[ATrpms-users] mythfilldatabase increased run time
Brian Long
briandlong at gmail.com
Fri Jan 20 21:16:07 CET 2012
On Fri, Jan 20, 2012 at 1:51 PM, John Welch <jrw3319 at gmail.com> wrote:
> Great, glad it worked out for you. My process *only* takes 20-25 minutes
> to run, so maybe I'm not as bad off as I thought. I don't know if I just
> became more aware of this process running after reading about it on the
> myth mailing list, but I just don't remember it hammering the system like
> it does now prior to the upgrade to F16.
>
> I never would have known about the change either if I hadn't seen a couple
> of threads about it over on myth list.
>
> Good luck!
>
>
Not sure if you guys saw the thread on mythtv-users about mythfilldatabase
+ MySQL hitting the disk hard. MySQL's config usually has "tmpdir = /tmp"
which is not tmpfs on many hosts (it's part of /) and mythfilldatabase
could create up to 500M of temp tables for each run. This was noticed when
someone was trying to extend the life of their SSD boot disk.
They changed /etc/my.cnf and set "tmpdir = /dev/shm" or remounted /tmp as
tmpfs and it sped things up plus reduced wear on the SSD. Other folks
suggested increasing the query cache, etc, but setting tmpdir to somewhere
tmpfs was the biggest win.
/Brian/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.atrpms.net/pipermail/atrpms-users/attachments/20120120/751d8e3e/attachment.html>
More information about the atrpms-users
mailing list