HPE Storage Users Group

Hi Guys,

I have upgraded my V400 from 2 nodes to 4 nodes and also added 320 disks to the system making a total 640 disks all of same type (FC).

Can someone suggest what is the best way to rebalance the system.
Has anyone tested tunesys for this.
The Inform version i am currently running on is 3.1.1 Mu1

I'm not on 3.1.1 yet, but wish I were for the tunesys command. We recently added drives to our T800s and I ended up writing a script that looped through all our VVs to tunevv them all.

Since you have tunesys available to you, that would be my first preference.

We have T400 and V400 on 3.1.1 MU1. Ran tunesys when we added new drives. It does balance out but may take a while and few iterations. Depends on how much data you currently have, it may be hours or days.

My last tunesys on V400 sitting at the DR site ran for 2 days. Current tunesys on T400 at the production site is 5 days in and still running. Hopefully be done by end of the week. Probably will have to run it again since there are so much data to distribute.

-Jason

Jason,

Did you enquire the support before rebalancing coz when i asked support they said it 'd be better to upgrade to 3.1.2 or 3.1.1 Mu2 p18 minimum (3.1.1 P13 is what i have currently)as they have taken out some bug which occurs while rebalancing. I have around 140 TB of data sitting on my box.

I tried a dry run which said i have to do 322 tune's for getting the data balanced.

Thanks for the update:Richard, Jason

Sunieal

We're on 3.1.1 MU1 P13 as well. I have not heard about a bug in regards with tunesys on that patch level. So far, we haven't had any issues with it.

Did the support tech provide a link about the tunesys bug?

Thanks
-Jason

I believe this is the link to the "bug".

http://h20000.www2.hp.com/bizsupport/Te ... ical_006_1

The way I read that, if you have plenty of free space, you should be OK. How much is plenty is the question. Even with VV based tunesys, you can watch the system consume new space as it moves the data from one CPG to another... then when done you can force a compaction on the old CPG, or if its empty, just delete it.

How many VV's do you have? It might be better to use tunecpg to move all your luns from the old CPGs to new ones to get the rebalance done. I posted a script I used to assist me in getting this done. viewtopic.php?f=17&t=215

If space is tight on the orginal node pairs and spindles, you may want to do a couple tunecpgs, followed by a compact CPG to free up the allocated space left behind to make room for more tunecpgs.

Nice find!

Read the advisory and I'm good. Bug affected 3.1.1.226 MU1 build.
We're on 3.1.1.342 MU1 build, which resolved the issue regards to tunesys.

Thanks
Jason

Thanks Richard,

We're on 3.1.1.342 MU1 build, and i have 140 VV's most of them are 1TB in size, and i have 200TB space sitting idle from the new disk + node upgrade.

So, enough resources are there to move the VV's, the only concern was the bug which is now off of my head, Thanks to you

Regards
Sunieal