[
https://issues.apache.org/jira/browse/ZOOKEEPER-4559?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17554215#comment-17554215
]
Chris Nauroth commented on ZOOKEEPER-4559:
------------------------------------------
Hello [~sergei.hanus] . As a starting point, I recommend reading through the
configuration options documented here:
[https://zookeeper.apache.org/doc/current/zookeeperAdmin.html#sc_advancedConfiguration]
In particular, snapCount, snapSizeLimitInKb, txnLogSizeLimitInKb and
preAllocSize all can influence the frequency of snapshots. Is it possible that
your deployment has reconfigured these settings in a way that triggers more
frequent snapshots?
> Zookeeper creates snapshots too often (every 5-10 minutes)
> ----------------------------------------------------------
>
> Key: ZOOKEEPER-4559
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-4559
> Project: ZooKeeper
> Issue Type: Bug
> Affects Versions: 3.6.1
> Reporter: Sergei Hanus
> Priority: Major
>
> I experience an issue, when zookeeper creates snapshots every 5-10 minutes.
> First, I suspected lots of transactions, which overload the ransaction log
> and rotate it - but, that's no the case. Each transaction log contains not
> more than 300-400 transactions (I opened one log file using zookeeper utility
> and at the end of this log there's a record "OF reached after 328 txns."
> Log dile size is also not big - less than 1 MB.
> Snapshot file is pretty big, though - it contains ~400k znodes. Also, there
> are about 2000 simultaneous sessions open to this zookeeper node. But that
> does not directly explain - why snapshots are rotated so fast.
>
> So, I'm wondering - what could be the reason for such frequent rotation?
>
> Thank you in advance.
>
>
>
>
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
