Space planning

[taylorb]

I know this is more theory than technical, but how do you all go about planning space requirements?  It seems I am always in the stage of panic as I run out, and I want to get to a more pro-active state.  But it seems like no matter how much space I add, it is gone in a flash.  Any ideas or resources on how to stay on top, or is this just the nature of the beast these days?  Do I just always tell management I need 5x what I think I need and hope they approve it? :smileygrin:

[vmroyale]

Hello.

Can you detail more about your current strategy?  Are you using thin provisioning?  I tend to be very stingy with storage and will contest requests that I think are wasteful, but that is not always an option.  With that being said, are your consumers really using up all of the storage, or are they wasting it?

Good Luck!

[AndreTheGiant]

Also which monitor tools are you using?

Andre

[taylorb]

Current strategy?

"Oh %$#% we are out of space!  Let's buy more before something bad happens!" 

I am sure the users are wasting space, but it is hard to tell them to stop making data or that they don't really need another application or product.  Every user's data is extremely critical if you ask them.   You know how it goes.    Plus, it is not like the departments tell me they need additional resources until they need it "ASAP", so it becomes hard to plan.   I guess I was wondering what others are doing to stay on top.

I'm not using any special monitoring tools, just what the Vmware performance tab gives me and basic OS level monitoring.  

[DSTAVERT]

You might waqnt to consider some form of chargeback. When there is budget money involved some projects loose steam and aren't quite so critical anymore.

[vmroyale]

Current strategy?

"Oh %$#% we are out of space!  Let's buy more before something bad happens!" 

:smileylaugh::smileylaugh:

Probably the best you can do is to start keeping tabs on the growth and try to use that data to justify spending ahead of time.  "Dept X requested Y GB of storage last year, and based off of that we should do this for the future or something bad will happen" type of stuff.  It is a bit of work, but getting out into the trenches and talking to the users about what they have coming up can also be a good strategy.

Good Luck!

[bryansemple]

First, capacity management tools should give you some measure of storage utilization trending that you can use to show upper management and at least get ahead of the demand.

Second, reclamation of underutilized space and/or use thin provisioning.

Finally - to the other posts, come measure of chargeback or at least showback.  Formal chargeback is difficult to implement without the supporting business systems and executive sponsorship. You also get into  a tough spot trying to determine a chargeback value. But simply reporting back to senior management the consumption of storage not by GB, but by $ will have a big impact.  Reporting what has been allocated vs. actually utilized is key to highlight any wasteful end users and put a dollar value behind it.

I am not as familair with CapacityIQ, but that should be on your list as would our product - vOperations Suite -- Chargeback and Reporting Module and also the Optimizer.

[mcowger]

Even if you can't do full chargeback, 'showbacks' demontrating who is being obnoxious helps curb the growrth rate.

But over all, data grows.  In my last admin job, we generally had a 500GB / day overall data growth rate, and purchased arrays (and arrays and arrays) based on that!