Ignite & KERNEL drivers/parameters

[Sean McConkey_3]

Using an Ignite tape from an N-Class system onto another N-Class system, I get the following panic on bootup after the recovery.

Boot
: disk(0/0/2/0.6.0.0.0.0.0;0)/stand/vmunix
7368704 + 893520 + 1231384 start 0x246868

real_page_alloc_addr() fails, status = -1

Console reset done.

Boot device reset done.

Stored message buffer up to panic:

System Panic:

linkstamp: Mon Apr 29 01:35:33 EDT 2002
_release_version: @(#)B2352B/9245XB HP-UX (B.11.00) #1:
Wed Nov 5 22:38:19 PST 1997
__kern_ci_revision: $Header: kern_sequence.c,v 1.6.106.512
97/11/05 18:01:46 msosa Exp $

panic: alloc_pdc_pages: real_page_alloc_addr() fails

PC-Offset Stack Trace (read across, top of stack is 1st):
0x00327454 0x00430238 0x0042f2f8
0x003368e4 0x00246da4
End Of Stack

I am sure this is caused by a driver, as error doesnt appear if I use the system file off the DR system, even when using the original tuneables. ie I rebuild the kernel using the dr systems system file.

I would have thought Ignite was capable of differentiating between hardware and building a Kernel with the necessary drivers in it.

It only seems to happen when using Ignite tapes from certain client systems.

Sean

[Clemens van Everdingen]

Hi,

You should look for HW differences !

What HW is in the system you are igniting and is not in the system the tape was made from ?

Then look for the driver's that should go with that HW.

C.

[Sridhar Bhaskarla]

Hi,

It looks to me like a PDC issue while trying to reallocate PDC into memory. Try upgrading the PDC on the new N server and then boot.

-Sri

[T G Manikandan]

check your memory boards whether it is properly seated.
I assume that the messages are due to a defect memory.

Thanks

[Paula J Frazer-Campbell]

Hi

This error msg suggest that a page in memory
was deallocated due to errors, and included in the system's page deallocation tables.

Check memory boards are appropriately seated.

Also ignite is/was designed to recover the server from which the ignite image was created.

It will have problems if the hardware and hardware paths do not match on another server.



HTH

Paula

[Tom Geudens]

Hi,
If I remember correctly Ignite makes a different kernel if the model string differs. You can force this by changing the control_cfg
file when creating the recovery tape.
You should make sure the _HP_CLONING parameter is false ...

Hope this helps,
Tom

[James Murtagh]

First of all, HP have already investigated this problem so the best bet is to contact the Response Centre......

I don't think there is PDC firmware or memory problems as from the message, I can tell page 0 of the IMM (Initial Memory Module) has already been initialised and the console and boot IODC have been read in after this.

From (my) memory, I think the problem occurs when the IMM size in the new server is smaller than the first and something is logged in the kernel about the size/addressing.....so when HPUXBOOT is reallocated at IMM - 2MB this tries to allocate it to an address past what is configured.

I think the problem occurs at the realmain() function.....possibly look in realmain.h although I don't have access to a HPUX server now.