[
https://issues.apache.org/jira/browse/ZOOKEEPER-1813?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13823356#comment-13823356
]
Vinay commented on ZOOKEEPER-1813:
----------------------------------
Scenario :
1. There are so many nodes under /test/subdir1/subdir2
2. One node is created under /test/subdir2/subdir. here the zxid is say X;
3. Now snapshotting with zxid X. started due to reaching snapCount threshold.
4. Now before snapshot was still writing nodes under /test/subdir1/subdir2,
children nodes under /test/subdir2/subdir are created and deleted. (
/test/subdir2/subdir/file)
/test/subdir2/subdir also deleted.
5. When snapshot comes to write /test/subdir2 it will have no childs, same is
stored in snapshot.
6. Now if the Zk restarted while loading data tree, after loading snapshot,
there will be txn for create /test/subdir2/subdir/file, but there wont be
parent node, because it was already missed from snapshot. And restart fails.
> Zookeeper restart fails due to missing node from snapshot
> ---------------------------------------------------------
>
> Key: ZOOKEEPER-1813
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-1813
> Project: ZooKeeper
> Issue Type: Bug
> Affects Versions: 3.4.5, 3.5.0
> Reporter: Vinay
> Priority: Blocker
>
> Due to following exception Zookeeper restart is failing
> {noformat}java.io.IOException: Failed to process transaction type: 1 error:
> KeeperErrorCode = NoNode for /test/subdir2/subdir2/subdir
> at
> org.apache.zookeeper.server.persistence.FileTxnSnapLog.restore(FileTxnSnapLog.java:183)
> at
> org.apache.zookeeper.server.ZKDatabase.loadDataBase(ZKDatabase.java:222)
> at
> org.apache.zookeeper.server.ZooKeeperServer.loadData(ZooKeeperServer.java:255)
> at
> org.apache.zookeeper.server.ZooKeeperServer.startdata(ZooKeeperServer.java:380)
> at
> org.apache.zookeeper.server.NIOServerCnxnFactory.startup(NIOServerCnxnFactory.java:748)
> at
> org.apache.zookeeper.server.ZooKeeperServerMain.runFromConfig(ZooKeeperServerMain.java:111)
> at
> org.apache.zookeeper.server.ZooKeeperServerMain.initializeAndRun(ZooKeeperServerMain.java:90)
> at
> org.apache.zookeeper.server.ZooKeeperServerMainTest$2.run(ZooKeeperServerMainTest.java:218)
> Caused by: org.apache.zookeeper.KeeperException$NoNodeException:
> KeeperErrorCode = NoNode for /test/subdir2/subdir2/subdir
> at
> org.apache.zookeeper.server.persistence.FileTxnSnapLog.processTransaction(FileTxnSnapLog.java:268)
> at
> org.apache.zookeeper.server.persistence.FileTxnSnapLog.restore(FileTxnSnapLog.java:181)
> ... 7 more{noformat}
--
This message was sent by Atlassian JIRA
(v6.1#6144)