riptano/cassandra-dtest/blob/master/repair_test.py#L32
> The output is here:
> http://cassci.datastax.com/job/cassandra-2.1_dtest/661/testReport/repair_test/TestRepair/simple_repair_test/
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
