[
https://issues.apache.org/jira/browse/CASSANDRA-10587?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15014345#comment-15014345
]
Ariel Weisberg commented on CASSANDRA-10587:
--------------------------------------------
In your reproduction steps are you not doing any upgrades? It doesn't reproduce
if you run the repair immediately? What role does OpsCenter play in
reproducing? If you run the repair manually using nodetool does it reproduce?
I tried it out with CCM. Created a 2 node Cassandra 2.2.2 cluster, ran stress
for a few seconds, ran repair on both nodes, and then tried
{{sstablemetadata}}. No errors, but that isn't surprising since I probably
haven't captured what you are doing.
> sstablemetadata NPE on cassandra 2.2
> ------------------------------------
>
> Key: CASSANDRA-10587
> URL: https://issues.apache.org/jira/browse/CASSANDRA-10587
> Project: Cassandra
> Issue Type: Bug
> Components: Tools
> Reporter: Tiago Batista
> Assignee: Ariel Weisberg
> Fix For: 2.2.x, 3.x
>
>
> I have recently upgraded my cassandra cluster to 2.2, currently running
> 2.2.3. After running the first repair, cassandra renames the sstables to the
> new naming schema that does not contain the keyspace name.
> This causes sstablemetadata to fail with the following stack trace:
> {noformat}
> Exception in thread "main" java.lang.NullPointerException
> at
> org.apache.cassandra.io.sstable.Descriptor.fromFilename(Descriptor.java:275)
> at
> org.apache.cassandra.io.sstable.Descriptor.fromFilename(Descriptor.java:172)
> at
> org.apache.cassandra.tools.SSTableMetadataViewer.main(SSTableMetadataViewer.java:52)
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
