[jira] [Commented] (CASSANDRA-8800) Cassandra fails to start with OOM after disk space problem occurred
[ https://issues.apache.org/jira/browse/CASSANDRA-8800?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14320206#comment-14320206 ] Maxim Podkolzine commented on CASSANDRA-8800: - Cassandra 2.1.1. We got this once. The DB is no longer available to check it unfortunately. > Cassandra fails to start with OOM after disk space problem occurred > --- > > Key: CASSANDRA-8800 > URL: https://issues.apache.org/jira/browse/CASSANDRA-8800 > Project: Cassandra > Issue Type: Bug >Reporter: Maxim Podkolzine > > It one point the Cassandra failed with > org.apache.cassandra.io.FSWriteError: java.io.IOException: No space left on > device > Which seems pretty normal situation. But after the disk space was increased, > Cassandra refused to start up again. The memory snapshot shows a big > ArrayList of org.apache.cassandra.io.sstable.SSTableScanner is allocated, > holding around 2.8Gb. > It looks like the DB is corrupted, but we're not entirely sure. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-8800) Cassandra fails to start with OOM after disk space problem occurred
[ https://issues.apache.org/jira/browse/CASSANDRA-8800?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14320204#comment-14320204 ] Philip Thompson commented on CASSANDRA-8800: What Cassandra version are you running? Nodetool scrub should fix corrupt sstables if you suspect that is the issue. > Cassandra fails to start with OOM after disk space problem occurred > --- > > Key: CASSANDRA-8800 > URL: https://issues.apache.org/jira/browse/CASSANDRA-8800 > Project: Cassandra > Issue Type: Bug >Reporter: Maxim Podkolzine > > It one point the Cassandra failed with > org.apache.cassandra.io.FSWriteError: java.io.IOException: No space left on > device > Which seems pretty normal situation. But after the disk space was increased, > Cassandra refused to start up again. The memory snapshot shows a big > ArrayList of org.apache.cassandra.io.sstable.SSTableScanner is allocated, > holding around 2.8Gb. > It looks like the DB is corrupted, but we're not entirely sure. -- This message was sent by Atlassian JIRA (v6.3.4#6332)