SUMMARY: Sparc 10 hangs about every two weeks

From: Mark H. Price (mhprice@sandia.gov)
Date: Wed Nov 20 1996 - 13:59:22 CST


> I have a SPARCstation 10 running Solaris 2.5.1 that hangs
> about every two weeks. (Stop-A won't even work when it
> does hang; I have to power cycle the system.)
> The message log shows the following:
>
> unix: WARNING: M-to-S Asynchronous fault due to
> unix: user write - non-fatal
> unix: WARNING: Multiple Errors
> unix: WARNING: Error during User mode cycle
> unix: WARNING: Timeout Error
> unix: WARNING: Requested transaction: byte at (space e) 2087ddc1
> unix: WARNING: Specific cycle: short xfer at e:2087ddc1
> unix: WARNING: AFSR = 0xa088120e, AFAR0 = 0x2087ddc1, AFAR1 = 0x0
>
> Does anyone know what these messages mean?

The general consensus is that the hang is caused by a hardware problem.
That's where the consensus ends, although knowing it is a hardware
problem definitely helps. I was told it might be caused by:

1. a weird SBUS card (although the only added SBUS card in the system
   is a SunPC accelerator), or

2. a motherboard problem, or

3. a SIMM failure, or

4. a disk failure.

One kind soul even told me that unplugging and plugging in the
keyboard might get me to the boot prompt so that I could sync
the disk and reboot (which is certainly preferable to power cycling
the machine).

So now I'll try swapping out various components of the system.
Fortunately, I have another Sparc 10 with which I can swap pieces.

Thanks to everyone who replied. (This group is incredible; I even got
one reply before my original post was echoed back to me.)

Mark Price
mhprice@sandia.gov



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:16 CDT