[
https://issues.apache.org/jira/browse/ZOOKEEPER-4332?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17372485#comment-17372485
]
Damien Diederen commented on ZOOKEEPER-4332:
--------------------------------------------
Hi Emil,
This is a long-standing issue, indeed, and one which is unfortunately still
relevant.
Suggestions which have been suggested include:
* [Rejecting node
creations|https://issues.apache.org/jira/browse/ZOOKEEPER-1162?focusedCommentId=13091100&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-13091100]
which would cause the {{GetChildren}} payload to overflow {{jute.maxbuffer}}
\(which is somewhat problematic, as the database does not specify the minimum
{{jute.maxbuffer}} needed at runtime);
* [Introducing a paginated
version|https://issues.apache.org/jira/browse/ZOOKEEPER-2260] of
{{GetChildren}}. I had missed that existing patch before; it would be
interesting to forward-port it to 3.7\+\!
You also noted:
bq. I am attaching some screenshots of the commands and their results. What's
interesting the numbers from getAllChildrenNumber and stat \(numChildren)
commands don't match.
Note that {{getAllChildrenNumber}} is a *recursive* computation, whereas
{{Stat.numChildren}} tracks the number of *direct* children—so I would expect
the former to be larger than the latter if some of the nodes have children.
> Cannot access children of znode that owns too many znodes
> ---------------------------------------------------------
>
> Key: ZOOKEEPER-4332
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-4332
> Project: ZooKeeper
> Issue Type: Bug
> Components: server
> Affects Versions: 3.6.1
> Reporter: Emil Kleszcz
> Priority: Critical
> Labels: zookeeper
> Attachments: Screen Shot 2021-06-30 at 16.52.17.png, Screen Shot
> 2021-06-30 at 16.52.42.png, Screen Shot 2021-06-30 at 16.53.04.png
>
>
> We experience problems with performing any operation (deleteall, get etc.) on
> a znode that has too many child nodes. In our case, it's above 200k. At the
> same time jute.max.buffer is 4194304. Increasing it by a few factors doesn't
> help. This should be either solved by limiting the number of direct znodes
> allowed by a parameter or by adding a hard limit by default.
> I am attaching some screenshots of the commands and their results. What's
> interesting the numbers from getAllChildrenNumber and stat (numChildren)
> commands don't match.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
