We (IBM Global Services) start the ADSM server with a perl script, like this:
open(SRVR,"dsmserv |")
The perl script can then monitor the output of the dsmserv process, and send
pages or email when it sees error messages.
We also have a perl script which reads through the activity log looking for
interesting events like:
migration begins / ends
expiration begins /ends
reclamation or move data begins/ends
node session begins/ends
tape mount/dismount
dsmserv restart
and prints out a timestamp and some numbers, so we can graph all this stuff.
It's hard to read 10,000 lines of activity log every day, but looking at a
graph is easier. We also graph storage pool size (both disk and tape) - that's
been useful for convincing customers that they really do need to buy more tape
storage.
|