QLA2300 wierdness

[Zaphoid1970]

Greetings:

We're having some very, very odd problems as we try to migrate our SLES9 machines off of a VA7400 on to an XP10,000, then eventually on to an MSA1500... We had a QLogic QLA2200 adapter in the machine, and were having problems with a PVMove... We were eventually told by HP support that QLA2200s are not supported with the XP10K, so we put a QLA2310 in the machine in place of the 2200... Now, the 200 could see BOTH the VA LUNs (0:0, 0:96, and 0:97) and the XP LUNs (1:0 and 1:9)... The 2300 can see the XP LUNs, but not the VA LUNs... We also get, in DMESG the following error:

scsi: host 0 channel 0 id 0 lun16384 has a LUN larger than allowed by the host adapter
scsi: host 0 channel 0 id 0 lun16480 has a LUN larger than allowed by the host adapter
scsi: host 0 channel 0 id 0 lun16481 has a LUN larger than allowed by the host adapter

I find it odd that a LUN can come up with such a large number... The VA doesn't even support LUN numbers that large!

This is apparently a documented, but not fixed problem... HAs anyone on the forums run into this, and come up with a fix???...

We're running an HPDL360 with SLES9, SP3...

Thanks,
Richard.

[Steven E. Protter]

Shalom Richard,

Your method of moving the data is not supported.

I would suggest a simpler approach that involves backing your data up to tape and restoring it after you set up nw logical volumes and luns on the target environment.

SEP

[James A. Donovan]

Did you ever figure out a solution to why the VA7400 LUNS are not being recognized properly?

I am seeing the exact same problem.

I have a Dell 1750 with two QLogic 2340 cards running RedHat 4.

scsi: unknown device type 12
Vendor: HP Model: A6189A Rev: HP20
Type: RAID ANSI SCSI revision: 03
qla2300 0000:01:04.0: scsi(4:0:0:0): Enabled tagged queuing, queue depth 32.
Attached scsi generic sg3 at scsi4, channel 0, id 0, lun 0, type 12
scsi: host 4 channel 0 id 0 lun16384 has a LUN larger than allowed by the host adapter
scsi: host 4 channel 0 id 0 lun16395 has a LUN larger than allowed by the host adapter
scsi: host 4 channel 0 id 0 lun16396 has a LUN larger than allowed by the host adapter
scsi: host 4 channel 0 id 0 lun16408 has a LUN larger than allowed by the host adapter
scsi: host 4 channel 0 id 0 lun16411 has a LUN larger than allowed by the host adapter

[IT Response]

A possible fix would be check what the array default host port behavior is set to. If both controllers are set to HPUX or windows for example, Linux os may not be able to see the array correctly.

You can see the contents of the armhost file from the array with command armhost -r -f filename
read the file filename and see if wwn of linux hba have entry in this table. If not Edited the file and added the node WWN Linux for each HBA add and then write file back to array armhost -w -f filename


Rebooted the Linux server and see if luns are reporting.

[Seth Kaplan]

I'd be curious to know if anyone's come up with a resolution for this. We're having the same issue w/RHEL4, Qlogic QLA234 HBA's on an HP DL585. Perhaps I should start a new thread?

thanks,
Seth

[James A. Donovan]

Creating the host port behavior file using the armhost command fixed this issue for me.

Create a file with entries like so:

200000e08b8ff43d Linux
200000e08b8fdf3b Linux
200000e08b8ff93e Linux
200000e08b8f5a3e Linux
200000e08b8ffb3e Linux
200000e08b8fee3d Linux

One entry per WWN. The default behavior for the VA7400 is "HP-UX", so you need to tell it that the operating system for the other cards is something else.

....rant on....
....It only took 72 hours for HP Support to finally come back with such a simple "fix" from the time I first called them about my problem....and only after escalating the call with both HP and QLogic support. You'd think something solved so simply would have been the first thing HP asked me to try, rather than telling me to call QLogic and RedHat for support....
....rant off....

[Seth Kaplan]

James,

thanks for your reply. I'm exploring this option. This is the first Linux server I'm adding to our SAN. Presently we've been homogeneous with HP-UX only. I currently have no entries list in the host port behaviour table, I presume that's typical. I'm being cautious as I don't want to interrupt our production servers. I've opened a call with HP but am also experiencing slowness in their response. RedHat's response to my problem was to reduce the Lun size. Thanks RedHat. I have to say I like the simplicity of dealing with one vendor for hardware and OS. I guess I got soft doing HP-UX.

Seth