Richard,

Thanks. I am monitoring TSM operations from a Perl script someone made a
long time ago (before I got here). What the script does is basically dump
the activity log to a file and then 'does stuff with it'. The stuff it does,
it is not doing very well. For some reason and for TDP sessions only, it is
now seeing lines with the word 'Failed' in it and tells me a backup has
failed. What it actually sees is 'Number of objects failed: 0'. Thus in my
overview I'm getting a lot of nodes that are 'Not OK', but in fact are.

I was trying to figure a way to rebuild this script, outputting the same
info but correctly, this time from the tsm command line with the correct
info... It is not all that easy, it appears.

Cheers,

Rick

On 10/18/07, Richard Sims <rbs AT bu DOT edu> wrote:
>
> Rick -
>
> I think you're perceiving the FAILED column of the Activity Summary
> table to be a kind of boolean, or condition code.  It is actually the
> number of objects which the (B/A) client reported to have failed in
> its attempt to back them up, as can occur when a transient file in a
> file system disappears before TSM gets to reading it for the backup.
> I don't run TDPs, but I don't believe that this "failed objects
> count" pertains to them, given that they operate in TSM API
> "streaming" mode.
>
> For my money, I leave the assessment of backups success to client
> administrators, who own and are responsible for the data being backed
> up, rather than try to "read tea leaves" on the TSM server.
>
>    Richard Sims    at Boston University
>