problem with EVA/Cluster

[meekrob]

Hi All,

i am having the following problem:
we have an HP-UX clustered system consisting of 2 rp3440 servers with an EVA 3000.
The fact is that morning the system showed malfunctions and the disks on the bays could not be detected: #ioscan -fnC disk only shows internal disks directly connected to the server.
Any help is much appreciated.

Regards

[g3jza · Accepted solution]

Hi,
Try to check status of your fc cards:
#ioscan -fnCfc to find DSF of your fc cards.

then for each FC port: #fcmsutil /dev/fcdX [fclpX]

is the status Online?

Also check status of the EVA array

[g3jza]

Check also /var/adm/syslog/syslog.log on each cluster node. Is there anything interesting?

[meekrob]

Hi g3jza and 10x for your prompt reply.
regarding the following command:
#fcmsutil /dev/fcdX [fclpX]

do i replace the "/dev/fcdX [fclpX]" in the upper line with the block DSF of each FC card?

Thanks in advance

[g3jza]

Hi,
yes,exactly

one example from my system:

:/root# ioscan -fnCfc
Class I H/W Path Driver S/W State H/W Type Description
=================================================================
fc 0 8/0/13/0/0/0/0/4/0/0/0 fclp CLAIMED INTERFACE HP AD221-60001 PCIe Fibre Channel 1-port 4Gb FC/1-port 1000B-T Combo Adapter
/dev/fclp0
fc 1 10/0/13/0/0/0/0/4/0/0/0 fclp CLAIMED INTERFACE HP AD222-60001 PCIe Fibre Channel 2-port 4Gb FC/2-port 1000B-T Combo Adapter
/dev/fclp1
fc 2 10/0/13/0/0/0/0/4/0/0/1 fclp CLAIMED INTERFACE HP AD222-60001 PCIe Fibre Channel 2-port 4Gb FC/2-port 1000B-T Combo Adapter
/dev/fclp2

:/root# fcmsutil /dev/fclp0

Vendor ID is = 0x10df
Device ID is = 0xfe00
PCI Sub-system Vendor ID is = 0x103c
PCI Sub-system ID is = 0x7042
Chip version = 2
Firmware Version = 2.72A2 SLI-3 (Z3D2.72A2)
EFI Version = ZE3.21A3
EFI Boot = ENABLED
Driver-Firmware Dump Available = NO
Driver-Firmware Dump Timestamp = N/A
Topology = PTTOPT_FABRIC
Link Speed = 4Gb
Local N_Port_id is = 0x8eea40
Previous N_Port_id is = None
N_Port Node World Wide Name = 0x50014380033a9445
N_Port Port World Wide Name = 0x50014380033a9444
Switch Port World Wide Name = 0x205600051eada302
Switch Node World Wide Name = 0x100000051eada302
Driver state = ONLINE
Hardware Path is = 8/0/13/0/0/0/0/4/0/0/0
Maximum Frame Size = 2048
Driver Version = @(#) FCLP: PCIe Fibre Channel driver (FibrChanl-02), B.11.31.0909, Jun 5 2009, FCLP_IFC (3,2)

[meekrob]

Hi,

apparently the management server which manages the EVA 3000 (HSV 100) is showing problems. Could that be the reason for the cluster showing malfunctions and the nodes not detecting the disks on the bay ??
how should i proceed in that case?

Thanks in advance

[g3jza]

I'm not an storage expert, but what does the status of your EVA says? Any logs etc.?

And what's the status of your FC connection from the server to the EVA array?
#fcmsutil /dev/fc....

What's the topology of your SAN? switched fabric, point-to-point? You should also check the status of your san-switches, if they are being used.

But first give me the fcmsutil /dev/fc... output.

Do you have any person responsible directly for storage? Might need to check with storage/SAN team....

[Pramod M]

Hi,

Check the Hardware status of the EVA3000 through management Serer Command View EVA. Is there any warnings on the Hardware lists? Is there any warnings on the VDISKs section? Is there any amber warning LEDs on controler / disks?

Thanks
Pramod

[likid0]

From the server point of view, your HBAs are linking ok:

Topology = PTTOPT_FABRIC
Link Speed = 4Gb

So you should check your fc switch/zonning/storage , for possible problems.

Once they are fixed, re-scan with ioscan again

[g3jza]

Daniel:
That was an example from my system. It was just to demonstrate how to get the output from his server :) .He still didn't provide his output...

[meekrob]

Hi All and many thanks for your support,
@ g3jza in fact i got a similar output as yours when i issue fcmsutil /dev/fcd0
however, i got a different output when doing fcmsutil /dev/fcd1 that mentions:
...........
Previous topology = UNINITIALIZED
Link Speed = UNKNOWN
Local N_Port_id is = None
Previous N_Port_id is = None
Driver state = AWAITING_LINK_UP
...........
and this is quite normal as the FC card connected to the host has 2 ports and 1 of them is up while the second is for redundancy.
One more thing i detected the following: both of the controllers were showing defective batteries does it cause all of this?? for the hosts not detecting the disks on the bay? ioscan -fnC disk on both hosts only detects internal disks.
In addition, i tried to create a virtual disk and present it to the management server (where i installed command view) the windows server does not detect it from disk management.

Any help is much appreciated

[Torsten.]

ioscan can see LUNs of an EVA, but never disks.

If both batteries in the EVA are failed, the array just stops to present LUNs, so you cannot see the LUNs anymore. Replace the cache batteries!

[Pramod M]

Hi,
Go ahead with replacing cache batteries with amber LED. Below is one document for the procedure.

http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?lang=en&cc=us&taskId=125&prodSeriesId=321347&prodTypeId=12169&prodSeriesId=321347&objectID=c00024681

Hope this will help you.

Thanks
Pramod

[meekrob]

Thanks for all of you
We proceeded by ordering replacement batteries; meanwhile thread closed.