[iKitsune]

I've recently been called to inflict service upon a p570 in a 4-drawer configuration (16-way, 80 GB RAM, unknown drive capacity). It's been having some oddball problems.

The customer initially reported that the system was running slow, and all 4 CPUs in the #2 drawer had deconfigured. I replaced the CPUs in the #2 drawer, and the system failed to boot. Last codes were C10099A0 and then B114E500, with the description of "Processor Bus error". Fearing damage to the flexible cable at the front of the system, I replaced that with a known-good spare with no apparent effect B114E500, with a description of "Processor Bus Error". Then things got weird.

After about a 4th IPL attempt after swapping hardware around to try and eliminate RAM/etc, the Service Processor RTC died. I replaced the Service Processor with a known-good spare.

All 4 drawers, operated independently (CPU bus and Service Processor cable disconnected, Processing Unit ID in ASMI set to B2 instead of B3) boot and pass diagnostics.

After I tested all the drawers independently, and put everything back in its full configuration, the boot problems changed.

I now get

C10099A0
C14120FF
C1C3C200
C1C3C201
(continues counting up to C1C3C220, and loops several times)
B170E500

I can't even find progress code C1C3C2xx in the IBM documentation. Time to throw in the towel and call, or am I missing something obvious?