* dm-thin COMMIT_PERIOD problem
@ 2014-11-01 14:40 lilofile
2014-11-01 18:45 ` Mike Snitzer
0 siblings, 1 reply; 2+ messages in thread
From: lilofile @ 2014-11-01 14:40 UTC (permalink / raw)
To: dm-devel
In dm-thin.c,COMMIT_PERIOD is set HZ, I think this value has a impact on system impact,too frequent commit will reduce the write performance of the system？
other method which it is submitted by the total size of dirty data is feasible？
--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: dm-thin COMMIT_PERIOD problem
2014-11-01 14:40 dm-thin COMMIT_PERIOD problem lilofile
@ 2014-11-01 18:45 ` Mike Snitzer
0 siblings, 0 replies; 2+ messages in thread
From: Mike Snitzer @ 2014-11-01 18:45 UTC (permalink / raw)
To: lilofile; +Cc: dm-devel
On Sat, Nov 01 2014 at 10:40am -0400,
lilofile <lilofile@aliyun.com> wrote:
> In dm-thin.c,COMMIT_PERIOD is set HZ, I think this value has a impact on system impact,too frequent commit will reduce the write performance of the system？
> other method which it is submitted by the total size of dirty data is feasible？
It only issues IO if changes have occured. It isn't writing to disk
every second unless changes need to get to disk anyway.
(There was a bug where it _did_ write the metadata superblock every
second even when it wasn't needed.. since fixed with this commit:
http://git.kernel.org/linus/4d1662a30 )
--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2014-11-01 18:45 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-11-01 14:40 dm-thin COMMIT_PERIOD problem lilofile
2014-11-01 18:45 ` Mike Snitzer
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.