Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am looking for some input on a Microsoft fail over cluster implementation on 2012. To use the storage and pools it apparently has to have JBODS if your using a raid card. But the LSI card we use (lsi megaraid SAS 9260cv-4i) does not do JBODS. For the fail over cluster What is the best way to get these disks on the home grown SAN server(s) (we have 2 of them) and want to have high availability on the storage and use it within fail over clustering. usable in the cluster?

Should we just not have it in the cluster and present Iscsi targets? What is the best practices in this case?

Any help is greatly appreciated.

share|improve this question
up vote 6 down vote accepted

Similar scenario here.

First, forget storage pools. Nice try - NOT THERE YET. Check google - they are a nice feature, but lack a lot of higher end stuff (expansion, rebalancing when you add discs) and have a SERIOUS performance problem. SERIOUS.

I work on similar stuff (8 disc, 24 disc servers) and I am just ignoring storage pools for the moment. Until MS handles capacity expansion nicely, and gets the performance under control, AND adds a 2nd level cache, I am just replacing my raid controllers with Adaptec 7805Q, add some SSD as transparent cache managed by the RAID controller, and keep going.

My main problem really is (a) the missing functionality along expansions and (b) - the horrific write speed (10x slower than a raid card):

Sorry, but I need performance.

Now, do not let me even start about the CRAPPY ISCSI service. Nice try - except... does not work from VHDS files, only from VHD. Who had that glorious idea? TONS of nice advantages in VHDX, tons of white papers how important they are blablabla - and then the on board ISCSI service can not deal with them. GREAT. Using it for booting servers soon, but I can not say I am happy about this decision MS did.

Depending what you do with them you do not need ISCSI at all - really depends. Hyper-V can cluster shared nothing from a file share.

Now, for high availability - if you need to go that way - you need either:

  • An ISCSI service that supports that on multiple machines (StarWind does that with replication between backends)
  • A proper SAN
  • 2 DAS with the accompanying LIS raid cards. THere is a cluster aware RAID functionality interface in windows, and LSI has the only raid controllers ready, but you need the supporting back end hardare (basically proper high available dual channel SAN enclosure).

Whatever you do with what you have in a server wont be highly available due to this server being able to fail.

share|improve this answer
So your suggesting just use shared smb3 then. – Dimentox Jan 3 '13 at 17:37
So Storage Spaces are very ZFS-like, right? – ewwhite Jan 3 '13 at 17:46
@ewwhite SADLY not - they are like ZFS version one. They have a TON of potential, but at the moment they lack seriously on the baseline. Enough to make them not usable for high performance stuff. – TomTom Jan 3 '13 at 17:52
@Dimentox This is what I do. Hyper-V from SMB shares. The SMB server has Raid controllers and comes from a NTFS high performance setup. I hope for a serious service pack. – TomTom Jan 3 '13 at 17:53
Fail over cluster manager is wonky at times. I cant add the raid disk not storage pool so i can have the file server use it lol. – Dimentox Jan 3 '13 at 18:06

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.