[jira] [Commented] (CASSANDRA-11996) SSTableSet.CANONICAL can miss sstables
[ https://issues.apache.org/jira/browse/CASSANDRA-11996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15360832#comment-15360832 ] Stefania commented on CASSANDRA-11996: -- utests look good now, all failures happening on unpatched branches as well. Ticket can be committed as soon as CASSANDRA-11944 is ready. > SSTableSet.CANONICAL can miss sstables > -- > > Key: CASSANDRA-11996 > URL: https://issues.apache.org/jira/browse/CASSANDRA-11996 > Project: Cassandra > Issue Type: Bug >Reporter: Marcus Eriksson >Assignee: Marcus Eriksson >Priority: Critical > Fix For: 3.0.x, 3.x > > > There is a race where we might miss sstables in SSTableSet.CANONICAL when we > finish up a compaction. > Reproducing unit test pushed > [here|https://github.com/krummas/cassandra/commit/1292aaa61b89730cff0c022ed1262f45afd493e5] -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-11996) SSTableSet.CANONICAL can miss sstables
[ https://issues.apache.org/jira/browse/CASSANDRA-11996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15360735#comment-15360735 ] Stefania commented on CASSANDRA-11996: -- Patch LGTM but the utests failed to launch on all branches so I've restarted them. dtest failures are all accounted for (either known failures, or connection timeouts). > SSTableSet.CANONICAL can miss sstables > -- > > Key: CASSANDRA-11996 > URL: https://issues.apache.org/jira/browse/CASSANDRA-11996 > Project: Cassandra > Issue Type: Bug >Reporter: Marcus Eriksson >Assignee: Marcus Eriksson >Priority: Critical > Fix For: 3.0.x, 3.x > > > There is a race where we might miss sstables in SSTableSet.CANONICAL when we > finish up a compaction. > Reproducing unit test pushed > [here|https://github.com/krummas/cassandra/commit/1292aaa61b89730cff0c022ed1262f45afd493e5] -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-11996) SSTableSet.CANONICAL can miss sstables
[ https://issues.apache.org/jira/browse/CASSANDRA-11996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15358457#comment-15358457 ] Marcus Eriksson commented on CASSANDRA-11996: - bq. The ticket you are referring to is actually CASSANDRA-11944? oops, yep > SSTableSet.CANONICAL can miss sstables > -- > > Key: CASSANDRA-11996 > URL: https://issues.apache.org/jira/browse/CASSANDRA-11996 > Project: Cassandra > Issue Type: Bug >Reporter: Marcus Eriksson >Assignee: Marcus Eriksson >Priority: Critical > Fix For: 3.0.x, 3.x > > > There is a race where we might miss sstables in SSTableSet.CANONICAL when we > finish up a compaction. > Reproducing unit test pushed > [here|https://github.com/krummas/cassandra/commit/1292aaa61b89730cff0c022ed1262f45afd493e5] -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-11996) SSTableSet.CANONICAL can miss sstables
[ https://issues.apache.org/jira/browse/CASSANDRA-11996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15358333#comment-15358333 ] Stefania commented on CASSANDRA-11996: -- Thanks, I'm out today but I will review this first thing next week. The ticket you are referring to is actually CASSANDRA-11944? > SSTableSet.CANONICAL can miss sstables > -- > > Key: CASSANDRA-11996 > URL: https://issues.apache.org/jira/browse/CASSANDRA-11996 > Project: Cassandra > Issue Type: Bug >Reporter: Marcus Eriksson >Assignee: Marcus Eriksson >Priority: Critical > Fix For: 3.0.x, 3.x > > > There is a race where we might miss sstables in SSTableSet.CANONICAL when we > finish up a compaction. > Reproducing unit test pushed > [here|https://github.com/krummas/cassandra/commit/1292aaa61b89730cff0c022ed1262f45afd493e5] -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-11996) SSTableSet.CANONICAL can miss sstables
[ https://issues.apache.org/jira/browse/CASSANDRA-11996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15356989#comment-15356989 ] Marcus Eriksson commented on CASSANDRA-11996: - try nr 2: ||branch||testall||dtest|| |[marcuse/canonical|https://github.com/krummas/cassandra/tree/marcuse/canonical]|[testall|http://cassci.datastax.com/view/Dev/view/krummas/job/krummas-marcuse-canonical-testall]|[dtest|http://cassci.datastax.com/view/Dev/view/krummas/job/krummas-marcuse-canonical-dtest]| |[marcuse/canonical-3.9|https://github.com/krummas/cassandra/tree/marcuse/canonical-3.9]|[testall|http://cassci.datastax.com/view/Dev/view/krummas/job/krummas-marcuse-canonical-3.9-testall]|[dtest|http://cassci.datastax.com/view/Dev/view/krummas/job/krummas-marcuse-canonical-3.9-dtest]| |[marcuse/canonical-trunk|https://github.com/krummas/cassandra/tree/marcuse/canonical-trunk]|[testall|http://cassci.datastax.com/view/Dev/view/krummas/job/krummas-marcuse-canonical-trunk-testall]|[dtest|http://cassci.datastax.com/view/Dev/view/krummas/job/krummas-marcuse-canonical-trunk-dtest]| same idea as before - get the sstables like in 2.2 but apply the filter *after* getting the requested SSTableSet based it on CASSANDRA-11994 since that changed how we get sstable in bounds. > SSTableSet.CANONICAL can miss sstables > -- > > Key: CASSANDRA-11996 > URL: https://issues.apache.org/jira/browse/CASSANDRA-11996 > Project: Cassandra > Issue Type: Bug >Reporter: Marcus Eriksson >Assignee: Marcus Eriksson >Priority: Critical > Fix For: 3.0.x, 3.x > > > There is a race where we might miss sstables in SSTableSet.CANONICAL when we > finish up a compaction. > Reproducing unit test pushed > [here|https://github.com/krummas/cassandra/commit/1292aaa61b89730cff0c022ed1262f45afd493e5] -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-11996) SSTableSet.CANONICAL can miss sstables
[
https://issues.apache.org/jira/browse/CASSANDRA-11996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15354775#comment-15354775
]
Marcus Eriksson commented on CASSANDRA-11996:
-
bq. Shouldn't we only return the canonical sstables that are passed into select
as a parameter?
true, argh.. this approach wont work at all - the sstables passed in do not
contain the original version of the sstable, JUST the early opened one when we
are finishing up the compaction...
bq. In the test, don't we need to call
DatabaseDescriptor.setSSTablePreempiveOpenIntervalInMB()
no, this happens since we always open an "early" sstable when finishing writing
an sstable
https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/io/sstable/SSTableRewriter.java#L316
- the {{setPreemptive...}} i had in the finally clause was a copy and paste
leftover
> SSTableSet.CANONICAL can miss sstables
> --
>
> Key: CASSANDRA-11996
> URL: https://issues.apache.org/jira/browse/CASSANDRA-11996
> Project: Cassandra
> Issue Type: Bug
>Reporter: Marcus Eriksson
>Assignee: Marcus Eriksson
>Priority: Critical
> Fix For: 3.0.x, 3.x
>
>
> There is a race where we might miss sstables in SSTableSet.CANONICAL when we
> finish up a compaction.
> Reproducing unit test pushed
> [here|https://github.com/krummas/cassandra/commit/1292aaa61b89730cff0c022ed1262f45afd493e5]
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (CASSANDRA-11996) SSTableSet.CANONICAL can miss sstables
[
https://issues.apache.org/jira/browse/CASSANDRA-11996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15354776#comment-15354776
]
Marcus Eriksson commented on CASSANDRA-11996:
-
bq. Shouldn't we only return the canonical sstables that are passed into select
as a parameter?
true, argh.. this approach wont work at all - the sstables passed in do not
contain the original version of the sstable, JUST the early opened one when we
are finishing up the compaction...
bq. In the test, don't we need to call
DatabaseDescriptor.setSSTablePreempiveOpenIntervalInMB()
no, this happens since we always open an "early" sstable when finishing writing
an sstable
https://github.com/apache/cassandra/blob/trunk/src/java/org/apache/cassandra/io/sstable/SSTableRewriter.java#L316
- the {{setPreemptive...}} i had in the finally clause was a copy and paste
leftover
> SSTableSet.CANONICAL can miss sstables
> --
>
> Key: CASSANDRA-11996
> URL: https://issues.apache.org/jira/browse/CASSANDRA-11996
> Project: Cassandra
> Issue Type: Bug
>Reporter: Marcus Eriksson
>Assignee: Marcus Eriksson
>Priority: Critical
> Fix For: 3.0.x, 3.x
>
>
> There is a race where we might miss sstables in SSTableSet.CANONICAL when we
> finish up a compaction.
> Reproducing unit test pushed
> [here|https://github.com/krummas/cassandra/commit/1292aaa61b89730cff0c022ed1262f45afd493e5]
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (CASSANDRA-11996) SSTableSet.CANONICAL can miss sstables
[
https://issues.apache.org/jira/browse/CASSANDRA-11996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15354722#comment-15354722
]
Stefania commented on CASSANDRA-11996:
--
Shouldn't we only return the canonical sstables that are passed into select as
a
[parameter|https://github.com/apache/cassandra/compare/trunk...krummas:marcuse/canonical#diff-6a3c38153a6ec0316c9192d40f71fb40R139]?
By looking
[here|https://github.com/apache/cassandra/compare/trunk...krummas:marcuse/canonical#diff-6a3c38153a6ec0316c9192d40f71fb40R151]
and
[here|https://github.com/apache/cassandra/compare/trunk...krummas:marcuse/canonical#diff-6a3c38153a6ec0316c9192d40f71fb40R159],
won't we return all compacting sstables that are not opened early, regardless
of the filter applied by the callers of select?
In the test, don't we need to call
{{DatabaseDescriptor.setSSTablePreempiveOpenIntervalInMB()}} with a lower
interval at the beginning of the test as well?
Nit: in the test, the
[runnable|https://github.com/apache/cassandra/compare/trunk...krummas:marcuse/canonical#diff-02c4b3ac06af7b36bb219956c41616f6R887]
can be a lambda.
> SSTableSet.CANONICAL can miss sstables
> --
>
> Key: CASSANDRA-11996
> URL: https://issues.apache.org/jira/browse/CASSANDRA-11996
> Project: Cassandra
> Issue Type: Bug
>Reporter: Marcus Eriksson
>Assignee: Marcus Eriksson
>Priority: Critical
> Fix For: 3.0.x, 3.x
>
>
> There is a race where we might miss sstables in SSTableSet.CANONICAL when we
> finish up a compaction.
> Reproducing unit test pushed
> [here|https://github.com/krummas/cassandra/commit/1292aaa61b89730cff0c022ed1262f45afd493e5]
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (CASSANDRA-11996) SSTableSet.CANONICAL can miss sstables
[ https://issues.apache.org/jira/browse/CASSANDRA-11996?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15327505#comment-15327505 ] Marcus Eriksson commented on CASSANDRA-11996: - note that the unit test will not fail every run, run it in a loop to reproduce > SSTableSet.CANONICAL can miss sstables > -- > > Key: CASSANDRA-11996 > URL: https://issues.apache.org/jira/browse/CASSANDRA-11996 > Project: Cassandra > Issue Type: Bug >Reporter: Marcus Eriksson >Priority: Critical > Fix For: 3.0.x, 3.x > > > There is a race where we might miss sstables in SSTableSet.CANONICAL when we > finish up a compaction. > Reproducing unit test pushed > [here|https://github.com/krummas/cassandra/commit/1292aaa61b89730cff0c022ed1262f45afd493e5] -- This message was sent by Atlassian JIRA (v6.3.4#6332)
