[
https://issues.apache.org/jira/browse/HBASE-17215?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15949419#comment-15949419
]
huaxiang sun commented on HBASE-17215:
--------------------------------------
Thanks [~carp84], it is definitely fine to go with a follow-up to address the
shared queue.
{quote}
I'm very much interested in this "some reason" (smile). Ever dig into it and
mind share if so? Thanks. huaxiang sun
{quote}
We have some rough idea about it. First, hdfs operation is slow, it takes 120
ms ~ 140 ms to delete a file from hdfs. I think this Jira will help to address
this issue (speed up + delete large files to free up the diskspace). Another
issue is related with SnapshotHFileCleaner, refresh snapshot file cache is
called very frequently, still looking into code to see if something can be done
there.
> Separate small/large file delete threads in HFileCleaner to accelerate
> archived hfile cleanup speed
> ---------------------------------------------------------------------------------------------------
>
> Key: HBASE-17215
> URL: https://issues.apache.org/jira/browse/HBASE-17215
> Project: HBase
> Issue Type: Improvement
> Reporter: Yu Li
> Assignee: Yu Li
> Attachments: HBASE-17215.patch, HBASE-17215.v2.patch
>
>
> When using PCIe-SSD the flush speed will be really quick, and although we
> have per CF flush, we still have the
> {{hbase.regionserver.optionalcacheflushinterval}} setting and some other
> mechanism to avoid data kept in memory for too long to flush small hfiles. In
> our online environment we found the single thread cleaner kept cleaning
> earlier flushed small files while large files got no chance, which caused
> disk full then many other problems.
> Deleting hfiles in parallel with too many threads will also increase the
> workload of namenode, so here we propose to separate large/small hfile
> cleaner threads just like we do for compaction, and it turned out to work
> well in our cluster.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
