[
https://issues.apache.org/jira/browse/HBASE-8329?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14272394#comment-14272394
]
zhangduo commented on HBASE-8329:
---------------------------------
{quote}
can we do with fewer config options? F.e. do need a separate control size? Can
that be the lower bound?
{quote}
You'are right. Will try to remove some of them :)
{quote}
compaction are in the end run per store. Should the speed decision be made on a
store by store basis, based on the number of store files for this store?
{quote}
As the comment above, we only have a small set of compaction threads which
means we can not do many compactions concurrently. We can not slow down the
compaction speed even if we only have 3 store files in store because there
maybe a store with 10 store files is waiting for compaction.
{quote}
should we take compaction min/max/blocking files into account? Maybe compact
faster when we're over the blocking file count.
{quote}
Yep, in this patch user can set a lower bound and a higher bound of throughput
limitation. Larger store files count means larger throughput limitation. Maybe
we could set to unlimited if we reach the blocking file count.
And I use a separated max store file config in DefaultThroughputController
because there is another meaning of "hbase.hstore.compaction.max" which is "max
store files count that we can compact in one time"(I'm not sure of it...).
{quote}
another way to configure is to specify or detect the speed on the NIC, and then
just limit compactions to a percentage of that. Would that be easier to
understand?
{quote}
Nice suggestion. I will try in next patch.
Thanks [~lhofhansl] ~
> Limit compaction speed
> ----------------------
>
> Key: HBASE-8329
> URL: https://issues.apache.org/jira/browse/HBASE-8329
> Project: HBase
> Issue Type: Improvement
> Components: Compaction
> Reporter: binlijin
> Assignee: zhangduo
> Fix For: 2.0.0, 1.1.0
>
> Attachments: HBASE-8329-10.patch, HBASE-8329-11.patch,
> HBASE-8329-12.patch, HBASE-8329-2-trunk.patch, HBASE-8329-3-trunk.patch,
> HBASE-8329-4-trunk.patch, HBASE-8329-5-trunk.patch, HBASE-8329-6-trunk.patch,
> HBASE-8329-7-trunk.patch, HBASE-8329-8-trunk.patch, HBASE-8329-9-trunk.patch,
> HBASE-8329-trunk.patch, HBASE-8329_13.patch
>
>
> There is no speed or resource limit for compaction，I think we should add this
> feature especially when request burst.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
