[ 
https://issues.apache.org/jira/browse/ZOOKEEPER-1107?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13095792#comment-13095792
 ] 

Laxman commented on ZOOKEEPER-1107:
-----------------------------------

Thanks very much for your review and update Pat.

{quote}having the period specified in minutes rather than hours{quote}

1) In our case we have gone for this approach as disk is running out of space 
after long run say 30 days and even longer. Initially we thought of providing 
purge interval in days. But, considering the testability and your point of 
finer level of granularity only we have gone for hours.

2) Say, for 3 days interval, user would prefer to configure 72 (hours) rather 
4320 (minutes). Just trying to balance usability and granularity.

Please let me know your thoughts on this. If you feel we need more granularity 
for some other usecase, surely I can update the patch.

{quote}(my reading of TimerTask indicates that only a single purge operation 
would be running at any time, ie they will not run concurrently if I set the 
interval to be 1 minute and it takes more than a minute to purge, can you 
confirm?){quote}

Yes, verified the behavior of Java Timer. Timer will be running in one single 
thread and no two invocations will be running concurrently. Verified the same 
through debug points.

> automating log and snapshot cleaning
> ------------------------------------
>
>                 Key: ZOOKEEPER-1107
>                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-1107
>             Project: ZooKeeper
>          Issue Type: Improvement
>          Components: server
>    Affects Versions: 3.3.3
>            Reporter: Jun Rao
>            Assignee: Laxman
>             Fix For: 3.4.0
>
>         Attachments: ZOOKEEPER-1107.1.patch, ZOOKEEPER-1107.2.patch, 
> ZOOKEEPER-1107.3.patch, ZOOKEEPER-1107.4.patch, ZOOKEEPER-1107.5.patch, 
> ZOOKEEPER-1107.5.patch, ZOOKEEPER-1107.6.patch, ZOOKEEPER-1107.patch
>
>
> I like to have ZK itself manage the amount of snapshots and logs kept,  
> instead of relying on the PurgeTxnLog utility.

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Reply via email to