RE: Backup performance issue.

[Evan Harris <spanner@xxxxxxxxxx>]

Rob

I don't think the performance data will tell you too much - I am guessing
that it will show you pretty much the same thing on each backup run if the
procedure is the same. I would be looking to see where there are variations
in the procedure rather than looking at performance data: apart from maybe
the start time of the backup each night, if you haven't seen it already it
probably isn't obvious in the performance data.

The point regarding jobs/conflicting jobs was more directed at other things
that are run periodically that might influence the backup rather than the
backup itself. Think of things like:

- are there any Domino housekeeping or agents that run periodically ?
- isn't there a Domino purge or something (cleanup ?) that has to be run
regularly ?
- What reporting processes or data extracts get run regularly ?
- Do any of these things affect shutdown time for Domino

The reason I mentioned BRMS is that I was wondering whether BRMS
maintenance or save cycles were causing variations in  your backup. If you
are not using BRMS then this can probably be discounted (same with Tivoli).

Keep at it :)

Regards
Evan Harris

> This is a multipart message in MIME format.
> --
> [ Picked text/plain from multipart/alternative ]
> All excellent ideas.
>
> The number of objects stays pretty consistent.  We do not find ourselves
> creating/deleting large numbers of Notes databases.
>
> It is not a save changed objects.  A bulk of these objects would get
> changed.
>
> I wondered about that pattern with conflicting jobs.  I will attempt to
> log these occurrences better.  However, when I first submitted this email
> I put in a lot of performance data and none of it seemed to indicate such.
>
> Yes we have BRMS installed.  I wouldn't exactly say we use it.  I am not
> sure that it would fit in with out current backup strategy.  And I don't
> see how adapting to the BRMS strategy would help us.
>
> Yes we also have Tivoli.
>
> I am not sure if any of these might be the culprit.  I would hope that
> they would appear in the mentioned performance data though.
>
> Rob Berendt