IBM Support

How to change reserve policy on one of two VIO Servers

Question & Answer


Question

The primary VIOS has the reserve policy set to no_reserve, and the secondary VIOS has it set to single path. The LUN is exported as a backing device. How do I change reserve policy to no_reserve on the secondary VIOS? This document applies to VIOS 2.x

Cause

DUAL VIOS with VIOC using virtual SCSI MPIO redundancy and reserve policies on VIOS are different.

Example:
LUNs on VIOS1 have the reserve policy set to "no_reserve", but LUNs on vios2 have it to "single_path" and need to know what to do to change them to "no_reserve".

Answer

In the following example, the VIO (AIX) Client has virtual SCSI disks, hdisk11 and hdisk12 that are part of a volume group called clientvg. Their corresponding backing devices (SAN disks) are known as hdisk22 and hdisk23 on the second VIOS (VIOS2).

On the VIO CLIENT, as root:

1. Ensure you have an up-to-date backup of clientvg BEFORE you do anything else.



2. Unmount all filesystems on clientvg and deactivate the volume group:
    # unmount /fsname
    # varyoffvg clientvg

3. Put target hdisk (for this example hdisk11 & 12) in a "Defined" state.
    # rmdev -l hdisk11
    (repeat for hdisk12)

    NOTE: Another option is power off the VIO Client.

On VIOS2 (where you need to change reserve policy to "no_reserve") as padmin user:

4. Remove the VTD for target hdisk (in this example hdisk22 & 23).
    $ rmvdev -vtd <VTDname>

To find out VTDnames run
    $ lsmap -vadapter vhost0
    (where vhost0 is the SCSI Server Adapter hdisk22 & 23 are exported through)
    =>look for VTD line associated with backing device hdisk22 & 23

5. Change the reserve policy.
    $ lsdev -dev hdisk22 -attr =>reserve_policy = single_path is expected
    $ chdev -dev hdisk22 -attr reserve_policy=no_reserve
    $ lsdev -dev hdisk22 -attr =>reserve_policy should now be set to no_reserve
    (repeat this for hdisk23)

6. Recreate VTDs for hdisk22 & 23
    $ mkvdev -vdev hdisk22 -vadapter vhost0 -dev <VTDname>
    (repeat for hdisk23)

NOTE: Another option is to put the VTD(s) in a "Defined" state before changing the reserve policy, and then make "Available" after the change has been made, instead or removing and recreating.


ON VIO CLIENT, as root:

7. At this point, bring hdisk11 & 12 back to "Available" state

    # cfgmgr
    # lsdev -Cc disk =>hdisk11 & 12 should be back available

8. Reactivate the volume group and remount the filesystem(s).
    # varyonvg clientvg
    # mount /fsname

    (or activate the VIO Client if it was shutdown)

[{"Product":{"code":"SSPHKW","label":"PowerVM Virtual I\/O Server"},"Business Unit":{"code":"BU058","label":"IBM Infrastructure w\/TPS"},"Component":"Not Applicable","Platform":[{"code":"PF002","label":"AIX"},{"code":"","label":"Other"}],"Version":"2.1.0.x-FP20.x;2.1.1.x-FP21.x;2.1.2.x-FP22.x","Edition":"Enterprise;Standard","Line of Business":{"code":"LOB57","label":"Power"}}]

Document Information

Modified date:
19 February 2022

UID

isg3T1010976