Hiya,

I have a robot with 4 drives and just set up ITC for our Vault jobs with
4.5FP3 running on Solaris8. I was first shocked to see that each of
these ITC jobs opened three jobs instead of the usual two, but found a
previous post to this list stating that one of those jobs is a "control
job".

These jobs are kicked off 10 minutes before other backups, so I expected
to see 2 drives being used for ITC and the other 2 used for "normal"
jobs. What I observed, however, was that one drive wrote the primary
copy, one wrote the duplicates, one did "normal" backups, and one sat
there doing absolutely nothing, although there were plenty of jobs
waiting in the queue.

When one of the tapes filled, a new one was inserted into the "unused"
drive and the drive that had just finished took over its state of apathy
- I can thereby rule out a problem with the drives.

When the number of ITC jobs (primary+duplicate+control) dropped below
the MPX for the schedule, which is lower than the global max MPX per
drive setting, a tape was mounted in the fourth drive and things began
working as I had originally expected.

MPX for all schedules is set to 10, the global max MPX per drive is set
to 25 (which, having noticed it, seems rather silly - wonder what I was
smoking) - I observed a maximum of 30 "real" jobs (ie, not including the
control jobs) running simultaneously instead of the usual 40.

Could it be that these "control jobs" count toward the maximum MPX of
the drives?

Any insight is appreciated.

Thanks,
Bruce