I agree that something to assist in the monitoring would be nice, but
you find very few IBM products that have that.  An exception in the VM
area is SQL Master for SQL/DS or DB2/VM.  Programmable Operator filters
or WAKEUP based code would be the most likely VM solution.  If anyone
has any to share, I wouldn't be too proud to accept it.

The missing pieces that you refer to are indeed a problem.  It would be
helpful to have more of the client log recorded at the server and also
to have more information about who initiated commands.

Jim Bohnsack
Raytheon E-Systems

>----------
>From:  John Schneider[SMTP:jdschn AT ibm DOT net]
>Sent:  Tuesday, March 31, 1998 10:54PM
>To:    ADSM-L AT VM.MARIST DOT EDU
>Subject:       Lack of monitoring and logging in ADSM?
>
>Greetings,
>        Is anyone else unsatisfied with the amount of logging and
>monitoring features in ADSM?  It seems like almost everything is
>roll-your-own.  I refer to ADSM v2.1 server under AIX for these
>questions.
>        Basic things like:
>
>1) A server log.  Yes, it is possible to write a script to set up a
>background process from "dsmadmc -consolemode -outfile=..." but why
>do you have to set this up yourself?  And the file is close to
>useless sometimes.  The file isn't even time-stamped!  Does anyone
>know a way to timestamp the file?  And a lot of information that
>you need is not logged.  For example, if someone starts a process,
>like "checkin libvol", or "audit library", the log will show that
>the process started, but not who started it?  Why not state that?
>
>2) Scheduler logs?  Yes, it does keep a log of each client, but on
>the client, not the server.  Am I supposed to log on to each client
>every day to check the log, and verify that the backup ran?  No
>message is produced anywhere if a scheduler dies, except on the
>client itself in the scheduler?
>
>3) I would really like some sort of reporting when things go wrong.
>Last week we went several days in a row with our ADSM database
>backup failing, and several systems' scheduler processes not
>running.  But there is no alarming or error reports of any kind
>that I can see.  Why not?  Sure, the stuff is buried in the logs,
>but who has time to wade through 1000s of lines of output every
>day, just looking for anything unusual.
>
>4) I would like some sort of warning when the database or log is
>close to full, so things don't come to a crashing halt without
>any warning.
>
>Have any of you thought of clever ways to:
>
>1) Timestamp the server log?
>
>2) Generate an alarm if a scheduler process isn't running?
>
>3) Warn a sys admin about other problem conditions?
>
>I know it is possible to monitor all this stuff by hand if you have
>lots of time on your hands, but in reality ADSM maintenance cannot
>really be anybody's full time job.  ADSM should provide you with
>tools.  Is anybody in ADSM development listening?  Other backup
>products (like Legato Networker) make it easy to look at a single
>summary log, and tell at a glance if all the backups ran last
>night, and why not for any that didn't.
>
>TIA for any replies,
>
>John Schneider
>
>***********************************************************************
>* John D. Schneider     * Email: jdschn AT ibm DOT net * Phone: 314-349-4556 *
>* Lowery Systems, Inc.  ***********************************************
>* 275 Axminister        * Disclaimer: Opinions expressed here are mine*
>* Fenton, MO 63026      * and mine alone.  My company is off the hook.*
>***********************************************************************
>