[
https://issues.apache.org/jira/browse/ZOOKEEPER-4397?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17432217#comment-17432217
]
Damien Diederen commented on ZOOKEEPER-4397:
--------------------------------------------
Hi [~ryan.ren],
Do you have clients which create large numbers (i.e., are leaking) ephemeral
nodes? If so, this may be a duplicate of ZOOKEEPER-4306.
HTH, -D
> Zookeeper crashes: Unable to load database on disk java.io.IOException:
> Unreasonable length
> -------------------------------------------------------------------------------------------
>
> Key: ZOOKEEPER-4397
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-4397
> Project: ZooKeeper
> Issue Type: Bug
> Components: jute
> Affects Versions: 3.6.3
> Environment: Linux
> OpenJDK 1.8
> Reporter: Ryan
> Priority: Major
> Attachments: Error_snapshot.jpg
>
>
> After running for a while, the entire cluster (3 zookeeper) crash suddenly
> ERROR－［main:QuorumPeer@1148］- Unable to load database on disk
> java.io.IOException: Unreasonable length = 3015236
> at
> org.apache.jute.BinaryInputArchive.checkLength(BinaryInputArchive.java:166)
>
> at.org.apache.jute.BinaryInputArchive.readBuffer(BinaryInputArchive.java:127)
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
