Nimble Storage

[XZim]

I'm thinking of investing in Nimble Storage to replace our current FC SAN infrastructure. It is does some pretty clever stuff with SSD and snapshots for backup ...however they are a relatively new player in the storage game with a small (but growing)client base.

Anyone using Nimble already and what has your experience been particularly with running VSphere on this SAN?

[ng_mo]

I heard that on a Nimble array, if power is physically removed from the active controller (accidentally removing a power cable) the controller will be powered off and the standby controller will assume active role.  Can anyone help to verify this?  (what I am trying to verify is if power is removed, if there is an ability in the enclosure to share power from the suriving controllers powersupply to prevent a controller failure event)

[slyfoxxx]

Yep

Happened here and support was rt on it even though it failed over successfully

Fran

[Josh26]

ng_mo wrote:

I heard that on a Nimble array, if power is physically removed from the active controller (accidentally removing a power cable) the controller will be powered off and the standby controller will assume active role.  Can anyone help to verify this?  (what I am trying to verify is if power is removed, if there is an ability in the enclosure to share power from the suriving controllers powersupply to prevent a controller failure event)

What's the concern?

Is the SAN really does have availability, a failover to the standby controller (if it has to happpen) shouldn't be an outage causing event anyway.

[dleary]

Just to clarify:  the redundant power supplies power the entire chassis. A power supply is NOT tied to a specific controller, and a single power supply provides more than adequate power for both controllers.  Either power supply can be removed with no impact whatsoever to either controller.  Loss of one power supply will NOT cause a controller failover.

[ng_mo]

I've asked this question on another  discussion board and the answer I received were that the controllers were independently powered.  So, this is the first time I have heard they are not.  Do you have anything (documention perhaps) that you can share to prove this?  I don't want to have to go through a PoC - so the more info I can share with my customer, the better off I am.

[ahahum]

I can confirm that the redundant power supplies are for the entire array - not 1 power supply per controller.  I had to re-route some power cables in our rack and I confirmed this with our Nimble Sales Engineer.  We had no issues doing this.

I can understand if you're still looking for somehting on Nimble letterhead,however.  I hope this helps you out.

[jmlmagsino]

hi!, I am a newbie to this forum. Nimble people approached us today. timely because we are currently evaluating our VMWare expansion. They wanted to do some POC however we are a bit hesitant to proceed hearing the sales talk they gave us (too good to be true). However, we wanted to see if it will deliver.

The reason I wrote is because maybe someone can provide a POC plan that they did during their storage evaluation.

We wanted to do a 50-60TB Usable Capacity requirement and a viability of a 2nd site for SRM. I hope someone can provide us their POC activities executed you can share your ideas as well.

Just in case, after our POC, we would be willing to post the technical results still subject to open discussion.

Message was edited by: JML

[Yps]

I´m on parental leave / vacation, back 26 Augusti.

/Magnus

[nobody8importan]

How does Nimble handle replication? Does replication rely on snapshots?  Can you have a cluster of two Nimbles (one primary, one in a secondary site), replicate between them, and then vMotion between hosts attached to those local storage arrays on each side?  (true live vMotion, not a crash recovery HA or something facilitated through SRM)

Also was reading one of their documents in regards to snapshots and it suggested only performing so many snapshots at once since it sends a snapshot request to vSphere and a quiesce request to the VSS writer.  Performing too many at once can cause "extensive interruption of service" due to so many snapshots being created at once.  How many is too many?  If I was wanting to replicate 400 VM guests and have a 5 minute or even a 1 minute or less consistancy delta between Nimble arrays, is it even possible?