LVM and boot problems after 11.0 split and 11.23 cold install

[Vball23]

I have an n4000 that was running 11.0. I split the 11.0 mirror and cold-installed 11.23 on the primary drive and left 11.0 on the secondary drive.

Now, I can only boot into 11.0 or 11.23 by interacting with the IPL and booting hpux -lm.

Otherwise, if I just let it boot normally into 11.0 or 11.23, it gives me a software failure, panic and reboot.


Boot
: disk(0/0/2/0.6.0.0.0.0.0;0)/stand/vmunix
14123008 + 6565888 + 2053096 start 0x22cc68

real_page_alloc_addr() fails, status = -1
Unable to allocate required memory between 0x7e380000 and 0x7ed00000.
Check the PDT data for deallocated pages within this range.

************* SYSTEM ALERT **************
SYSTEM NAME: con_so88
DATE: 03/22/2006 TIME: 16:21:25
ALERT LEVEL: 12 = Software failure

REASON FOR ALERT
SOURCE: 1 = processor
SOURCE DETAIL: 1 = processor general SOURCE ID: 0
PROBLEM DETAIL: 0 = no problem detail

.
.
.
System Panic:

linkstamp: Mon Mar 20 18:12:20 EST 2006
_release_version: @(#) $Revision: vmunix: B11.23_LR FLAVOR=perf Fri Mar 19 15:24:05 PST 200
4 $
panic: alloc_pdc_pages: Unable to allocate memory for PDC.

PC-Offset Stack Trace (read down, top of stack is 1st):
0x00783a98 (panic+0xc0)
0x007b57dc (alloc_pdc_pages+0x2fc)
0x00805658 (DoCalllist+0x100)
0x0060adfc (realmain+0x24)
0x0022d230 ($locore+0x68)
End Of Stack


Any thoughts?

[melvyn burnard]

This looks like a hardware problem. Your server appears to have a full PDT or page Deallocation Table, indicating you have had memory errors and pages have been deallocated, and it is trying to to it again at boot time, and this fails.
I woul dsuggest you log a call with your local HP Response Centre

[Vball23]

It either has 3rd party memory or incorrectly installed memory, but if I clear the PDT, it still gives this failure immediately.

I believe these are 2 different issues and even if the memory issue is cleared up, I think the lvm problem will remain.

[Vball23]

I have done these tests:

I booted from the 11.0 disk and specified hpux -lm. The server booted up fine.

I booted from the 11.0 disk and did not interact. The server crashed as above.

I booted from the 11.23 disk and specified hpux -lm. The server booted up fine.

I booted from the 11.23 disk and did not interact. The server crashed as above.

So, I can only boot into lvm maintenance mode and then into multiuser, so whatever lvm maintenance mode does to the filesystems automatically fixes the problem.

Any thoughts/suggestions?