[Sunhelp] Faulty SCSI Controller?

[S Condit]

Hi,
	One of our machines is an ultra 5 fitted with a PCI symbios 53c875 dual 
channel UW SCSI controller. Earlier today the machine couldn't see any of 
the SCSI drives we had mounted. The console was filled with timeout errors 
and after reboot doing a probe-scsi-all shows nothing - just returning to 
the ok prompt. A reboot using the second channel rather than the first made 
no difference. A test scsi shows no devices to test (is this because SCSI 
isn't built into the Ultra5 ??) and a test all hangs when it tries to test 
the SCSI controller. I haven't looked into the OS itself too much as I 
can't get results from the OK prompt but it's Solaris 2.6 for those who 
have to know :-)

Console Message-------------

/pci at 1f,0/pci at 1/scsi at 1 (glm0)
Connected command timeout for target 0.0   (there is no target 0 BTW)


The drives themselves are fine as they've been mounted on a SPARC 5 (slower 
but at least it works!)  and are giving no problems.
My feeling is that the controller itself may have failed but any other 
theories are welcome. I've included the error messages below.

What is a little too coincidental is that /var/mail (mounted on one of the 
SCSI drives) filled just before this happened (due to a user receiving 
600MB of email over the weekend - naturally I've taken out a contract on 
the sender )

anyway sequence of events -

/var/adm/messages ---------------

Mar 27 09:00:26 phoenix unix: NOTICE: alloc: /var/mail: file system full
Mar 27 09:05:02 phoenix unix: glm0:     Cmd (0x605f97a0) dump for Target 3 
Lun 0
:
Mar 27 09:05:02 phoenix unix: glm0:             cdb=[ 0x2a 0x0 0x0 0x20 0x1 
0x6a
  0x0 0x0 0x10 0x0 ]
Mar 27 09:05:02 phoenix unix: glm0:     pkt_flags=0x4000 
pkt_statistics=0x60 pkt
_state=0x7
Mar 27 09:05:02 phoenix unix: glm0:     pkt_scbp=0x0 cmd_flags=0x1860
Mar 27 09:05:02 phoenix unix: WARNING: /pci at 1f,0/pci at 1/scsi at 1 (glm0):
Mar 27 09:05:02 phoenix unix:   Connected command timeout for Target 3.0
Mar 27 09:05:02 phoenix unix: WARNING: ID[SUNWpd.glm.cmd_timeout.6017]

it then tries to revert to non wide mode then to async mode (failing to do so)

The messages repeat but even now the machine does find the scsi controller 
on boot.

Mar 27 12:11:21 phoenix unix: glm0:     Rev. 4 Symbios 53c875 found.
Mar 27 12:11:21 phoenix unix: glm0 is /pci at 1f,0/pci at 1/scsi at 1
Mar 27 12:11:21 phoenix unix: glm1:     Rev. 4 Symbios 53c875 found.
Mar 27 12:11:21 phoenix unix: glm1 is /pci at 1f,0/pci at 1/scsi at 1,1


any ideas or is it new controller time ??

Thanks,
Stephen Condit