[Koetje]

I hope anyone can help me out ...

I have 2 vio servers on 1 P550 for MPIO. and 1 Client LPAR
I installed 2 scsi controllers with 4 hd's each for rootvg and rootvg of the lpar's (through virtual device sharing and mirroring)
The data is put on a HP san.
So far i have setup 1 lun and enabled MPIO on the client which works now.

So far so good

Then i had to install another lun, and everything worked fine till i wanted to create the second virtual target device with mkvdev.
So on vioserver0 i made a virtual target device and it says available
on vioserver1 i create the virtual target device and it fails with:

some error messages may contain invalid information for the Virtual I/O Server environment

Method error (/usr/lib/methods/cfg_vt_scdisk) :
0514-047 Cannot access a device.

If i look on the lpar where the disk is for, i see a hardisk with 1 path wich is basicly what i configured.

Now i googled on it and people keep saying that you need to change the reserve_policy but the parameter doesnt excist on the Hdisk atribute.

It doesnt exsist on the hardisk i setup before too but those work with MPIO and show up on the client LPAR with 2 paths

Anyone any ideas what is going on?

p.s. if i delete the disk on vioserver0 and create it on vioserver1 first the creation on vioserver0 fails so seems like it locks the disk if you make a virtual target device

[sha012]

I am very sure that the reserve policy of hdisk on vio is causing your problem.

Which command are you using on vio server to display the hdisk attributes?...

The following is on one of my vio server...and I see the "reserve_policy " attribute being displayed on the hdisk6.

$ lsdev -attr -dev hdisk6
attribute value description user_settable
PCM PCM/friend/fcpother Path Control Module False
algorithm fail_over Algorithm True
clr_q no Device CLEARS its Queue on error True
dist_err_pcnt 0 Distributed Error Percentage True
dist_tw_width 50 Distributed Error Sample Time True
hcheck_cmd test_unit_rdy Health Check Command True
hcheck_interval 0 Health Check Interval True
hcheck_mode nonactive Health Check Mode True
location Location Label True
lun_id 0x551c000000000000 Logical Unit Number ID False
max_transfer 0x40000 Maximum TRANSFER Size True
node_name 0x5005076300c00cf7 FC Node Name False
pvid 00c3eece2b5ef56c0000000000000000 Physical volume identifier False
q_err yes Use QERR bit True
q_type simple Queuing TYPE True
queue_depth 1 Queue DEPTH True
reassign_to 120 REASSIGN time out value True
reserve_policy no_reserve Reserve Policy True
rw_timeout 30 READ/WRITE time out value True
scsi_id 0x660007 SCSI ID False
start_timeout 60 START unit time out value True
ww_name 0x5005076300d00cf7 FC World Wide Name False

[Koetje]

$ lsdev -attr -dev hdisk6
attribute value description user_settable

clr_q no Device CLEARS its Queue on error True
location Location Label True
lun_id 0x2000000000000 Logical Unit Number ID False
max_transfer 0x40000 Maximum TRANSFER Size True
node_name 0x50001fe15009e820 FC Node Name False
pvid 000c1750f6c3b3e30000000000000000 Physical volume identifier False
q_err yes Use QERR bit True
q_type simple Queuing TYPE True
queue_depth 1 Queue DEPTH True
reassign_to 120 REASSIGN time out value True
rw_timeout 30 READ/WRITE time out value True
scsi_id 0x650200 SCSI ID False
start_timeout 60 START unit time out value True
ww_name 0x50001fe15009e829 FC World Wide Name False

This is my output. As you see i dont have the option on the disk so when i wanna set it to no_reserve it says the paramter isnt there and therefor not editable.
What i noticed was that the pvid paramater wasnt there either till i enabled it. So i first had to add the paramater and after that i enabled the id and got one.
This was on the lun that worked fine the same as the disk that is giving me problems
The disk/lun that is working with MPIO has the same values as the one above
The ouput of the same command on hdisk0 (which is a 'normal' scsi internal disk) i get the values like you had.

$ lsdev -attr -dev hdisk0
/.../
reserve_policy single_path
/.../

In my setup I have 8 disks (hdisk0 to hdisk3) on 2 scsi controllers, so every VIO server has 4 local disks and 1 hba to a san. The disks are used for the root dirs of the VIO server and the rootvg of the Logical Partitions, i mirror them to both VIO server for failover. The real data and applications are on a SAN storage through the lba's. So the disks localy in the server have the reserve_policy option and the disk presented by the san dont have the option.

any clues?

[sha012]

Hi,

few more things to check,

what is the storage software running on the VIO servers? do you remember installing AutoPath software from HP on vio servers?

if not, try installing it on one of the vio server and try running the command again to see if it makes any difference....

on VIO, HP disk can only be handled by AutoPath software. Aix Clients can acheive MPIO through this way.

for example:
EMC disk can be handled by either PowerPath or mpio(built in vio)
HDS disk can be handled by either HDLM or mpio(built in vio)

[Koetje]

Hi thanks for your reply,

I never installed any software from HP. So Autopath isnt installed on the VIO.
This should propably be the problem.

I dont know what it exactly does but it sounds like something you need to have multiple paths....
I have 1 disk working with MPIO on the client partition, so both VIO servers have 2 disks from wich i enabled 1 through mkvdev to the client LPAR.
This is propably also the reason i see 2 disks on 1 lun (and 4 disks on 2 lun's)
2 x VIO disk => 1 Disk on client LPAR that runs MPIO.

The problem is with the second diskset ...

What i'm gonna do is install the software if I can get my hands on it and give that a try. If its the sollution, i will tell you guys here so other people can use it as a reference... And i stick around this forum as its nice to be among people with the same AIX bug in their head :P

Thanks in advance for helping

[sha012]

can you send me the following from the vio server?

lspv
lsdev -type disk
lsmap -all
lspath
lsdev -vpd -dev hdisk##


onthe client

lsdev -Cc disk
lsdev -Cc adapter
lspath
lsattr -El hdisk##

[Koetje]

I have put it in a .txt file and attached it as it was quite some data.

I had a few days off so i replied a bit late. But having off does goed things to your brain
So what i wanna ask ...

I use the same virtual scsi controller for both the virtual hardisk san_jde_data which can be MPIO'd

and for the san_ver_data, which cant be mpio'd and gets the error the moment i made the second virtual device on the second VIOServer