We had a two drive failure on our AS400 and had to reinstall the operating system. Can we install A7.3 from scratch and then copy in our libraries from our latest backup and JDE function properly? Any ideas would help!
Well, it may be that I am not correctly understanding your disaster situation. I am assuming that you have the same IBM CPU and same IBM hardware serial number. It sounds like you had to replace a disk drive and a tape drive, but not the actual CPU itself. Now it does sound like your operating system backup was eaten up by the tape drive and lost, and thus you are saying you had to reinstall the IBM operating system? Would you not have a slightly older backup tape that you could recover from? Even if you did not and had to reinstall the operating system, JDE World is fairly independent of the OS/400 itself (other than needing to run on it). So I am thinking that you do NOT need to reinstall JD Edwards. You should be able to restore the JD Edwards libraries, though you would need to have the QGPL library as part of that restore process, and run okay.
If my assumptions are wrong, then what I said may not apply. Certainly if you restore onto a new CPU, with a different serial number, you will have to contact JDE support to get a new license key. When doing a hot site test, I restore the operating system, and then restore all the user libaries, get new license keys, and I am up and running. Note that in a hot site test I did not have to do a reinstall of JDE. But certainly the difference maybe that in a hot site test, I am restoring the operating system from backup tape (so not reinstalling OS/400). So I have not personally been involved in the exact situation that you may be facing (I doubt many folks have such experience).
You probably want to be in touch with JDE support and look to them for guidance on what to do, giving them your exact restore situation that you are facing.
Hope this helps a little bit. Bit surprised that these days a disk failure would cause a system crash (have not seen that since the 1980s) - most systems have disk mirroring or RAID of some kind now. So the system can still run while waiting for the disk drive to be replaced (perfomance may suffer, but better that than losing the whole system).
Thanks for your response John. In answer to one of your questions the machine was mirrored and had raid/ parity protected. That works fine as long as you only lose a single disk drive. Our problem came when the the IBM tech changed the failed drive. The system got to 31% storage restored and stopped. The IBM tech a second drive failed because it was missing from the disk configuration display. Instead of explaining to "we the customer" what he planned to do he took it upon himself to format all of our disk drives. His logic was that the second drive had stopped spinning so it must have failed. With a two drive failure raid protection is worthless. Now me being one who works on PC's when a drive doesn't boot once I don't concern it dead. I would have tried a few more attempts but that's water under the bridge. Hense my statement "a perfect storm". Oh yeah I don't know if I mentioned the second two disk crash mid week last week. Same two disk failure only this time a different tech and he recovered everything we had restored so that we didn't have to go back to square one again. Makes me think with the right IBM tech from the start we wouldn't be going through this now.
You may have touched on something when you mentioned the QGPL library. Is there any way you could list the contents of QGPL so that I can see if I am missing something? Again thanks for your help.
Yep, a mutiple disk drive failure will cause problems even when RAID is being used, no doubt about that. But it is unusual to have a multi-disk drive failure. This is the first time I have seen someone run into that kind of situation. We have had instances where a single drive has failed, but just a single drive. What a lot of companies do now is buy a second CPU (IBM used to have a special deal on buying a DR CPU, but don't know if it still exists - you would have to ask the IBM sales rep about that) and locate it in a second location, and then mirror the primary production machine to that DR machine. In case of a failure, the second machine would take over. You can also use this switchover ability to do maintenance on the primary machine (say operating system upgrade) while still keeping the system available for users. Given your experience, your management might be receptive to spending some money to do something like that.
QGPL library is unique to each installation. Normally there are not a lot of objects in that library. Usually output queues and job descriptions and sometimes device descriptions. Maybe job queues. JDE might have some licensing related objects in there. QGPL does get saved as part of the SAVSYS command, but we also save it as part of our user data save. But you cannot look at someone else's QGPL and be sure that you can spot something that you may or may not be missing.