Curtis Larsen wrote:
>
> Thanks very much to everyone who replied to my earlier question about
> possible ways to determine which DBs are causing (additional) XM
> activity.
>
> For those who asked, here's a little more background: The system is a
> 969KS/120 w/1GB RAM, running MPE/iX 5.5 (yeah, I know - long story), and
> for the most part has only two applications that use TurboImage DBs, and
> all main application have their own volumesets.  The abnormal XM
> activity I'm seeing is intermittent and strong.  Everything is fine,
> then suddenly XM is kicking in every 20-30 seconds, performing 50-80
> I/Os/sec during that time  (all based on Glance thus far).

Assuming you have XM checkpoint activity going on here, can you pinpoint
which disks are the busiest during the high IO load ? does it point to
a particular volumeset ?

One thing I would watch out for is if you have any groups (or HFS
directories) that have a huge number of files in them, and from
which files are frequently deleted or frequently inserted at the
beginning (alphabetically speaking).

HTH,
Goetz.

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *