[
https://issues.apache.org/jira/browse/HBASE-13388?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14392628#comment-14392628
]
Vikas Vishwakarma commented on HBASE-13388:
-------------------------------------------
ZKUtil.getData can return null because of which the next debug log can hit
NullPointer at data.length
{noformat}
byte[] data = ZKUtil.getData(zkController.getWatcher(), path);
LOG.debug("start proc data length is " + data.length);
if (!ProtobufUtil.isPBMagicPrefix(data)) {
{noformat}
Null check for data is handled in ProtobufUtil.isPBMagicPrefix(data) which will
ultimately throw IllegalArgumentException, so moving the debug log to after
this check should avoid the Nullpointer. Does not have a functional impact but
is cleaner.
> Handling NullPointer in ZKProcedureMemberRpcs while getting ZNode data
> ----------------------------------------------------------------------
>
> Key: HBASE-13388
> URL: https://issues.apache.org/jira/browse/HBASE-13388
> Project: HBase
> Issue Type: Bug
> Affects Versions: 2.0.0
> Reporter: Vikas Vishwakarma
> Assignee: Vikas Vishwakarma
> Priority: Minor
> Fix For: 2.0.0
>
> Attachments: HBASE-13388.patch
>
>
> Handling a minor NullPointer in ZKProcedureMemberRpcs while getting ZNode
> data. Does not have any functional impact is just a little cleaner
> 2015-04-01 10:04:32,913 INFO [ver60020-EventThread]
> procedure.ZKProcedureMemberRpcs - Received procedure start children changed
> event: /hbase/online-snapshot/acquired
> 2015-04-01 10:04:32,913 DEBUG [ver60020-EventThread]
> procedure.ZKProcedureMemberRpcs - Looking for new procedures under
> znode:'/hbase/online-snapshot/acquired'
> 2015-04-01 10:04:32,916 DEBUG [ver60020-EventThread]
> procedure.ZKProcedureMemberRpcs - Found procedure znode:
> /hbase/online-snapshot/acquired/test13
> 2015-04-01 10:04:32,917 ERROR [ver60020-EventThread] zookeeper.ClientCnxn -
> Error while calling watcher
> java.lang.NullPointerException
> at
> org.apache.hadoop.hbase.procedure.ZKProcedureMemberRpcs.startNewSubprocedure(ZKProcedureMemberRpcs.java:203)
> at
> org.apache.hadoop.hbase.procedure.ZKProcedureMemberRpcs.waitForNewProcedures(ZKProcedureMemberRpcs.java:172)
> at
> org.apache.hadoop.hbase.procedure.ZKProcedureMemberRpcs.access$100(ZKProcedureMemberRpcs.java:55)
> at
> org.apache.hadoop.hbase.procedure.ZKProcedureMemberRpcs$1.nodeChildrenChanged(ZKProcedureMemberRpcs.java:107)
> at
> org.apache.hadoop.hbase.zookeeper.ZooKeeperWatcher.process(ZooKeeperWatcher.java:351)
> at
> org.apache.zookeeper.ClientCnxn$EventThread.processEvent(ClientCnxn.java:522)
> at
> org.apache.zookeeper.ClientCnxn$EventThread.run(ClientCnxn.java:498)
> 2015-04-01 10:04:32,917 INFO [ver60020-EventThread]
> procedure.ZKProcedureMemberRpcs - Received procedure start children changed
> event: /hbase/online-snapshot/acquired
> 2015-04-01 10:04:32,917 DEBUG [ver60020-EventThread]
> procedure.ZKProcedureMemberRpcs - Looking for new procedures under
> znode:'/hbase/online-snapshot/acquired'
> 2015-04-01 10:04:32,918 INFO [ver60020-EventThread]
> procedure.ZKProcedureMemberRpcs - Received procedure abort children changed
> event: /hbase/online-snapshot/abort
> 2015-04-01 10:04:32,918 DEBUG [ver60020-EventThread]
> procedure.ZKProcedureMemberRpcs - Checking for aborted procedures on node:
> '/hbase/online-snapshot/abort'
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
