Apologies if this is in the incorrect forum...
Anyway, I am experimenting in the EC2 cloud. I have (after considerable messing about...) got an 11.2 oracle database running with a pair of ASM diskgroups. I decided to try and increase the size of one of my disk groups.
I couldn't really find any "how-to's" on this subject so decided to jump right in. I took the following approach.
unmount asm diskgroup
detach the ebs volumes (there are 6 in my diskgroup).
snapshot the volumes to s3
create new bigger volumes from the snapshots taken above.
attatch the new volumes
-- Here, the plan was to remount the asm diskgroup and finish up.
HOWEVER...it didn't really happen that way.
ORA-15017: diskgroup "RECO" cannot be mounted
ORA-15063: ASM discovered an insufficient number of disks for diskgroup "RECO"
I messed a bit with oracleasm, checking it could see the disks etc, reloading it etc all to no avail.
I restarted the server and, lo and behold, all my diskgroups came back. With the new volumes. I was just about to congratulate myself when I noticed that the diskgroup was still the same size as before!!
I tried a resize of the disk group, but this didn't help either.
So, my questions are
1) How do I increase the size of the disks?
2) Is this approach valid at all?
3) Is there any documents anywhere which tell me exactly how I should be carrying out this sort of operation?
Thanks for your time.
Edited by: user13024707 on 05-Feb-2011 08:44
Well I guess I really am on my own here...however, I can contribute further.
In order to get ASM to see the disks as the correct size, we needed to actually increase the size of the partition. This was done by using fdisk, once this has been done and an "alter diskgroup blah resize all;" the correct sizes were reported.
Having now succeeded in doing this I am still not sure that this is the "right" way? Any comments, better ideas more than welcome.
This may be too late but here goes...
An ASM disk group can have volumes that are different in size.
Add new larger volumes to the existing disk group.
drop old smaller volumes that you want to resize.
old volume is now out of ASM and can be delt with at OS level to make it what you want.
repeat until all old volumes are gone.
re-integrate new volumes after they have been rebuilt.
All with no outage and ASM preserves all of the data.
It works better if you have enough new larger volumes to take over all of the data from the old volumes where you can drop all old volumes at the same time.
Thanks for your reply.
I had considered your suggestion, though shied away because of the limited number of volumes that i can attatch at one time. Currently I have 15 EBS volumes and that the limit is 20 (I think...) being unable to do the transition in one go led me down the path I took.
I am only experimenting at the moment, so not too late, but it looks like in the future production increases will need to be done in the way you suggested.