CPU error message HELP

[NavinB]

Hi All,

I got this error message this weekend.....

Mar 26 21:59:04 ospf unix: panic[cpu0]/thread=700f1ba0:
Mar 26 21:59:04 ospf unix: CPU0 Ecache Writeback Data Parity Error: AFSR 0x0000000
0.00800008 AFAR 0x000001ff.00000000
Mar 26 22:00:33 ospf savecore: reboot after panic: CPU0 Ecache Writeback Data
Parity Error: AFSR 0x00000000.00800008 AFAR 0x000001ff.00000000

How serious is this ???And what steps(commands) should I take to see/diagnosis should be done???

Any help would be greatly appreciated ....

Thanx in advance.....

 

[octorian]

Naming your machine after a routing protocol, eh?

In any case, from those messages I'd guess your CPU has bad cache. Out of curiosity, what sort of CPU module is this? (I know there were known issues with the cache on some models of the UltraSparc-II modules in the bigger enterprise servers a few years ago)

 

[NavinB]

Hi ,

The server is Ultra 10(Ultrasparc 143MHz) and CPU is working fine now..I checked it using prtdiag command....Cache status is also fine....Is there any precaution that I should take ???/

 

[octorian]

I think you're getting machines confused... The Ultra 10 CPUs were UltraSPARC-IIi's, and came in speeds of 300, 333, and 440MHz. If you mean an Ultra 1, then 143MHz sounds plausible. In any case, I think I've heard that the Ultra 1 might run hot, so make sure the cooling is good. Otherwise, I don't know what to say.

 

[NavinB]

I again got this error message but diff from earlier one.....Pls help to solve it ...

Mar 31 10:43:36 ospf unix: panic[cpu0]/thread=700f0c80:
Mar 31 10:43:36 ospf unix: mutex_enter: bad mutex, lp=711c3444 owner=700f0c80
thread=700f0c80


And u r right ...server is ultra 1 ....

 

[johngiggs]

NavinB,

I looked around online and I believe this panic indicates some type of failure in the kernel code. Try installing Sun's recommended patches.

John

 

[cndcadams]

I have been a technician for Sun systems for 8 years and replaceing the cpu is what we would do, and what Sun used to tell us to do. The CPU0 Ecache Writeback Data Parity Error is the key. The crashing of the cpu is going to at some point corrupt your OS and you will start have software issues.

 

[octorian]

And since the CPU on an Ultra 1 was soldered to the motherboard, last I checked, I guess that means effectively a whole new machine is in order (or a new system board, which is probably harder to get a hold of than a new box).