Autoswitch not autoswitching

[Dan Hardison]

I have the following configuration for /dev/vg02. According to the output "Autoswitch" is "on". My assumption here is that if the controller for disks c13t*d* was to go down it would autoswitch to the controller for disks c16t*d*. However if I manually disconnect the Fibre cable from the c13 side, the system just hangs. I thought maybe it was a timeout issue, but I've waited for over 30 minutes just to be sure. Is there something else I need to set/configure?

TIA,
Dan

# vgdisplay -v /dev/vg02
--- Volume groups ---
VG Name /dev/vg02
VG Write Access read/write
VG Status available
Max LV 255
Cur LV 1
Open LV 1
Max PV 16
Cur PV 4
Act PV 4
Max PE per PV 8683
VGDA 8
PE Size (Mbytes) 2
Total PE 34724
Alloc PE 34724
Free PE 0
Total PVG 4
Total Spare PVs 0
Total Spare PVs in use 0

--- Logical volumes ---
LV Name /dev/vg02/lvol1
LV Status available/syncd
LV Size (Mbytes) 34724
Current LE 17362
Allocated PE 34724
Used PV 4


--- Physical volumes ---
PV Name /dev/dsk/c13t0d0
PV Name /dev/dsk/c16t0d0 Alternate Link
PV Status available
Total PE 8681
Free PE 0
Autoswitch On

PV Name /dev/dsk/c13t1d0
PV Name /dev/dsk/c16t1d0 Alternate Link
PV Status available
Total PE 8681
Free PE 0
Autoswitch On

PV Name /dev/dsk/c13t2d0
PV Name /dev/dsk/c16t2d0 Alternate Link
PV Status available
Total PE 8681
Free PE 0
Autoswitch On

PV Name /dev/dsk/c13t3d0
PV Name /dev/dsk/c16t3d0 Alternate Link
PV Status available
Total PE 8681
Free PE 0
Autoswitch On


--- Physical volume groups ---
PVG Name PVG1
PV Name /dev/dsk/c13t0d0
PV Name /dev/dsk/c13t1d0

PVG Name PVG2
PV Name /dev/dsk/c13t2d0
PV Name /dev/dsk/c13t3d0

PVG Name PVG3
PV Name /dev/dsk/c16t0d0
PV Name /dev/dsk/c16t1d0

PVG Name PVG4
PV Name /dev/dsk/c16t2d0
PV Name /dev/dsk/c16t3d0

[Thayanidhi]

Hi,

LVM cumulative patch may help you!
Do you have this already.

PHCO_27369 (for 11.x)

TT

[Jerry Zhang]

Try to switch manually by
# pvchange -s /dev/dsk/c16t*d*
This will set c16t*d* as a primary path and c13t*d* as a secondary one. If you can do it, then it is not a hardware problem.

[Dan Hardison]

Jerry,

The pvchange worked fine and I'm still able to access the LVOL. So, since it's not a H/W problem........... Maybe a patch? I'm running 11.11

Dan

[Eugeny Brychkov]

Dan, what's hardware connected? Which FC HBA do you use? FC switches? Disk array / JBOD?
Eugeny

[Dan Hardison]

Eugeny,

I'm running a rp5470 with 2-PCI 2GB Fibre Channel Adapters (A6795A) connected to 3-DS2405 JBOD disk systems (A6250AZ). Total of 32-18Gb disks (15-15-2). No FC switches. It is configured with 1 FC adapter feeding each "side" of the JBOD.

Dan

[Eugeny Brychkov]

Try more tests: disconnect from another HBA (c16) - the same symptom?
What tdutil /dev/tdX shows you? Are both 2GB/s? Does 'ioscan -fn' show all disks correctly?
Can you attach all these to you next reply?
Eugeny

[Dan Hardison]

Eugeny,

Regardless of which FC I disconnect, the c13 side or the c16 side, the system just hangs. Both FC are set at 2Gb. Attached is the output of ioscan and tdutil.

Thanks,
Dan

[Eugeny Brychkov · Accepted solution]

EMS bundle Sep 2002: http://www.software.hp.com/cgi-bin/swdepot_parser.cgi/cgi/displayProductInfo.pl?productNumber=B6191AAE
Check that you have Sep 2002 GR amd HWE patches installed (0209). I suspect that not FC disks are causing problem, but A6255A LCC controller and maybe EMS monitoring services hang when LCC dissapears. Try disabling EMS monitoring.
Try disconnecting BOTH links! Will your system get live? Try with only one JBOD connected.
If you're sure that your system is configured properly you can disconnect one of links, and press TOC button. Then ask HP to analyze crash dump. What do you think?
Eugeny

[Dan Hardison]

Eugeny,

Sounds like a plan to me! I'll save these steps until after the Holidays and give them a try. I'm getting ready to drive 10 hours to get back home. Thank you very much for your assistance. Have a great holiday!!

Dan