[
https://issues.apache.org/jira/browse/ZOOKEEPER-2430?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15311297#comment-15311297
]
Patrick Hunt commented on ZOOKEEPER-2430:
-----------------------------------------
I'm afraid i wasn't clear enough in my previous comment. This is what I meant
to highlight:
1) you were removing the max packet length, which is a bad idea. We need to
validate that the packet length doesn't exceed the max
2) similarly we need to ensure that the length is not < 0 - iirc we added this
a while back to handle the case with a malformed packet, say someone runs a
security scanner which sends in "junk". We don't want that to crash the process.
You might take a look at "git blame" or svn history on that file/lines in order
to get more insight.
> Remove jute.maxbuffer limit packetLen in client side
> ----------------------------------------------------
>
> Key: ZOOKEEPER-2430
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-2430
> Project: ZooKeeper
> Issue Type: Improvement
> Reporter: Yong Zhang
> Assignee: Yong Zhang
> Attachments: ZOOKEEPER-2430.001.patch, ZOOKEEPER-2430.002.patch
>
>
> jute.maxbuffer can be configured in both client and server side, if we try to
> getChildren
> from a parent znode with large number of children znode, client may failed to
> get because
> of packetLen is more than jute.maxbuffer configured.
> even if we can change the value in java system property, but we have to
> restart the application,
> and more all data has been in zookeeper client, check the length/size is
> unnecessary.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
