Re: [parisc-linux] HP B132L hangs

[Joel Soete <joel.soete@tiscali.be>]

Sebastian Brueckner wrote:

> Hi!
>
> In the last few days the B132l I use as DSL router completely hung 3 
> times. Every time requiring a power cycle to revive it. It is running 
> Debian testing with kernel 2.4.18-hppa.
>
> To test the machine I tried to compile a new kernel. After compiling a 
> few files everything hung and spit out the following messages:
>
> scsi0: (2:0) Synchronous at offset 8, period 100ns
> scsi0: (2:0), UNEXPECTED PHASE after command phase (CD BSY REQ CMD_OUT)
>          len = 6, cmd =Read (10) 00 00 35 e2 00 00 00 18 00
> scsi0: Bus Reset detected, executing command 10081c00, slot 00010520, 
> dsp 000902 failing command because of reset, slot 00010520, cmnd 10081c00
>  failing command because of reset, slot 00010654, cmnd 10081a00
>  failing command because of reset, slot 00010788, cmnd 10081800
>  failing command because of reset, slot 000108bc, cmnd 10081600
>  failing command because of reset, slot 000109f0, cmnd 10081400
>  failing command because of reset, slot 00010b24, cmnd 10081200
>  failing command because of reset, slot 00010c58, cmnd 10081000
>  failing command because of reset, slot 00010d8c, cmnd 10082e00
>  failing command because of reset, slot 00010ec0, cmnd 10082c00
>  failing command because of reset, slot 00010ff4, cmnd 10082a00
>  failing command because of reset, slot 00011128, cmnd 10082800
>  failing command because of reset, slot 0001125c, cmnd 10082600
>  failing command because of reset, slot 00011390, cmnd 10082400
>  failing command because of reset, slot 000114c4, cmnd 10082200
>  failing command because of reset, slot 000115f8, cmnd 10082000
>  failing command because of reset, slot 0001172c, cmnd 100a4e00
> scsi0: (2:0) Synchronous at offset 8, period 100ns
> scsi0: (2:0), UNEXPECTED PHASE after command phase (CD BSY REQ CMD_OUT)
>          len = 6, cmd =Read (10) 00 00 35 e2 00 00 00 18 00
> scsi0: Bus Reset detected, executing command 10081c00, slot 00010520, 
> dsp 000902 failing command because of reset, slot 00010520, cmnd 10081c00
>  failing command because of reset, slot 00010654, cmnd 10081a00
>  failing command because of reset, slot 00010788, cmnd 10081800
>
> ... and so on and so forth ...
>
> The power seitch does nothing, only power cycling helps. I tried 
> compiling the kernel again with the same results...
>
> What do these messages mean? Is the hd defective or is it some problem 
> with the scsi bus?

Yes it looks like a disk becoming defective. The best way to be sure is 
to use a hp cdrom containing Diagnostic tools (iirc the last hpux 
support + would contains it). This cd is bootable and contains tools 
allowing you test your disk.

If you have another system on which you can connect your suspected disk 
you can also try a dd cmd like:
dd if=/dev/rdsk/c0txd0 of=/dev/null bs=2048k (where you replace 
/dev/rdsk/c0txd0 by the actual disk path)

>
> The machine worked flawlessly for half a year... I have no idea what 
> went wrong!

It seems to confirm your suspition (i have to manage remotely some 300 
hp server: a few b132, b180, b2000, A500, K250, L, N and a lot of D; and 
the most peaces we have to replace are disk becoming defective)

hth,
    joel