[
https://issues.apache.org/jira/browse/CASSANDRA-19007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17784904#comment-17784904
]
Andres de la Peña commented on CASSANDRA-19007:
-----------------------------------------------
We have mentioned in Slack conversation that read repair on a subset of the
columns might create the kind of partial writes that produce the consistency
issue. I'm not sure that's so. The issue is due to partial writes that are only
seen by a replica. The partial writes introduced by read repair can propagate
the most recent changes of a column to other replicas. So after RR we have a
partial write, yes, but that value ends up in both source and destination
replicas, making the data more consistent than it was. So I think it won't
produce the scenario we are dealing with.
I haven't figured out an example where user writes are not partial and RR
introduces a consistency issue. I think that if we aren't able to produce such
an example any guidance we provide about this issue can indicate that users are
safe from consistency issues on multi-column replica-side filtering queries if
they don't use partial writes.
> Queries with multi-column replica-side filtering can miss rows
> --------------------------------------------------------------
>
> Key: CASSANDRA-19007
> URL: https://issues.apache.org/jira/browse/CASSANDRA-19007
> Project: Cassandra
> Issue Type: Bug
> Components: Consistency/Coordination
> Reporter: Andres de la Peña
> Assignee: Caleb Rackliffe
> Priority: Normal
> Fix For: 5.0.x, 5.x
>
>
> {{SELECT}} queries with multi-column replica-side filtering can miss rows if
> the filtered columns are spread across out-of-sync replicas. This dtest
> reproduces the issue:
> {code:java}
> @Test
> public void testMultiColumnReplicaSideFiltering() throws IOException
> {
> try (Cluster cluster = init(Cluster.build().withNodes(2).start()))
> {
> cluster.schemaChange(withKeyspace("CREATE TABLE %s.t (k int PRIMARY
> KEY, a int, b int)"));
> // insert a split row
> cluster.get(1).executeInternal(withKeyspace("INSERT INTO %s.t(k, a)
> VALUES (0, 1)"));
> cluster.get(2).executeInternal(withKeyspace("INSERT INTO %s.t(k, b)
> VALUES (0, 2)"));
> String select = withKeyspace("SELECT * FROM %s.t WHERE a = 1 AND b =
> 2 ALLOW FILTERING");
> Object[][] initialRows = cluster.coordinator(1).execute(select, ALL);
> assertRows(initialRows, row(0, 1, 2)); // not found!!
> }
> }
> {code}
> This edge case affects queries using {{ALLOW FILTERING}} or any index
> implementation.
> It affects all branches since multi-column replica-side filtering queries
> were introduced, long before 3.0.
> The protection mechanism added by CASSANDRA-8272/8273 won't deal with this
> case, since it only solves single-column conflicts where stale rows could
> resurrect. This bug however doesn't resurrect data, it can only miss rows
> while the replicas are out-of-sync.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
