Fast Data Access MMU Miss

[reinstalled]

Getting the error Fast Data Access MMU Miss when trying to do a probe-scsi. probe-scsi-all works but just for the external tape drive then returns the MMU error again. I know this is a catch all error message and hard to troubleshoot.

Does anyone have any ideas?

Thanks

 

[bfitzmai]

I've been looking for my link to the Kernel Debugger Instructions (kdb). When I was having this error message with some of my Themis Workstations, I used kdb to capture the problem for the techs.

I had 3 of my Themis workstations display this error message over the last 5 years. The fix for this was a CPU replacement.

 

[reinstalled]

That particular error is not very descriptive and according to Sun it means a lot of different possibilities. They also say that it's most likely a third party application fault and to contact the vendor. Since this is a brand new box with no third party app's installed I'm not buying their explaination.

So it's a catch-all error message that doesn't give any indication as to the problem, I've heard memory, cpu, hardware, the way the OS was installed, etc......
and all of those issue's were the actual problem per case.

So how do I troubleshoot this error message without trying this or that?

 

[bfitzmai]

Have you tried POST Diagnostics. You can capture the output of this diagnostics via /dev/ttya connected through a null modem to another Sun Box or a Laptop using hyperterminal. Not sure if it will help, but it may point you in the right direction.

 

[daFranze]

some OBP Errors are caused by just halting the OS and run a OBP Command.

Please follow this procedure:

init 0
ok setenv auto-boot? false
#depending on which Hardware and OBP Version you are running it is ether or
ok reset
ok reset-all

when the OBP is back you can run eg
ok probe-scsi-all

do not forget to set auto-boot? variable to the value you set before testing (usually it is set to true)

as bfitzmai pointed out: so many data mmu miss error's are "bad cpu"s

Best Regards, Franz
--
Solaris System Manager from Munich, Germany
I used to work for Sun Microsystems Support (EMEA) for 5 years

 

[reinstalled]

Thanks Franz,

One question though: Setting my env to not auto boot, wouldn't
a stop+A after the the banner do the same thing and halt the boot process? or is there another reason?
Just curious really...I just finished my Solaris cert's last fall and am now learning what they don't teach you in class

Anyway, I have reinstalled and do not see the error any longer. I am chaulking this one up to mis-configured software, package, or patch since I don't see any obvious hardware error's. It could be that the error show's up only after a period of operation but I will monitor and see.

Thanks all for your responses!!!

 

[daFranze]

no STOP-A is not the same, it leaves OBP (and some registers) in an undefined state, the auto-boot? false + reset - Procedure stops booting in a definded state.

after finishing schools/classes you will learn most things, because only real life can invent so many errors...
(I hope you can pick up, doing Smalltalk is not that easy for a non native speaker)

Best Regards, Franz
--
Solaris System Manager from Munich, Germany
I used to work for Sun Microsystems Support (EMEA) for 5 years