iTera is just doing what it does knows how to do, watching everything on the system, kind like a phantom, looking (and apparently locking) things as he stealths around checking for changed objects.

After talking to iTera, I pulled the library out of the replication. I added code following the save to tape to push the objects to the backup machine using an iTera command for objects not being replicated. For the first time in more than a month, the backup ran without a hitch.

Thanks to everyone for their suggestions – there are some I still plan to investigate now that I’ve got the main problem taken care of.

Unexpected/unexplained locks seem to top the list.

The vendors almost always seem plenty knowledgeable on recovery or avoidance or whatever is needed.

Tom

I didn’t mention it because I couldn’t see how it was related in any way; I’ve been able to trace its effects elsewhere, so I assumed (mistakenly) that as I couldn’t see any interaction there was none.

I’ve been in touch with iTera & will be trying a different approach to the backups.

BTW, iTera didn’t have an explanation for why WRKOBJLCK didn’t show anything.

This is a HA environment? Well, in that case, that’s the first place to check. for future reference, that’s an element that should be mentioned in the beginning for many (unexpected/unexplained) lock-related issues.

Vendor contact is the best step.

Tom

IBM claims the problem lies with the iTera HA software. I’ve given Vision the information and am waiting to hear back.

Start by saving one library to one virtual tape cartridge. Then copy the virtual tape image to a real tape cartridge using the DUPTAP command. Then delete your test library then restore the test library from tape.

Once you have done that exercise you will understand and trust virtual tape.

I realize that this may seem like a significant change, but try virtual tape with a single TEST library then analyze how much work it would be to change your current backup process. It might surprise you. It might surprise me. ??

Another advantage is that virtual tape images can be FTP’s to another server for archival.

IPLed the system on Sunday, with a RCLSTG tossed in for good measure – no problems with the RCLSTG or the system save.

Checked the job log for the nightly backup on Monday; the problem persists. Contacted IBM & sent the job log, program dump, etc. for them to look over.

To ensure I had a backup of the file library in question, I added code to create a second save file (EC#FILE_1) and save the file library a second time. I figured this way I could be reasonably confident that I had something to fall back on if things hit the fan.

Looking over the job log this morning, it wasn’t surprising to see the familiar CPF3761 message. What was surprising was to see that it now read Cannot use FILE EC#FILE_1 in SAVELIBS.

Sent the job log, program dump, etc. so IBM can puzzle over this development.

Peter, thanks for posting the DMPSYSOBJ command; having it here has saved me some scrounging.