ADSM-L

Re: query archive /*

2000-12-06 23:21:43
Subject: Re: query archive /*
From: Joel Fuhrman <joelf AT CAC.WASHINGTON DOT EDU>
Date: Wed, 6 Dec 2000 20:22:06 -0800
Reversing the options had no affect.

The problem is that the "client" can not find all occurrences of an archive
unless they remember every filesystem that ever existed.

The dba's dump the database to a file and then archive it, re-using the
description for a previous archives of this database.  If the dump/archive
is done using different file systems due to space constraints (or it was
unmounted and it went to root), the client has to remember both filesystem
names.  If it happens to be a 3 year archive and one of the filesystems is
renamed or removed, the client is going to have to have a good memory and
g-d forbid they should quit.

It seems to me that there should be some syntax that the "client" can use to
find all their archive.  A simple 'q arch' and 'q arch desc="some desc"'
would do nicely.

On Wed, 6 Dec 2000, Richard Sims wrote:

> > TSM 3.7.2.0 client on AIX; TSM 3.7.3.8 server on AIX
> >
> > The command
> >    dsmc q arch '/*' -subdir=yes
> >
> > shows the archives for all but one filesystem.  To see the missing
> > filesystem requires
> >    dsmc q arch '/advscratch2/*' -subdir=yes
> >
> > Any thoughts as to why the /advscratch2 archive would not be listed by the
> > first command?
>
> The command documentation specifies coding options before filespace
> names, which may make a difference in your case.
> Always observe invocation syntax regarding the order of operands:
> some applications simply take and process each operand as it
> occurs in the command line, rather than going through them all
> first before figuring out what to do.  Keep in mind also that
> applications change from release to release, and so deviations
> from syntax that might work in one release might not in another.
>    Richard Sims, BU
>
<Prev in Thread] Current Thread [Next in Thread>