We had some fun last week with RoadRunner having device time out errors
on several DLTs on multiple machines. Could all 14 of our DLTs developed
problems at the same time on five different machines? Yea, I didn't
think so either! I used TurboStore and no longer received device time
out errors. I did run into another situation, but that is another story
for the second half of this e-mail. Of course it prompt me to call the
HP Response Center. It was determined that we needed to install a new
version of DLT firmware. The new version of RoadRunner brought the
problem into full view instead of the problem nipping us in the heals
now and then. Thank you, Tivoli! HP upgraded out firmware from version
CD3C to v115. This was on Saturday and Sunday. Not only did the device
time out error go away, but so did those pesky device read and write
errors that seemed populate the RoadRunner job. I would recommend
talking to HP about an upgrade.

Second Half: While trying to solve the above problem we ran into a
situation were if you try to restore an KSAM file that is quite large
and it is replacing the file on disc, expect to see the job take high
CPU time with zero IOs. It is caused by a lock contention on the file.
We also found that if a PURGE or PRINT is done on the same file it
causes the same error. The only way to get the job/session to release
the file and abort is to reboot the system. A Ctl-A SHUTDOWN will never
reach (Shut 6) because the job/session will not go away. In this case
our file had a capacity of over twenty million records with an end of
file over nineteen million. I have sent HP a copy of the file and a
memory dump of the last time I hung some sessions. When I hear back I
will post something on the list.

We having fun now.

Rich Holloway
Systems Administrator
Providence Health Plans
Phone: (503) 574-7457
Fax: (503) 574-8614
[log in to unmask]

The opinions expressed are not necessarily those of my company. They
belong solely to me.