[
https://issues.apache.org/jira/browse/HBASE-14969?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15063370#comment-15063370
]
Yu Li commented on HBASE-14969:
-------------------------------
Thanks [~Apache9] for the follow up.
bq. do you have some perf results already?
I tried to load 100GB data into a 3 nodes dev cluster using PE (--presplit=100
--rows=10485760 randomWrite 10), and could observe less slow sync (threshold
100ms) in the RS log, but didn't see much difference on the latency number in
PE result (similar, no downgrade, but no obvious enhancement also). I guess
it's not easy to reproduce the online situation and will retry. Will paste some
detailed perf number here later.
> Add throughput controller for flush
> -----------------------------------
>
> Key: HBASE-14969
> URL: https://issues.apache.org/jira/browse/HBASE-14969
> Project: HBase
> Issue Type: Improvement
> Affects Versions: 2.0.0
> Reporter: Yu Li
> Assignee: Yu Li
> Attachments: HBASE-14969.patch
>
>
> In HBASE-8329 we added a throughput controller for compaction, to avoid spike
> caused by huge IO pressure like network/disk overflow. However, even with
> this control on, we are still observing disk utils near 100%, and by analysis
> we think this is caused by flush, especially when we increase the setting of
> {{hbase.hstore.flusher.count}}
> In this JIRA, we propose to add throughput control feature for flush, as a
> supplement of HBASE-8329 to better control IO pressure.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
