On 15-Jul-99 Dwight Cook wrote:
>      Our adsm servers are running on AIX so we just set up a little script
>      that gets kicked off as a cron job that goes around to every adsm
>      server and performs a
>         q event * * begind=-1 endd=today  begint=00:00 endt=06:00
>      run the results through an
>         egrep '(Missed|Failed)'
>      then match the node with a node admin. & send out a mail message.

 We do exactly the same. Plus: checking "backup storage pool's", which
 tapes are unavailable/destroyed/readonly, does the backup database
 succeed, how many empty volumes in each storage pool, which node's are
 not in an association (i.e. not backup-ed automatically), (some) errors
 from the activity log, and some more.
 Oh, it's also handsome to display the 'days since last access' when a
 backup is missed, we don't worry if it is less then three days :)
 And a check on the database is also handsome. Send a mail if the
 "PCT_UTILIZED" is higher then 90 (databases will always grow and faster
 then you wish...)

 We use unique prefixes for our nodenames, and every person who is
 responsible for a bunch of node's (i.e. backup's), gets his own email
 about the status of the backup's of his nodes.

 Cheers,
 Henk ten Have.