cancel
Showing results for 
Search instead for 
Did you mean: 
Highlighted

missing disks and reboot wont solve it

I am very new to veritas. We have AIX 7.1 server using veritas DMP. When I look at the VIO all of the virtual fibre channel adapters are logged in, but on the lpar it is failing to see any disks on fscsi0 and fscsi1. I have been going back and forth with IBM and symantec and cannot get this resolved, so decided to pick your brains here.

# lsdev| grep fscsi
fscsi0          Available 01-T1-01    FC SCSI I/O Controller Protocol Device
fscsi1          Available 02-T1-01    FC SCSI I/O Controller Protocol Device
fscsi2          Available 03-T1-01    FC SCSI I/O Controller Protocol Device
fscsi3          Available 04-T1-01    FC SCSI I/O Controller Protocol Device
fscsi4          Available 05-T1-01    FC SCSI I/O Controller Protocol Device
fscsi5          Available 06-T1-01    FC SCSI I/O Controller Protocol Device
fscsi6          Available 07-T1-01    FC SCSI I/O Controller Protocol Device
fscsi7          Available 08-T1-01    FC SCSI I/O Controller Protocol Device


# vxdmpadm listctlr
CTLR_NAME       ENCLR_TYPE      STATE        ENCLR_NAME      PATH_COUNT
=========================================================================
fscsi2          Hitachi_VSP     ENABLED      hitachi_vsp0         44
fscsi3          Hitachi_VSP     ENABLED      hitachi_vsp0         44
fscsi4          Hitachi_VSP     ENABLED      hitachi_vsp0         44
fscsi5          Hitachi_VSP     ENABLED      hitachi_vsp0         44
fscsi6          Hitachi_VSP     ENABLED      hitachi_vsp0         44
fscsi7          Hitachi_VSP     ENABLED      hitachi_vsp0         44

^ Above you will see that fscsiX seen by OS is not being seen by veritas.

How can I force them into veritas? I have already tried rebooting the VIO and LPAR and that doesnt seem to help. FWIW, i deleted the disks that were in defined state. Usually when MPIO is being used and we lose path, deleting the disks and virtual fibrechannel adapter and running cfgmgr solves the issue, but that doesnt seem to help here.

 

1 Solution

Accepted Solutions
Highlighted
Accepted Solution!

Hi,  First you should confirm

 

Hi, 

First you should confirm the disk array in HCL

 

http://www.symantec.com/business/support/index?page=content&id=TECH211575

 

Second you should confirm the configuration in the disk array should follow hardware notes:

 

http://www.symantec.com/business/support/index?page=content&id=TECH47728

 

Many requirement settings need to do , like on AIX :

 

1.Enable fail_fast on all FC initiators:

chdev -l fscsix -a fc_err_recov=fast_fail -P where x is the specific HBA (fscsi0, fscsi1, and so forth).

 

2. Enable Dynamic Tracking of FC devices for all FC initiators :

chdev -l fscsix -a dyntrk=yes -P where x is the specific HBA (fscsi0, fscsi1, and so forth)

 

3.Disable MPIO if it is enabled for certain devices:

Install the vendor specified ODM, referring to the steps provided by the vendor. (Reboots may be required)

 

4. If any path to the target disk has SCSI reserve ODM attribute set (reserve_policy or reserve_lock), then change the attributes to release the SCSI reservation from the paths, in the event to run DMP in VIOS, the DMP device will be used in cluster set-up having fencing enabled, or the shared DMP device will be used as  SANboot :

 chdev –l hdiskx –a reserve_policy=no_reserve –P Or chdev –l hdiskx –a reserve_lock=no –P Reboot the system for the changes to take effect. 

 

 

...

 

View solution in original post

1 Reply
Highlighted
Accepted Solution!

Hi,  First you should confirm

 

Hi, 

First you should confirm the disk array in HCL

 

http://www.symantec.com/business/support/index?page=content&id=TECH211575

 

Second you should confirm the configuration in the disk array should follow hardware notes:

 

http://www.symantec.com/business/support/index?page=content&id=TECH47728

 

Many requirement settings need to do , like on AIX :

 

1.Enable fail_fast on all FC initiators:

chdev -l fscsix -a fc_err_recov=fast_fail -P where x is the specific HBA (fscsi0, fscsi1, and so forth).

 

2. Enable Dynamic Tracking of FC devices for all FC initiators :

chdev -l fscsix -a dyntrk=yes -P where x is the specific HBA (fscsi0, fscsi1, and so forth)

 

3.Disable MPIO if it is enabled for certain devices:

Install the vendor specified ODM, referring to the steps provided by the vendor. (Reboots may be required)

 

4. If any path to the target disk has SCSI reserve ODM attribute set (reserve_policy or reserve_lock), then change the attributes to release the SCSI reservation from the paths, in the event to run DMP in VIOS, the DMP device will be used in cluster set-up having fencing enabled, or the shared DMP device will be used as  SANboot :

 chdev –l hdiskx –a reserve_policy=no_reserve –P Or chdev –l hdiskx –a reserve_lock=no –P Reboot the system for the changes to take effect. 

 

 

...

 

View solution in original post