[ 
https://issues.apache.org/jira/browse/HBASE-5930?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Devaraj Das updated HBASE-5930:
-------------------------------

    Attachment: 5930-2.3.patch

Attaching another rev of the patch. Changed the policies for flushing some. 
Here is a summary:
0. If the last flush happened in the last 60 minutes don't request another 
flush now for the region in question.
1. For a region, a flush is requested if the following is true:
   1.1. If the region is the meta region, and if there is at least one edit 
after the last flush request another flush now.
   1.2. If the region is not one of the meta regions, a flush is requested,
     1.2.1. if the last edit is more than 1 minute old, or,
     1.2.2. if the region wasn't flushed in the last two flush cycles (this was 
Enis's suggestion)

I also added a new API in FlushRequester to do flush requests with a delay. 
That is used by the chore (earlier I used to throttle the flushes by doing a 
sleep in the chore and that seemed a little odd).

Thoughts?
                
> Periodically flush the Memstore?
> --------------------------------
>
>                 Key: HBASE-5930
>                 URL: https://issues.apache.org/jira/browse/HBASE-5930
>             Project: HBase
>          Issue Type: Improvement
>            Reporter: Lars Hofhansl
>            Assignee: Devaraj Das
>            Priority: Minor
>             Fix For: 0.96.0
>
>         Attachments: 5930-1.patch, 5930-2.1.patch, 5930-2.2.patch, 
> 5930-2.3.patch, 5930-wip.patch
>
>
> A colleague of mine ran into an interesting issue.
> He inserted some data with the WAL disabled, which happened to fit in the 
> aggregate Memstores memory.
> Two weeks later he a had problem with the HDFS cluster, which caused the 
> region servers to abort. He found that his data was lost. Looking at the log 
> we found that the Memstores were not flushed at all during these two weeks.
> Should we have an option to flush memstores periodically. There are obvious 
> downsides to this, like many small storefiles, etc.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Reply via email to