[
https://issues.apache.org/jira/browse/HBASE-17738?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16083756#comment-16083756
]
Anoop Sam John commented on HBASE-17738:
----------------------------------------
The math in calculating the number of BBs per thread seems some thing wrong.
We can do this way. Say for eg: if we have to create 101 BBs and 10 threads,
lets give 10 BBs per thread and the last one get one extra means 11.
If we have to create 98 BBs and 10 threads, lets give 10 BBs per thread and the
last one get 2 less ie. 8 BBs.
You can round off the bufferCount/threadsCount.
Lets make the Callable return a BB[] so that u can avoid the start and end
index math and changing a shared variable from diff threads.
> BucketCache startup is slow
> ---------------------------
>
> Key: HBASE-17738
> URL: https://issues.apache.org/jira/browse/HBASE-17738
> Project: HBase
> Issue Type: Sub-task
> Components: BucketCache
> Affects Versions: 2.0.0
> Reporter: stack
> Assignee: ramkrishna.s.vasudevan
> Fix For: 2.0.0
>
> Attachments: HBASE-17738_2.patch, HBASE-17738_2.patch,
> HBASE-17738_3.patch, HBASE-17738_4.patch, HBASE-17738_5_withoutUnsafe.patch,
> HBASE-17738_6_withoutUnsafe.patch, HBASE-17738.patch
>
>
> If you set bucketcache size at 64G say and then start hbase, it takes a long
> time. Can we do the allocations in parallel and not inline with the server
> startup?
> Related, prefetching on a bucketcache is slow. Speed it up.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
