Solaris

I received a new error that I have not seen before on my Solaris 10 server last night:

From /var/adm/messages
Jan 8 04:59:26 nediaccomp005j scsi:[ID 365881 kern.info] /pci@780/pci@0/pci@9/scsi@0 (mpt0):
Jan 8 04:59:26 nediaccomp005j Log info 31010000c received for target 1.
Jan 8 04:59:26 nediaccomp005j scsi_status=0, ioc_status=8043, scsi_state=0
Jan 8 04:59:26 nediaccomp005j scsi:[ID 107833 kern.notice] Device is gone


What does this mean? How is the device "gone"? It's an internal disk.

Any input appreciated!

-Fred

firstly check all your mount points with df -k and try accessing them. Check format output and iostat -En and see if there are any hard errors on the disk. "Device is gone" not necessarily be an internal disk.We need more outputs to verify.

I've had a similar experience, see below. The last contributor suggested that the problem may not be caused by the hard disk itself, but rather by an external device. In my case all servers have a cd drive. Could it be that under certain circumstances a cd/dvd could trigger this error on the hard disk.
The 3 messages listed below are from /var/adm/messages at the time the T2000 server hung. The same error occured on several T2000 servers over a period of 6 weeks. The submirrors d16 and d17 of mirror d15 were in a state of needs maintenance. The slice 0 of one disk was somehow corrupt and a simple metareplace wasn't sufficient to bring the system back up.
I also would really like to find the route cause of this error

<date> <hostname> scsi_status=0, ioc_status=8043, scsi_state=0
<date> <hostname> scsi: [ID 107833 kern.notice] Device is gone
<date> <hostname> md_stripe: [ID 641072 kern.warning] WARNING: md: d17: write error on /dev/dsk/c0t3d0s6
<date> <hostname> 1 md_mirror: [ID 104909 kern.warning] WARNING: md: d17: /dev/dsk/c0t3d0s6 needs maintenance

The servers have in the meantime been updated to the firmware listed below, also the problem has not reoccurred since then but I still dont know what lies behind the error.
System Firmware 6.6.4 Netra[TM] T2000 2008/07/01 02:01
---------------------------------------------------------
ALOM-CMT v1.6.4 Jun 6 2008 16:52:02
VBSC 1.6.4.a Jun 6 2008 05:19:51
OBP 4.28.9 2008/06/30 21:26
Hypervisor 1.6.4 2008/06/06 04:57
Netra[TM] T2000 POST 4.28.6 2008/05/23 12:34

Hi FredSmith..........
Greatings.


Once if ur working in Production environment plz check the hardware parts(like disk's) is there any red amber light disk will be fault.

Otherwise ur in remot support,

check the
#df -h
#iostat -En,
#prtconf
#dmesg aslo.
then u can find the result.Once still ur not getting any info u have to check
#prtdiag -v ----- it will display all faulty devices.
According to that u should troubleshoot it.
Otherwise plz send ur output for above cmds(must prtdiag -v).


=============================================
=============================================

Hi conor01,
Greatings,

yours also same problem.Plz execute above commands.
According to ur scrap(md: d17: /dev/dsk/c0t3d0s6 needs maintenance
)...ur facing some problem in raid configuration.
it might be adding extra disk size/disk partition.......
u can run prtdiag -v...it vll display ur hardware problems.


Best of Luck........


Regards,
Arjun Reddy.

Thanks for you reply.I checked the commands, everything looks fine. The thing is the the server is no longer in this state. It was recovered using metareplace commands etc. I don't believe it was a hardware error as it occured on several severs over a perid of time. But I wiuld like to find the root cause of the error.