Extending a VG in an AutoRAID

[Joe Profaizer]

I have a VG that is 100GB. It has 27495 Total PE and 25000 allocated PE. Is there any way to extend this volume group to 200GB? I'm not sure how to add another disk to the volume group since it's an AutoRAID. Currently it has to Physical volumes, /dev/dsk/c4t1d1 and /dev/dsk/c6t0d1 (which is the Alternate Link).

Please advise.

Thanks,.

..Joe

[Denver Osborn]

The AutoRAID wouldn't make this vg any different. To make the VG 200GB you could either add another 100GB worth of disks to the volume group are recreate the volume group with a 200GB disk.

[Joe Profaizer]

I removed a VG that was 100GB just so that I could make this one 200GB. I've made a little progress, but the most it would extent to was 134GB. Is there a limitation?

..Joe

[Carlo Henrico_1]

Are you using SAM or command line LVM commands? SAM immediately shows you how much is used/available as well as physical and logical volumes. This should then also help you well on your way to extend the VG.

Carlo

[Dave Seibel]

I have an N4000 with an AutoRAID 12H here that I currently babysit. Ours is 12x18.2Gb drives with 6x20Gb LUNS, nothing near the size that you are talkin'. The readers here will need to know what your AutoRAID situation is to help you. It sounds like you will need to enlarge a current LUN or create a new one and add it to the existing volume group. There is an 8 LUN limit with AutoRAID and you do not want to take up all the disk space on the AutoRAID with LUNS. Your performance will suffer BIGTIME. We have 220Gb of disk (about) and will not go over 180Gb of space used for LUNS to assure that we stay in Raid 0/1 as much as possible.

Good Luck

Dave

[Joseph C. Denman]

Sounds like you need to extend your existing vg to include the lun that was used by the vg you removed? I am unclear on all the facts.

Hope this helps.


...jcd...

[Bill McNAMARA]

Create another LUN of 100G on the AutoRAID.. let's say it ends up LUN 3, your device files
would be /dev/dsk/c4t1d3 and /dev/dsk/c6t0d3 based on your earlier example.
Then

# pvcreate /dev/rdsk/c4t1d3

(no need to pvcreate on the alt link..
it's the same disk!)
If your VG is vgauto and the primary path to the other pv is through c4t1d1 you should not
bottleneck c4 with LUN3s traffic so extend the vg this way

# vgextend vgauto /dev/dsk/c6t0d3 /dev/dsk/c4t1d3


# strings /etc/lvmtab

should look something like this

/dev/vgauto
/dev/dsk/c6t0d1
/dev/dsk/c4t1d1
/dev/dsk/c6t0d3
/dev/dsk/c4t1d3


Now you can stripe your lvs across the LUNs.

# lvcreate -i 2 -I 64 -n mylvol /dev/vgauto
# lvextend -L 400 /dev/vgauto/mylvol


Later,
Bill

[Thierry Poels_1]

Hi Joe,

I'm afraid that your 134GB limitation is caused by the "maximum physical extent" parameter (vgcreate -e n ).

The bad thing is that there is no way to increase this setting. You will need to drop and recreate the volume group.

For your sake I hope I'm wrong.

Thierry.

[Bill McNAMARA_1]

I don't think so Thierry,
If you look at his PE # he's just running
out of unallocate free space.

[Thierry Poels_1]

hmmmz,
- he HAD a 100GB VG (25000 PE of 27495 PE used)
- 100 GB diskspace was added (from dropped VG)
- maximum extendable to 134 GB ?!?

I still have a hunch I'm right.
Please show us some vgdisplay output Joe.

Thierry.

[Bill McNAMARA_1]

I believe that the PE numbers will increase when he vgextends the new Lun, but he won't be able to change the extent size, default 4M.
Without recreating.

Yea, send the vgdisplay anyway, it'll do no harm.

Bill

[Thierry Poels_1]

hmmmz (2)

vgcreate -e defines the maximum number of physical extents possible per disk in the volume group, NOT for the whole VG.

is there an option to delete some replies in this forum ? ;-)
Thierry.