cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alex Petrov (JIRA)" <>
Subject [jira] [Comment Edited] (CASSANDRA-11922) BatchlogManagerTest is failing intermittently
Date Mon, 30 May 2016 12:45:12 GMT


Alex Petrov edited comment on CASSANDRA-11922 at 5/30/16 12:44 PM:

It seems that there's a race condition between the write to {{pause}} which sets {{isActive}}
to {{false}} and the check for {{isEnabled}} which reads this flag, which results into compactions
activated despite the fact they shouldn't.

I've ran the test with additional write lock 10 times and all 10 times it passed. Without
the same lock, it was failing almost all the time on my machine. Added a multiplexed test
to run {{BatchlogManagerTest}} 50 times, too:


was (Author: ifesdjeen):
|[trunk|] |[utest|]

> BatchlogManagerTest is failing intermittently
> ---------------------------------------------
>                 Key: CASSANDRA-11922
>                 URL:
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Alex Petrov
>            Assignee: Alex Petrov
> {{BatchlogManagerTest}} is failing sometimes with the following trace:
> {code}
>     [junit] Testcase: testLegacyReplay(org.apache.cassandra.batchlog.BatchlogManagerTest):
>     [junit] [BigTableReader(path='/Users/oleksandrpetrov/foss/java/cassandra/build/test/cassandra/data:0/system/batches-919a4bc57a333573b03e13fc3f68b465/mb-2-big-Data.db'),
>     [junit] junit.framework.AssertionFailedError: [BigTableReader(path='/Users/oleksandrpetrov/foss/java/cassandra/build/test/cassandra/data:0/system/batches-919a4bc57a333573b03e13fc3f68b465/mb-2-big-Data.db'),
>     [junit]     at org.apache.cassandra.db.ColumnFamilyStore.discardSSTables(
>     [junit]     at org.apache.cassandra.db.ColumnFamilyStore$
>     [junit]     at java.util.concurrent.Executors$
>     [junit]     at org.apache.cassandra.db.ColumnFamilyStore.runWithCompactionsDisabled(
>     [junit]     at org.apache.cassandra.db.ColumnFamilyStore.truncateBlocking(
>     [junit]     at org.apache.cassandra.batchlog.BatchlogManagerTest.setUp(
> {code}
> Sometimes there are running compactions despite the fact they should have been waited
upon before truncate. {{pause}} and {{resume}} are called on the {{CompactionStrategyManager}}

This message was sent by Atlassian JIRA

View raw message