SAN EVA 5000 with disk increase

[Mike Mager]

The upgrade of physical disks from 36.3 GB to 146 GB and increase in count from 14 drives to 16 increased the capacity by 450% going from 508.2 GB to 2,336 GB for each grouping of DATABASE and SYSAPPS per location. Each grouping is broken down to virtual disks and has data installed. There are two locations and one location at a time can be shut down: I unnderstand the commands such as:
SET VOLUME dsa0:[,...] /LIMIT=134217728
SET VOLUME dsa0:[,...] /SIZE=134217728

We are running 6.2 version of HP Management and already increased the size of the backup set since there is only one location for this. DATABASE and SYSAPPS are another kettle of fish though. How long should this process take to run with me having 16 disks per group per site? If I do a remote site and bring it back online, will the shadowmerge from the main site overwrite the sizing commands?

I have a 120 minute window with no activity on the system to complete one site if I have to do both before going live and can leave second site down for a day to perform the size change there. There are 13 virtual sets per site on both the BEP and FEP level.

[The Brit]

Mike,
The first question is "do you have dynamic volume expansion enabled??"

check the "Expansion Size Limit" on your volume.

Note this can only be set when the volume is mounted privately to your own process. By default, if you do not specify a value for /Limit it sets it to 1TB. So it should look like

Logical Volume Size 75497472 Expansion Size Limit 2147475456

(in my case, the disk is currently 36GB).

In any case, Expansion Size Limit must be greater than the new proposed size. If this is true, then

Step 1. - on the EVA increase the size of all LUNs (units) which make up the shadowset.
Step 2. - on all cluster nodes (best via sysman) set volume /size = new_size DSAxxx

It isn't clear from your post whether you shadow between sites (i.e. units at different sites), However the thing to keep in mind is that the settings on the mounted volume take precedence. So if you make the changes on your backup site, and then merge those units back into a (already mounted) shadowset, the values on the already mounted shadow set will prevail.

I would think that 2 hours would be plenty of time (although your copy/merge will take longer)

An alternative would be to have your backup drives shadowed in when you make the changes then they can be dismounted immediately afterwards.

Note that the LUN size changes can all be done ahead of time since this has no effect on your systems, data or apps.
-------------------------------

If DVE is not enabled, you can easily take care of it with a script (like the attachment). This will take ~2 minutes.

HTH

Dave.

[Mike Mager]

Thanks Dave,

We do not have "volume expansion enabled" so this will have to be done manually.

The multi-site is mirrored in real time as one large cluster minus one grouping. We had been using the 36GB disks and increased all disks to 146GB to make the system last until replacement in 2013.

If we bring down a site, we use shadowmerge to catch back up so doing one site at a time will not be an option then but will be required to have at least one site done before we "go live".

Thanks for the information regarding the EVA increase and noting that it can be done prior with no impact on the systems, data or apps.

Mike

[The Brit · Accepted solution]

just realized that my attachment has extension .com, so you cant open it to read.

Try again with extension .txt.

This script reads a file with one record for each shadow set. Record looks like

DSAnnn /Shad=($1$DGAxxx[,$1$DGAyyy[,$1$DGAzzz]]) []

(set can be 1, 2, or 3 units)

This process only needs to be followed once, on the first node to boot.

1. Log in as SYSTEM.
2. Dismount all of your volumes except obviously the system disk. Beware of open files (particularly secondary pagefiles.) which will stop your volumes dismounting.

If you have secondary pagefiles installed from other disks, you will probably have to either comment out the mount/install of these files and reboot, or reboot and halt the startup process before it gets to that.

Once all of your disks are dismounted, then

1. $ Mount DSAzz/Shad=($1$...,$1$...) Label
2. Check current Expansion limit
3. $ Set volume /Limit DSAzz
4. Check new Expansion Limit.
5. $ Dismount DSAzz

Once this is done for all volumes, you can now mount them normally. Always set the expansion limit to the max (i.e. do not give /Limit a value). This is the only way to increase the limit, so if you need to do it again and you have NOT allowed the limit to be large enough, you will have to do this all over again.

If you are ready to increase the size, you can probably set the new size as part of the procedure above, i.e. before dismount enter

$ Set Volume /size DSAzz

This will increase the size to the physical size of the LUN.


Dave

[Mike Mager]

I had opened it as a text document and it looks okay but prefer the one you gave in your second response. I am sure of the answer to the next one but was asked to confirm.

When you speak of the LUN size, we found in the virtual disk VMS we see the area labelled Capacity and noticed we can change the amounts (shown with 240), is this equal to the LUN area you mention?

Thank you again for your help.

Mike

[The Brit]

As I recall (we have moved on to XP stuff now), on any virtual disk, the Capacity window is at the bottom left. It (used to) show "Requested" size and "Current" size.

The requested size can be changed. Type in the required size, and click "Save Changes" at the top. A window will pop up telling you to be sure that your OS knows how to do dynamic expansion (or words to that effect).

If you proceed, the LUN will grow to the requested size and you will see the Current Size change.

The time taken is fairly quick (depending on how much increase you want and how many disks are growing simultaneously).

Dave

[Mike Mager]

Thank you.

I was sure of the answer, but powers that be wanted to be sure.

Mike