zookeeper-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Hadoop QA (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ZOOKEEPER-2101) Transaction larger than max buffer of jute makes zookeeper unavailable
Date Tue, 19 May 2015 04:15:00 GMT

    [ https://issues.apache.org/jira/browse/ZOOKEEPER-2101?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14549754#comment-14549754

Hadoop QA commented on ZOOKEEPER-2101:

+1 overall.  Here are the results of testing the latest attachment 
  against trunk revision 1679951.

    +1 @author.  The patch does not contain any @author tags.

    +1 tests included.  The patch appears to include 5 new or modified tests.

    +1 javadoc.  The javadoc tool did not generate any warning messages.

    +1 javac.  The applied patch does not increase the total number of javac compiler warnings.

    +1 findbugs.  The patch does not introduce any new Findbugs (version 2.0.3) warnings.

    +1 release audit.  The applied patch does not increase the total number of release audit

    +1 core tests.  The patch passed core unit tests.

    +1 contrib tests.  The patch passed contrib unit tests.

Test results: https://builds.apache.org/job/PreCommit-ZOOKEEPER-Build/2702//testReport/
Findbugs warnings: https://builds.apache.org/job/PreCommit-ZOOKEEPER-Build/2702//artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html
Console output: https://builds.apache.org/job/PreCommit-ZOOKEEPER-Build/2702//console

This message is automatically generated.

> Transaction larger than max buffer of jute makes zookeeper unavailable
> ----------------------------------------------------------------------
>                 Key: ZOOKEEPER-2101
>                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-2101
>             Project: ZooKeeper
>          Issue Type: Bug
>          Components: jute
>    Affects Versions: 3.4.4
>            Reporter: Liu Shaohui
>             Fix For: 3.5.2, 3.6.0
>         Attachments: ZOOKEEPER-2101-v1.diff, ZOOKEEPER-2101-v2.diff, ZOOKEEPER-2101-v3.diff,
ZOOKEEPER-2101-v4.diff, ZOOKEEPER-2101-v5.diff, test.diff
> *Problem*
> For multi operation, PrepRequestProcessor may produce a large transaction whose size
may be larger than the max buffer size of jute. There is check of buffer size in readBuffer
method  of BinaryInputArchive, but no check in writeBuffer method  of BinaryOutputArchive,
which will cause that 
> 1, Leader can sync transaction to txn log and send the large transaction to the followers,
but the followers failed to read the transaction and can't sync with leader.
> {code}
> 2015-01-04,12:42:26,474 WARN org.apache.zookeeper.server.quorum.Learner: [myid:2] Exception
when following the leader
> java.io.IOException: Unreasonable length = 2054758
>         at org.apache.jute.BinaryInputArchive.readBuffer(BinaryInputArchive.java:100)
>         at org.apache.zookeeper.server.quorum.QuorumPacket.deserialize(QuorumPacket.java:85)
>         at org.apache.jute.BinaryInputArchive.readRecord(BinaryInputArchive.java:108)
>         at org.apache.zookeeper.server.quorum.Learner.readPacket(Learner.java:152)
>         at org.apache.zookeeper.server.quorum.Follower.followLeader(Follower.java:85)
>         at org.apache.zookeeper.server.quorum.QuorumPeer.run(QuorumPeer.java:740)
> 2015-01-04,12:42:26,475 INFO org.apache.zookeeper.server.quorum.Learner: [myid:2] shutdown
> java.lang.Exception: shutdown Follower
>         at org.apache.zookeeper.server.quorum.Follower.shutdown(Follower.java:166)
>         at org.apache.zookeeper.server.quorum.QuorumPeer.run(QuorumPeer.java:744)
> {code}
> 2, The leader lose all followers, which trigger the leader election. The old leader will
become leader again for it has up-to-date data.
> {code}
> 2015-01-04,12:42:28,502 INFO org.apache.zookeeper.server.quorum.Leader: [myid:3] Shutting
> 2015-01-04,12:42:28,502 INFO org.apache.zookeeper.server.quorum.Leader: [myid:3] Shutdown
> java.lang.Exception: shutdown Leader! reason: Only 1 followers, need 2
>         at org.apache.zookeeper.server.quorum.Leader.shutdown(Leader.java:496)
>         at org.apache.zookeeper.server.quorum.Leader.lead(Leader.java:471)
>         at org.apache.zookeeper.server.quorum.QuorumPeer.run(QuorumPeer.java:753)
> {code}
> 3, The leader can not load the transaction from the txn log for the length of data is
larger than the max buffer of jute.
> {code}
> 2015-01-04,12:42:31,282 ERROR org.apache.zookeeper.server.quorum.QuorumPeer: [myid:3]
Unable to load database on disk
> java.io.IOException: Unreasonable length = 2054758
>         at org.apache.jute.BinaryInputArchive.readBuffer(BinaryInputArchive.java:100)
>         at org.apache.zookeeper.server.persistence.Util.readTxnBytes(Util.java:233)
>         at org.apache.zookeeper.server.persistence.FileTxnLog$FileTxnIterator.next(FileTxnLog.java:602)
>         at org.apache.zookeeper.server.persistence.FileTxnSnapLog.restore(FileTxnSnapLog.java:157)
>         at org.apache.zookeeper.server.ZKDatabase.loadDataBase(ZKDatabase.java:223)
>         at org.apache.zookeeper.server.quorum.QuorumPeer.loadDataBase(QuorumPeer.java:417)
>         at org.apache.zookeeper.server.quorum.QuorumPeer.getLastLoggedZxid(QuorumPeer.java:546)
>         at org.apache.zookeeper.server.quorum.FastLeaderElection.getInitLastLoggedZxid(FastLeaderElection.java:690)
>         at org.apache.zookeeper.server.quorum.FastLeaderElection.lookForLeader(FastLeaderElection.java:737)
>         at org.apache.zookeeper.server.quorum.QuorumPeer.run(QuorumPeer.java:716)
> {code}
> The zookeeper service will be unavailable until we enlarge the jute.maxbuffer and restart
zookeeper hbase cluster.
> *Solution*
> Add buffer size check in BinaryOutputArchive to avoid large transaction be written to
log and sent to followers.
> But I am not sure if there are side-effects of throwing an IOException in BinaryOutputArchive
 and RequestProcessors

This message was sent by Atlassian JIRA

View raw message