AIX New Harddrive Configuration

Install ISCSI on AIX, we see the harddrives in lsdev -Cc disk
The drives are presented from a netapp appliance to AIX 6.1 server
hdisk2-5 were created.
 
What are the steps after that so that we can mount the drives,
do I need to create a VG? when I have tried mount //dev/hdisk2 /test,
it will error about about not having a logfile, if I give it a log file via the -o
option, it fails on mount i/o.
 
I have looked at smit, but not sure of the VG options, create normal, big.
are these what needs to be done next?
mkvg -y  iscsi_vg hdisk1
mklv -t jfs2log -y logiscsi_vg iscsi_vg 1
mklv -t jfs2 -y iscsilv iscsi_vg 64G
then make file system

What I am worried about is overwriting any data on the Netapp drives, I am told
there is data there.

Thanks
FITFSCAsked:
Who is Participating?

Improve company productivity with a Business Account.Sign Up

x
 
woolmilkporcConnect With a Mentor Commented:
Hi,

creating an AIX volume group on a hdisk or on several hdisks always destroys the data on the disk(s).
There is no difference between hdisks accessed via iSCSI and locally attached disks or disks accessed via SAN.

If there is data on the disks there must be or have been a host machine owning/managing them!
In other words: Are these disks already attached to a different host?
If so, and if you want to share the data with that host you can't use iSCSI, you must revert to some file sharing protocol like NFS instead.

If there is no other host (anymore) you can access the disks without destroying the data only if there is an AIX VG already present on them.
In this case you can use "importvg" to make this VG known to AIX. Use "man importvg" to see how it works.
Please take note that all hdisks making up the VG should be accessible by the new AIX system.

By the way, "mkvg" only destroys data on the LUNs presented to AIX. The remainder of your NetApp box will of course be left unchanged!

Anyway, the commands you mentioned in your post are nearly OK. Some remarks/corrections:

1)
- You should use "mkvg -S" to create a "scalable" VG. This will better allow for future growth.
- You can add all hdisks to the new VG in one go, not only one. In your example you use "hdisk1",
but in your Q you stated "hdisk2-5 were created". Is this a mistake or should hdisk1 become part of the new VG as well?
Assuming a mistake the command should be: mkvg -S -y iscsi_vg hdisk2 hdisk3 hdisk4 hdisk5

2)
The command to define the log volume is correct, but you must format it afterwards:
mklv -t jfs2log -y logiscsi_vg iscsi_vg 1
logform /dev/logiscsi_vg
Answer "y" to the confirmation prompt.

3)
- AIX formats its hdisks in chunks called "physical partitions". "mklv" expects the number of PPs to specify the size, not GB. Find out
which PPsize you have with "lsvg iscsi_vg". Look for "PP size" in the upper right area.
Assuming a PP size of 128 MB the "mklv" command should be:
mklv -t jfs2 -y iscsilv iscsivg 512

Please note that "mklv" can take hdisk names as parameters, so you  could control on which hdisks(s) the LV should be placed.

4)
- Use "crfs" to create a filesystem.
crfs -v jfs2 -d iscsilv -m /mount/point
mount /mount/point


Note: you can use "smitty storage" instead of the command line for all of the above tasks.

Good luck!

wmp




 
0
 
FITFSCAuthor Commented:
the hdisk are presented from a netapp box that was configured offsite and sent to our location.
I used lqueryvg -Atp hdisk2 and I can see that there are file systems on the various h/d. I was looking
for any information regarding previous VG, but did not see any i will try the importvg.

Thanks

0
 
woolmilkporcConnect With a Mentor Commented:
Good luck!

Take care to have all hdisks available whose PVIDs show up in "lqueryvg -Atp ..." under "Physical:"
Check with "lspv"!

Please note too that importvg will change the name of a logical volume if this name already exists!

Consider using the "-n" flag of importvg to not varyon the VG after importing. This way you can examine /etc/filesystems beforehand to check what filesystems came with the new VG. Making a backup copy of /etc/filesystems could also be a good idea.

wmp
0
 
FITFSCAuthor Commented:
closing,,, the issue turned out is that the LUNs were a "live" replication from another site, so, AIX was unable to mount them and considered the VID as corrupt.
 
0
 
FITFSCAuthor Commented:
provided good help so awarding points, but the final solution was that the LUN was open by another process.
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.