1748275 Members
3723 Online
108761 Solutions
New Discussion юеВ

Re: BL465 G1 VC-FC

 
Andrew Hellyer
Advisor

BL465 G1 VC-FC

Hi,
I have a blade which has been giving us on going issues for the past year or so.
It is a BL465 G1 in a C7000 chassis in slot 10.

Randomly, its HBA will lose connectivity to the SAN only on port1.
No other blades experience this issue in the same enclosure.
How we online the port again, is to 'unassign' the VC SAN profile for this blade for the VC module which is in Bay 3, and reapply.
This will bring the port back online where it will operate happily for an unknown period of time.
It doesnt appear to be load dependand, as it will do it during the day OR during backups at night.

VC SAN firmware is at 1.32
OA firmware is at 2.51
Blade ROM is A13
iLO2 firmware is 1.82
PMC firmware is 0.5
Active/Passive VCM firmware is 2.10

Server OS is 2008 and is running up to date MPIO/Storport and HBA firmware.
STORport 9.1.8.27
Driver 5.03.02
BIOS 2.15
HBA type QMH2462

VC SAN interconnect module is HP 4Gb VC-FC Module.

There are no logs in the VCM pertaining to link faults of any sort, and nothing on the core switches.

The VC SAN is configured in STATIC login distribution, and all speeds are set to auto.

No target persistent binding is used throughout the enclosure.

Any hints or direction as to where to look would be helpful - we replaced the HBA and checked all the seatings but to no avail.
I am now looking at replacing the VC SAN module.

Thanks
6 REPLIES 6
Johan Guldmyr
Honored Contributor

Re: BL465 G1 VC-FC

Hey,

"and nothing on the core switches." - What kind of SAN-switches would that be? Is there nothing visible in the fabstateshow (if brocade) when the port goes offline either?

If you have a large environment this log fills up pretty fast.

Firmwares:

http://h18000.www1.hp.com/products/blades/components/c-class.html

(for compatibility list)

Your blade ROM is A15 - but what date is it? There are lots of fixes on the BIOS for the bl465 g1 as well.

OA 2.51 - old - there has been maaany new firmware releases after this one.
VC - 1.32 - do you mean 2.32? 2.32 is listed as supported in the release set that has OA firmware version over 2.60.

Fix in VC 2.34:

├в ┬в Resolved an issue where a VC network outage would occur due to the inability to detect the state of physical servers.
Andrew Hellyer
Advisor

Re: BL465 G1 VC-FC

Core switches are HP Brocade Silkworm 4100 (32 port) - does not have 'fabstateshow' command (FOS 6.4.0b)
The core switchport doesnt go offline at all.
We do see some errors on linux hosts within the same enclosure however:

Jan 28 17:26:43 chclnxora2 multipathd: 8:0: mark as failed
Jan 28 17:26:43 chclnxora2 multipathd: mpath13: remaining active paths: 7
Jan 28 17:26:43 chclnxora2 multipathd: dm-6: add map (uevent)
Jan 28 17:26:43 chclnxora2 multipathd: dm-6: devmap already registered
Jan 28 17:26:44 chclnxora2 multipathd: dm-13: add map (uevent)
Jan 28 17:26:44 chclnxora2 multipathd: dm-13: devmap already registered
Jan 28 17:26:44 chclnxora2 multipathd: 66:16: mark as failed
Jan 28 17:26:44 chclnxora2 multipathd: mpath20: remaining active paths: 7
Jan 28 17:26:51 chclnxora2 multipathd: sda: tur checker reports path is up
Jan 28 17:26:51 chclnxora2 multipathd: 8:0: reinstated
Jan 28 17:26:51 chclnxora2 multipathd: mpath13: remaining active paths: 8
Jan 28 17:26:51 chclnxora2 multipathd: dm-6: add map (uevent)
Jan 28 17:26:51 chclnxora2 multipathd: dm-6: devmap already registered
Jan 28 17:27:44 chclnxora2 multipathd: sdah: tur checker reports path is up
Jan 28 17:27:44 chclnxora2 multipathd: 66:16: reinstated

seems they are having path failures, but recovering.

A13 bios date is 12/08/2009

Johan Guldmyr
Honored Contributor

Re: BL465 G1 VC-FC

it's fabriclog -s on newer FOS
Andrew Hellyer
Advisor

Re: BL465 G1 VC-FC

Nothing showing up for the dates when the failures occured. There are no events on the VC either on the same dates/times relating to the path failure.
Johan Guldmyr
Honored Contributor

Re: BL465 G1 VC-FC

Hm..

Do you have a free slot in a c-class enclosure? Would you be able to try the blade somewhere else? Maybe so that it uses the other VC SAN module?

Andrew Hellyer
Advisor

Re: BL465 G1 VC-FC

yes! we just did that last night as the server lost 3 of its disks which it hasnt done in a while. I discovered that 5 of the 8 servers were all using the same VC module, and same uplink port. the other three servers were using the other one, but of the other three, only two use SAN disk, and 1 is fairly dormant.

Will see if this solves it - as it would indicate a load issue.
Also ran a brocade health report while it failed which hasnt happened before, so we will be able to see on the core switches the port utilisation of the VC uplink ports.

fingers crossed