In September 2013, EMC
released its next generation of mid-market storage arrays, the VNX2. This
release is a complete re-write of the code used in the previous generation. Among other things, this array is designed specifically for
use of optimizing performance via tiered storage, especially flash or solid
state drive (SSD). (For information on other storage tiers, read Scott Lowe’s article on how SAS, Near Line SAS, and SATA disks compare.)

The VNX2 comes in five models. All of the models can be
unified, meaning they can have both File and Block capabilities.

  • VNX5400
    has up to 250 drives (2.5″ and 3.5″ Flash, SAS, and NL-SAS) and 1000 GB
    Fast Cache.
  • VNX5600
    has up to 500 drives (2.5″ and 3.5″ Flash, SAS, and NL-SAS) and 2000 GB
    Fast Cache.
  • VNX5800
    has up to 750 drives (2.5″ and 3.5″ Flash, SAS, and NL-SAS) and 3000 GB
    Fast Cache.
  • VNX7600
    has up to 1000 drives (2.5″ and 3.5″ Flash, SAS, and NL-SAS) and 4200
    GB Fast Cache.
  • VNX8000
    has up to 1500 drives (2.5″ and 3.5″ Flash, SAS, and NL-SAS) and 4200
    GB Fast Cache.

The VNX has always come with solution enablers, or software
suites, for which you can buy licenses. 
All VNX arrays have the capability to use these features, but the
licenses need to be purchased to enable them. The software suites are:

During a recent installation of a VNX2, I noticed there are major differences if you’re used to the CX series or VNX1. A big one is the way hot spares are used.

In the older versions, you specified a
particular disk as your hot spare. For example, if a disk died in a RAID group
or storage pool, you knew that Disk 14 in DAE 1 was the hot spare and would
become part of that RAID group temporarily. This is no longer the case. There’s
now a Hot Spare Policy that you need to set up. The administrator is not in control of which disk is chosen.
You can pick how many disks to reserve for each disk type, but you can’t say
which one. Also, these disks remain unused; they are not configured as hot
spares until it comes time to use them. Look at how the Hot Spare Policy is set
up in Figure A.

Figure A

Also, the hot spares are no longer temporary. So if a disk fails and one of your
unused disks is used as the hot spare, it will stay in that RAID group or
storage pool unless you manually take it out. This could mean a few things. If you’re using a larger disk
as a hot spare, you may lose some of your capacity. Also, if you have configured redundancy based on your
shelves and the buses they’re on, this might get tricky; you may lose control of that redundancy
as well.

A pretty cool feature is that after you create your storage
pools, RAID groups, and/or luns, you can use the disks while they are zeroing. In the older arrays, these disks would
be unavailable. I don’t know if it might run a little slower, but the performance hit would be temporary.

Something to keep in mind is that the new VNX has different power requirements. While the older arrays
could use 110v or 220v and it would auto-detect, the new arrays require
220v.

Summary

The VNX2 is built to be much faster than any of
the earlier arrays. It’s still
early, so many customers are waiting until the bugs get worked out before
purchasing a VNX2.

If you’ve had any experiences with a VNX2, please describe them in the discussion. I’m curious to hear what
people think of this release.