<br><br><div class="gmail_quote">On Mon, Apr 18, 2011 at 11:02 AM, Rob Smith <span dir="ltr"><<a href="mailto:kormoc@mythtv.org">kormoc@mythtv.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im">On Mon, Apr 18, 2011 at 6:29 AM, Mark Lord <<a href="mailto:mythtv@rtr.ca">mythtv@rtr.ca</a>> wrote:<br>
> But in the end, it just might be mysql doing an excessive number of "fsync()"<br>
> or "fdatasync()" calls. Databases like to do things that way, but it does<br>
> mean they run very slowly when updating stuff<br>
<br>
</div>I wouldn't call myisam's usage of fsync excessive... I'd call it<br>
majorly lacking.<br>
<div><div></div><div class="h5">_______________________________________________<br>
mythtv-users mailing list<br>
<a href="mailto:mythtv-users@mythtv.org">mythtv-users@mythtv.org</a><br>
<a href="http://www.mythtv.org/mailman/listinfo/mythtv-users" target="_blank">http://www.mythtv.org/mailman/listinfo/mythtv-users</a><br>
</div></div></blockquote></div><br><div>I've been researching this problem and appear to have gotten better performance by adding the barrier=0 option to the line in /etc/fstab that mounts the ext4 volume containing the mythconverg database. My understanding is that you need to ensure that you don't have power blips (use a UPS) when disabling write barriers. I'm also using LVM. </div>
<div><br></div><div>My per/schedule day process times in mythfilldatabase dropped from 10+ minutes to 2 minutes. With the --dd-grab-all option, total time was 17.5 minutes. Before the change it was 1-2 hours if it didn't timeout. IO wait times have gone from 20% to 5.5% when running "iostat 5". My database type is myisam.</div>
<div><br></div><div>--</div><div>David McMillan </div>