[
https://issues.apache.org/jira/browse/CASSANDRA-4227?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13281030#comment-13281030
]
Jonathan Ellis commented on CASSANDRA-4227:
-------------------------------------------
So the situation is, I've upgraded to 1.2 so I'm supposed to have node IDs, but
one of the nodes starts slower than the others (for instance) so it hasn't
broadcast its nodeid to the rest of the cluster yet?
> StorageProxy throws NPEs for when there's no hostids for a target
> -----------------------------------------------------------------
>
> Key: CASSANDRA-4227
> URL: https://issues.apache.org/jira/browse/CASSANDRA-4227
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Reporter: Dave Brosius
> Priority: Trivial
> Fix For: 1.2
>
> Attachments: 4227_guard_against_npes_for_old_gossip_versions.diff
>
>
> On trunk...
> if there is no host id due to an old node, an info log is generated, but the
> code continues to use the null host id causing NPEs in decompose... Should
> this bypass this code, or perhaps can the plain ip address be used in this
> case? don't know.
> as follows...
> UUID hostId =
> StorageService.instance.getTokenMetadata().getHostId(target);
> if ((hostId == null) &&
> (Gossiper.instance.getVersion(target) < MessagingService.VERSION_12))
> logger.info("Unable to store hint for host with
> missing ID, {} (old node?)", target.toString());
> RowMutation hintedMutation =
> RowMutation.hintFor(mutation, ByteBuffer.wrap(UUIDGen.decompose(hostId)));
> hintedMutation.apply();
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
