I'm not prepared to post our scripts yet but will tell you some of the
things that we run daily:

Report on failed client schedules
q event * * enddate=today begintime=NOW-23:59 endtime=NOW f=d
        - filter and report on anything other than "completed"

Report on Client errors (mostly failed files)
select MESSAGE,DOMAINNAME,NODENAME,DATE_TIME from actlog where -
(date_time>current_timestamp - 1 days and originator='CLIENT' AND
SEVERITY='E') -
ORDER BY DOMAINNAME,NODENAME

Report on file spaces not backed up in over x days
q fi f=d
        - filter and report on specific nodes and days, depending on
complexity could be easier done with SQL

Report on all server error messags and some warning messages reported over
last 24 hours
        - we just do a q actlog then filter for specific messages
        - we stated with all error messages then slowly filtered out ones we
didn't want
        - we started with no warning\informational and then slowly added
ones we wanted
        - this might be easier with sql but we are using a perl script!

we also run an online error monitor that emails, net sends, or pages us on
various errors throughout the day

Daily\weekly processing report
        - client report from accounting log, reports on gb, transfer rate,
media wait ... for backup/restore, archive/retrieve

TSM Health Check every 5 minutes - email and page if TSM is down

TSM System check - checks for things like DB > %utilization, # of scratch
tapes, stgpool > %, db and log volumes synched, db cache hit > 98%, log < %,
disk volumes all online and available, no tapes in error state or specific
ones in read only that should not be, cleaning cart close to no uses ....

report of tapes not written to in over x amount of time (we then reclaim
these)


Tim Rushforth
City of Winnipeg