[
https://issues.apache.org/jira/browse/HBASE-18294?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16305066#comment-16305066
]
ramkrishna.s.vasudevan commented on HBASE-18294:
------------------------------------------------
Again am not seeing the code. Seeing these discussions, I think the intent of
[~eshcar] I get it. Initially to avoid this confusion of offheap and onheap
tracking she tried to add a new param at the HTD level so that at global level
alone we ensure this new param is honored. But that was asking too much from
user. Hence she simplified it with the old way. Left global region server level
pressure unchanged and only track the region level with heap + data size. It
does not help offheap cases for sure because we clearly distinguished what is
heapSize and that heapSize was purely based on what java is going to occupy for
that given cell.
One question - since tracking data size is much better in region level, why
don't we change the global size to work with heap+dataSize || dataSize > 128 MB?
This will solve the blocking updates issue and at the same time works for
offheap cases also right? I t hink this what Anoop also asked in one of
yesterday's comment. Again am not a blocker just saying.
> Reduce global heap pressure: flush based on heap occupancy
> ----------------------------------------------------------
>
> Key: HBASE-18294
> URL: https://issues.apache.org/jira/browse/HBASE-18294
> Project: HBase
> Issue Type: Improvement
> Affects Versions: 3.0.0
> Reporter: Eshcar Hillel
> Assignee: Eshcar Hillel
> Fix For: 2.0.0-beta-2
>
> Attachments: HBASE-18294.01.patch, HBASE-18294.02.patch,
> HBASE-18294.03.patch, HBASE-18294.04.patch, HBASE-18294.05.patch,
> HBASE-18294.06.patch, HBASE-18294.07.patch, HBASE-18294.07.patch,
> HBASE-18294.08.patch, HBASE-18294.09.patch, HBASE-18294.10.patch,
> HBASE-18294.11.patch, HBASE-18294.11.patch, HBASE-18294.12.patch,
> HBASE-18294.13.patch
>
>
> A region is flushed if its memory component exceed a threshold (default size
> is 128MB).
> A flush policy decides whether to flush a store by comparing the size of the
> store to another threshold (that can be configured with
> hbase.hregion.percolumnfamilyflush.size.lower.bound).
> Currently the implementation (in both cases) compares the data size
> (key-value only) to the threshold where it should compare the heap size
> (which includes index size, and metadata).
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
