ADSM-L

Re: One more time: admin schedules?

1998-07-31 08:22:41
Subject: Re: One more time: admin schedules?
From: Hilton Tina <HiltonT AT TCE DOT COM>
Date: Fri, 31 Jul 1998 07:22:41 -0500
Pat,

I don't understand your comment "DRM probably won't catch the cron job".
I've used a script from day one to issue all these commands after the
backups have finished (v2 and v3) and everything works fine.  At least with
version 3 we have the wait option and don't have to query the log to see
when the process finishes.

Please explain why you think DRM won't know about the database and storage
pool backups.

Tina Hilton

> -----Original Message-----
> From: Pat Wilson [SMTP:paw AT NORTHSTAR.DARTMOUTH DOT EDU]
> Sent: Thursday, July 30, 1998 3:39 PM
> To:   ADSM-L AT VM.MARIST DOT EDU
> Subject:      One more time: admin schedules?
>
> As far as I can tell from reading the archives, the "best"
> answer for "how do I run a bunch of server commands which want to be
> sequaential" is to go completely outside ADSM and run a cron job that
> starts at a designated time and uses the -wait option on the commands.
> For example:
>  a file called "05am_tasks" containing
>
>         dsmadmc -id yyy -pa=xxx  expire inventory wait=yes
>         dsmadmc -id yyy -pa=xxx  backup stg backuppool copypool wait=yes
>         dsmadmc -id yyy -pa=xxx  backup stg autodlt_pool copypool wait=yes
>         dsmadmc -id yyy -pa=xxx  backup db
>
>  can't be scheduled as
>     def sch 05am_tasks t=a cmd="05am_tasks" startt=05:00
>  as would be intuitive; in order for things to work as expected, one would
> have to run 05am_tasks as a cron job.
>
> Am I missing something here?  This seems *really* inelegant, and DRM
> probably won't catch the cron job...
>
> It's been suggested that this might be doable as a client schedule via
> the "objects" argument, but that feels like a hack, and client schedules
> are bound to a domain (though what effect that would have on this case is
> unclear to me).
>
> Thoughts?
>
> Thanks.
>
>
> Pat Wilson
> paw AT dartmouth DOT edu
<Prev in Thread] Current Thread [Next in Thread>