[
https://issues.apache.org/jira/browse/ZOOKEEPER-4229?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17314643#comment-17314643
]
Zhewei Hu commented on ZOOKEEPER-4229:
--------------------------------------
Hi Alexandre,
I think you could also set below to options to control the creation of
snapshots:
1. snapCount: The number of transactions recorded in the transaction log before
a snapshot can be taken, the default value is 100k.
2. snapSizeLimitInKb: The size limit of one snapshot file, the default value is
4G.
Alternatively, I think you could also set up a cronjob to clean up the
snapshots?
> Authorize less than one hour of snapshot retention
> --------------------------------------------------
>
> Key: ZOOKEEPER-4229
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-4229
> Project: ZooKeeper
> Issue Type: Improvement
> Reporter: Alexandre DUVAL
> Priority: Major
>
> When ZK is used with Pulsar, when a lot of data changes occur in ZK, it
> conducts ZK to snapshot a lot. I had a disk issue on creating 900 000
> producers and 1 500 000 consumers from scratch in 20 minutes.
> This causes the used disk to increase really fast. As each snapshot costs a
> 2.5GB disk.
> Auto purge is run each hour and this is the current minimum configuration
> authorized:
> {color:#859900}autopurge.purgeInterval{color}{color:#333333}=1{color}
> {color:#333333}Can we authorize define this value with minutes?{color}
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
