[
https://issues.apache.org/jira/browse/HBASE-14978?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15057783#comment-15057783
]
ramkrishna.s.vasudevan commented on HBASE-14978:
------------------------------------------------
bq.We won't get an OOME instead we'll kill the cache and be un-able to read
anything from disk
Just wanted to understand your concern here. In HBase branch-1.3 even if we
configure the BucketCache (L2) offheap every time we bring in the block onheap
before serving any read request. So in case of multi Gets there could be n
number of blocks being copied to the onheap and all those will have references
in the Result till the results are transferred to the client. So the heap can
grow upto 2GB as per your decription.
But in the latest trunk this copy to onheap will never happen so the Cells that
are created in the Results will only refer directly to the offheap area and as
you said we do the ref counting of these blocks in offheap.
The only thing that could happen now is that till all the Results could be
transferred to the client we may hold the block and so any such blocks cannot
be evicted till the reference is cleared.
> Don't allow Multi to retain too many blocks
> -------------------------------------------
>
> Key: HBASE-14978
> URL: https://issues.apache.org/jira/browse/HBASE-14978
> Project: HBase
> Issue Type: Improvement
> Affects Versions: 2.0.0, 1.2.0, 1.3.0
> Reporter: Elliott Clark
> Assignee: Elliott Clark
> Priority: Critical
> Attachments: HBASE-14978-v1.patch, HBASE-14978-v2.patch,
> HBASE-14978.patch
>
>
> Scans and Multi's have limits on the total size of cells that can be
> returned. However if those requests are not all pointing at the same blocks
> then the KeyValues can keep alive a lot more data than their size.
> Take the following example:
> A multi with a list of 10000 gets to a fat row. Each column being returned in
> in a different block. Each column is small 32 bytes or so.
> So the total cell size will be 32 * 10000 = ~320kb. However if each block is
> 128k then total retained heap size will be almost 2gigs.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
