Using Hitachi disk with Native AIX mpio ( no HDLM )

1.  Check MPIO filesets available:

lslpp -l devices.common.IBM.mpio.rte

2. Get Hitachi Specific MPIO ODM Updates from

https://tuf.hds.com/gsc/bin/view/Main/AIXODMUpdates

Collect MPIO Base and updates: List below current at 2 Feb 2013

https://tuf.hds.com/wiki/pub/Main/AIXODMUpdates/aix_odm_5400.zip

https://tuf.hds.com/wiki/pub/Main/AIXODMUpdates/aix_odm_5401U.zip

https://tuf.hds.com/wiki/pub/Main/AIXODMUpdates/aix_odm_5402U.zip

https://tuf.hds.com/wiki/pub/Main/AIXODMUpdates/aix_odm_5403U.zip

https://tuf.hds.com/wiki/pub/Main/AIXODMUpdates/aix_odm_5404U.zip

 

1)      Unzip all files and move *.tar and *U to your target server in path /usr/sys/inst.images. 

2)      cd /usr/sys/inst.images

3)      extract tar file

tar xvf *tar

4)       inutoc .

5)      Remove packages Hitachi.aix.support.rte and its dependent package Hitachi.hacmp.support.rte  which relates to no MPIO driver ( e.g. HDLM  or PowerPath ).  These are incompatible with native MPIO driver for HDS disks.  Make sure to change remove dependent software to yes or action will fail

6)       smitty install: Select and install the software in /usr/sys/inst.images

7)      check installed filesets

8)  if any Hitachi disks have been detected already, remove them with rmdev -dRl hdisk

9)  Reboot server at soonest opportunity

10)  Check for detected MPIO disks: "lsdev -Cc disk"

11)  Check that 2 or more paths are seen to the Hitachi disk: "lspath" 

12)  can also check paths and disk sizes by

echo “ No hdiskxx size mb No-paths”

p="/usr/sbin/lspath";for i in `$p| awk ' !/Missing/ {print $2}'|sort|uniq `;do echo "$i; `getconf DISK_SIZE /dev/$i` mb; `$p| awk ' !/Missing/ &&/'$i' / {print $2}'|wc -l|sed 's/ //g'`" ; done|cat -n

      13)  Default setup of multipath is failover, reserved single path, qdepth=2. you can check this with lsattr -EL hdisk4 for example. 

     14) To configure round-robin with queue depth of 8 we need to change each disk as follows:

              chdev -l hdisk4 -areserve_policy=no_reserve -a algorithm=round_robin -a queue_depth=8

  This will also have to be done for any new disks that get added.

15)  Validate that new settings are applied using lsattr -EL hdisk

16)  Thats it.  You can build your volume groups and filesystems as needed

17)  you can check to ensure data is going down both fibre adapters by running "iostat -a | grep fcs"