not for traditional raid groups, for pools yes (especially the ones with FAST).

I'd say that this depends on your usage of the disks, and your data growth...

Filling any Raid Group to capacity could end up in a bad situation if your data i growing, and the RG running up to a 100%.

Normally I would say fill it up to around 70% and when you reach that point, start looking for more capacity.

With Raid Groups in Pools with FAT VP enabled, you need to have som free space to move data from tier to tier and to rebalance data across the Tier....but then in pools you will fiond it very difficult to monitor where the data is on the Raid groups in a tier.

Jim_Hegner wrote: I'd say that this depends on your usage of the disks, and your data growth... Filling any Raid Group to capacity could end up in a bad situation if your data i growing, and the RG running up to a 100%.

how is that a "bad situation" ? You use all available capacity in the raid group, there is no risk of oversubscribing. You need to manage capacity with flare LUNs and pool LUNs ...there is more chance of "bad situation" with oversubscribed pools as you will impact every single (thin) LUN in the pool versus one host filling up their file system and only impacting that host. It's all about capacity management.

I agree with Dynamox. Filling a traditional RAID Group to full capacity should not be a problem. Pools should be left with 10% free space but traditional RAID Groups do not use FAST VP so data does not need to move around. In Traditional RAID Groups all LUNs are fully provisioned devices as well so there is no chance for growth with those devices.

If your company continues to grow and your customers need more space for their use you would need to provision a new LUN or you would need to "expand" by using/creating a metaLUN which should absolutely not be created on the same RAID Group/set of disks anyway. So filling a RAID Group up completely should not be a problem.

Jim,

The RAID groups will be for a GPFS.  So is what your talking about still apply?

you configure hot spares, they are global and can be used by disks in traditional raid groups and pools.

for VNX2 - http://www.emc.com/collateral/software/white-papers/h10938-vnx-best-practices-wp.pdf

for Clariion/VNX - http://www.emc.com/collateral/software/white-papers/c1069-clariion-global-hot-spares-ldv.pdf

Hey,

Another question about RAID groups.  They are protected by the host spares correct?  I know the Pools are, but are the Raid groups protected too?

Thank you,

I dont see a reason to leave space empty on traditional raidgroups

I think that Dynamox and Rainer is much more experienced and knowledgeable about this than I'm...so Better listen to those two...

But what I meant was:

do not run out of free space if your capacity is growing,

what we see at almost all customers,is  that the amount of data is growing, and they because of this need free capacity some where on the Box, be that in a pool a RG or as disks on a shelf waiting to be used....

If you fill up your RG and do this for all RG you have, and needs extra Capacity, you could have a problem and be in a bad situation....

Hope that Explans my former comment...