Problem with cfgadm and fc tape library

[Rafal Zawierta]

Hello.

After rebooting tape library I get on one of my Sol10 server such things:
-bash-3.00# cfgadm -al
Ap_Id Type Receptacle Occupant Condition
c0 scsi-bus connected configured unknown
c0::dsk/c0t0d0 disk connected configured unknown
c0::dsk/c0t1d0 disk connected configured unknown
c0::emcpsf0 unknown connected configured unknown
c1 scsi-bus connected configured unknown
c1::dsk/c1t4d0 CD-ROM connected configured unknown
c1::emcpsf1 unknown connected configured unknown
c2 fc-fabric connected configured unknown
c2::2103000d7768a3ef unavailable connected unconfigured failed
c2::50060161302219dc disk connected configured unknown
c2::50060169302219dc disk connected configured unknown

c2::2103000d7768a3ef should appear as tape

But if I try:
-bash-3.00# cfgadm -c unconfigure c2::2103000d7768a3ef
cfgadm: Attachment point not found

And:
-bash-3.00# cfgadm -c unconfigure c2
cfgadm: Library error: remove operation failed:
/devices/pci@84,4000/lpfc@2/fp@0,0/ssd@w50060169302219dc,2: I/O error
remove operation failed:
/devices/pci@84,4000/lpfc@2/fp@0,0/ssd@w50060161302219dc,2: I/O error
remove operation failed:
/devices/pci@84,4000/lpfc@2/fp@0,0/ssd@w50060161302219dc,2: I/O error
remove operation failed:
/devices/pci@84,4000/lpfc@2/fp@0,0/ssd@w50060161302219dc,1: I/O error
remove operation failed:
/devices/pci@84,4000/lpfc@2/fp@0,0/ssd@w50060161302219dc,0: I/O error
remove operation failed:
/devices/pci@84,4000/lpfc@2/fp@0,0/ssd@w50060169302219dc,2: I/O error
remove operation failed:
/devices/pci@84,4000/lpfc@2/fp@0,0/ssd@w50060169302219dc,1: I/O error
remove operation failed:
/devices/pci@84,4000/lpfc@2/fp@0,0/ssd@w50060169302219dc,0: I/O error
failed to unconfigure ANY device on FCA port

#dmesg
Jul 18 09:39:36 bazap-base.pwr.wroc.pl scsi: [ID 107833 kern.warning]
WARNING: /pci@84,4000/lpfc@2/fp@0,0/ssd@w50060161302219dc,1 (ssd4):
Jul 18 09:39:36 bazap-base.pwr.wroc.pl Error for Command: read(10)
Error Level: Retryable
Jul 18 09:39:36 bazap-base.pwr.wroc.pl scsi: [ID 107833 kern.notice]
Requested Block: 59483200 Error Block: 59483200
Jul 18 09:39:36 bazap-base.pwr.wroc.pl scsi: [ID 107833 kern.notice]
Vendor: DGC Serial Number: 1F0000F5D4CL
Jul 18 09:39:36 bazap-base.pwr.wroc.pl scsi: [ID 107833 kern.notice]
Sense Key: Unit Attention
Jul 18 09:39:36 bazap-base.pwr.wroc.pl scsi: [ID 107833 kern.notice]
ASC: 0x29 (power on, reset, or bus reset occurred), ASCQ: 0x0, FRU: 0x0
Jul 18 09:39:37 bazap-base.pwr.wroc.pl scsi: [ID 107833 kern.warning]
WARNING: /pci@84,4000/lpfc@2/fp@0,0/ssd@w50060169302219dc,0 (ssd2):
Jul 18 09:39:37 bazap-base.pwr.wroc.pl Error for Command: write(10)
Error Level: Retryable
Jul 18 09:39:37 bazap-base.pwr.wroc.pl scsi: [ID 107833 kern.notice]
Requested Block: 26179 Error Block: 26179
Jul 18 09:39:37 bazap-base.pwr.wroc.pl scsi: [ID 107833 kern.notice]
Vendor: DGC Serial Number: 1E00000DD1CL
Jul 18 09:39:37 bazap-base.pwr.wroc.pl scsi: [ID 107833 kern.notice]
Sense Key: Unit Attention
Jul 18 09:39:37 bazap-base.pwr.wroc.pl scsi: [ID 107833 kern.notice]
ASC: 0x29 (power on, reset, or bus reset occurred), ASCQ: 0x0, FRU: 0x0

Does anybody have an idea what to do - now it is impossible to reboot
that server.

Best regards
Rafal Zawierta

[Blazej]

Rafal Zawierta pisze:

Hi,

did you found a solution for that? What is library vendor/model/type? Do
you have persistent binding configured for tapes? What kind of FC
interfaces you have? What settings of configuration files: qla,
emlx.conf, st.conf?
Did you checked zonning if its working properly?
I/O error may suggest driver or FC problems - as well as HW problem. Did
you checked this drive from library RMU?
If whole controller cannot be reconfigured - this can indicate FC card
problem - check leds and/or management soft for your FC cards. Any older
messages problems?

Regards,
Blazej

[Rafal Zawierta]

> Hi,
>
> did you found a solution for that? What is library vendor/model/type? Do
> you have persistent binding configured for tapes? What kind of FC
> interfaces you have? What settings of configuration files: qla,
> emlx.conf, st.conf?

Not yet ;)
The library is a VTL from EMC, but Solaris 'see' some kind of ULTRIUM
device.

> Did you checked zonning if its working properly?
> I/O error may suggest driver or FC problems - as well as HW problem. Did
> you checked this drive from library RMU?
> If whole controller cannot be reconfigured - this can indicate FC card
> problem - check leds and/or management soft for your FC cards. Any older
> messages problems?

My VTL is connected to 4 Solaris servers. On 3 of them everything is OK,
only that one seems to be 'broken'. Maybe the solution is to reboot that
server, but now it is impossible, maybe next week. I'll give an
additional informations as soon as I'll reboot my server.

> Regards,
> Blazej

Rafal

[kmem]

If you're not seeing the problem on other hosts that are connected and
configured the same way, then I wouldn't suspect a SAN configuration
problem.
(unless there's something funky going on with the zone for this host).
If other hosts see the device with no problems, there's probably
nothing wrong with the device itself.

As the OP suggested, perhaps the HBA seen on this c1 attachment
point: /pci@84,4000/lpfc@2/fp@0,0/ is suspect.
The HBA cards are Emulex right?

With the Qlogic cards, you can run luxadm -e port to see the physical
path of the HBA and see if it's connected or not.
What's the similar command for the Emulex cards?

For each HBA on the host, you should at least see some message in the /
var/adm/message file when the driver is successfully loaded.
Do you see such a message for the HBA at this path: /pci@84,4000/
lpfc@2/fp@0,0/?

On Jul 25, 9:40 am, Rafal Zawierta <zawi...@mampocztenagmail.com>
wrote: