cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Michael Shuler (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (CASSANDRA-6414) BlacklistingCompactionsTest fails in trunk
Date Thu, 05 Dec 2013 23:52:36 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-6414?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13840702#comment-13840702
] 

Michael Shuler edited comment on CASSANDRA-6414 at 12/5/13 11:51 PM:
---------------------------------------------------------------------

(edit: added commit, since it's a one-liner)

I misunderstood the logback spam as the WARN entries from CASSANDRA-3578.  So ignoring all
the ERROR traces in the output (which I was marking bad above), just straight SUCESSFUL/FAILED
test gets me:

((f3dc188...)|BISECTING)mshuler@hana:~/git/cassandra$ git bisect bad
f3dc188e203b3db980ee81df05390968043cb601 is the first bad commit
commit f3dc188e203b3db980ee81df05390968043cb601
Author: Jonathan Ellis <jbellis@apache.org>
Date:   Fri Nov 22 17:33:07 2013 -0600

    move setting lastCompactedKey to before the return-if-nothing-added

:040000 040000 0c616036c3aac24861a0f43b21f5812faecf2c92 33fb0a0761168d605a224dc87522785afe8cb5a1
M      src
----
$ git show f3dc188e203b3db980ee81df05390968043cb601
commit f3dc188e203b3db980ee81df05390968043cb601 (HEAD, refs/bisect/bad)
Author: Jonathan Ellis <jbellis@apache.org>
Date:   Fri Nov 22 17:33:07 2013 -0600

    move setting lastCompactedKey to before the return-if-nothing-added

diff --git a/src/java/org/apache/cassandra/db/compaction/LeveledManifest.java b/src/java/org/apache/cassandra/
index 76f51d1..232d1f7 100644
--- a/src/java/org/apache/cassandra/db/compaction/LeveledManifest.java
+++ b/src/java/org/apache/cassandra/db/compaction/LeveledManifest.java
@@ -142,6 +142,7 @@ public class LeveledManifest
             int thisLevel = remove(sstable);
             minLevel = Math.min(minLevel, thisLevel);
         }
+        lastCompactedKeys[minLevel] = SSTableReader.sstableOrdering.max(added).last;
 
         // it's valid to do a remove w/o an add (e.g. on truncate)
         if (added.isEmpty())
@@ -152,7 +153,6 @@ public class LeveledManifest
 
         for (SSTableReader ssTableReader : added)
             add(ssTableReader);
-        lastCompactedKeys[minLevel] = SSTableReader.sstableOrdering.max(added).last;
     }
 
     public synchronized void repairOverlappingSSTables(int level)


was (Author: mshuler):
I misunderstood the logback spam as the WARN entries from CASSANDRA-3578.  So ignoring all
the ERROR traces in the output (which I was marking bad above), just straight SUCESSFUL/FAILED
test gets me:

((f3dc188...)|BISECTING)mshuler@hana:~/git/cassandra$ git bisect bad
f3dc188e203b3db980ee81df05390968043cb601 is the first bad commit
commit f3dc188e203b3db980ee81df05390968043cb601
Author: Jonathan Ellis <jbellis@apache.org>
Date:   Fri Nov 22 17:33:07 2013 -0600

    move setting lastCompactedKey to before the return-if-nothing-added

:040000 040000 0c616036c3aac24861a0f43b21f5812faecf2c92 33fb0a0761168d605a224dc87522785afe8cb5a1
M      src

> BlacklistingCompactionsTest fails in trunk
> ------------------------------------------
>
>                 Key: CASSANDRA-6414
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-6414
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Tests
>            Reporter: Jonathan Ellis
>            Assignee: Michael Shuler
>             Fix For: 2.1
>
>         Attachments: 6414_out.txt
>
>
> Passes in 2.0 HEAD.  Bisect should be relatively easy.



--
This message was sent by Atlassian JIRA
(v6.1#6144)

Mime
View raw message