[
https://issues.apache.org/jira/browse/CASSANDRA-11349?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15278032#comment-15278032
]
Branimir Lambov commented on CASSANDRA-11349:
---------------------------------------------
Not precisely: One part of the problem is that we cannot ensure that e.g. the
same range tombstone will not come twice from the same sstable (in which case
{{MergeIterator}} would issue to separate {{getReduced}} calls) or from two
different sstables (in which case {{MergeIterator}} would call {{getReduced}}
once with both) or some complex combination of these. Another is that while
compaction uses the tracker to identify when a tombstone is redundant and can
be omitted, {{getReduced}} does not have that information at the time it
processes that tombstone because the covering tombstone has not arrived yet.
The tracker can properly resolve these situations, but it can't do it without
delaying which causes the necessity for abusing the serializer.
The reducer only adds RTs to the tracker if it would not return them in the
output for some reason (e.g. expiration), the point being to always pass on the
full stream of RTs; the order should not be affected by it choosing to do that.
> MerkleTree mismatch when multiple range tombstones exists for the same
> partition and interval
> ---------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-11349
> URL: https://issues.apache.org/jira/browse/CASSANDRA-11349
> Project: Cassandra
> Issue Type: Bug
> Reporter: Fabien Rousseau
> Assignee: Stefan Podkowinski
> Labels: repair
> Fix For: 2.1.x, 2.2.x
>
> Attachments: 11349-2.1-v2.patch, 11349-2.1-v3.patch, 11349-2.1.patch
>
>
> We observed that repair, for some of our clusters, streamed a lot of data and
> many partitions were "out of sync".
> Moreover, the read repair mismatch ratio is around 3% on those clusters,
> which is really high.
> After investigation, it appears that, if two range tombstones exists for a
> partition for the same range/interval, they're both included in the merkle
> tree computation.
> But, if for some reason, on another node, the two range tombstones were
> already compacted into a single range tombstone, this will result in a merkle
> tree difference.
> Currently, this is clearly bad because MerkleTree differences are dependent
> on compactions (and if a partition is deleted and created multiple times, the
> only way to ensure that repair "works correctly"/"don't overstream data" is
> to major compact before each repair... which is not really feasible).
> Below is a list of steps allowing to easily reproduce this case:
> {noformat}
> ccm create test -v 2.1.13 -n 2 -s
> ccm node1 cqlsh
> CREATE KEYSPACE test_rt WITH replication = {'class': 'SimpleStrategy',
> 'replication_factor': 2};
> USE test_rt;
> CREATE TABLE IF NOT EXISTS table1 (
> c1 text,
> c2 text,
> c3 float,
> c4 float,
> PRIMARY KEY ((c1), c2)
> );
> INSERT INTO table1 (c1, c2, c3, c4) VALUES ( 'a', 'b', 1, 2);
> DELETE FROM table1 WHERE c1 = 'a' AND c2 = 'b';
> ctrl ^d
> # now flush only one of the two nodes
> ccm node1 flush
> ccm node1 cqlsh
> USE test_rt;
> INSERT INTO table1 (c1, c2, c3, c4) VALUES ( 'a', 'b', 1, 3);
> DELETE FROM table1 WHERE c1 = 'a' AND c2 = 'b';
> ctrl ^d
> ccm node1 repair
> # now grep the log and observe that there was some inconstencies detected
> between nodes (while it shouldn't have detected any)
> ccm node1 showlog | grep "out of sync"
> {noformat}
> Consequences of this are a costly repair, accumulating many small SSTables
> (up to thousands for a rather short period of time when using VNodes, the
> time for compaction to absorb those small files), but also an increased size
> on disk.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
