Running Synchronous Replicated SAN's in VMWare?

I am looking into purchasing a completely new Virtual Infrastructure. Initially I am looking at an IBM H Chassis, 4 blades with 2x CPU and 64GB RAM in each. This will be connected via 8GB FC to two IBM DS3512 SAN's. I believe this will be running VMWare 5 and ESXi. We only have a single data center and all of the equipment and both SAN's will be sitting ont he same rack connected to the same FC switches.

Part of the requirement is for this solution to be 100% redundant which is why I need two SAN's in synchronous or near synchronous replication. IBM has suggested their "Remote Mirroring Base" software to accomplish this asychonously (within a couple of minutes) but the cost is around $40k (which seems a bit rediculous)

Is there any other ways to accomplish a synchronous or near synchronous replication using either VMWare, VEEAM, or another solution? Also, what VMWare "add on" would be required for a "seemless" failover is the event that the primary SAN went down?
Who is Participating?
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
Why not try Veeam Backup and Replication v6, and replicate the VMs at 15 mins intervals, or less.

At approx $500 per processor.
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
Andrew Hancock (VMware vExpert / EE MVE^2)VMware and Virtualization ConsultantCommented:
You may want to look at maybe a different technology like vSphere VSA, HP LeftHand VSA, with clustered SANs, and replication between SANS, to prevent the single point of failure of the SAN.
Put Your Flow Data to Work

SolarWinds® Flow Tool Bundle combines three easy-to-download, easy-to-use flow analysis tools that can help you quickly distribute, test, and configure your flow traffic.

andyalderSaggar maker's framemakerCommented:
Umm, isn't there 100% redundancy with just one dual controller DS3512, 2 SAN switches and two or more servers? I can't see a SPOF there, the interconnect between the controllers maybe but I believe it's passive. For that matter the backplane of your bladecenter is a SPOF, it may have redundant paths but I bet I could kill the whole thing by forcing a mechanically faulty blade (or a screwdriver) in hard enough.

100% redundancy isn't 100% uptime of course, you can get two faults but there's nothing that'll protect against that in a single data centre.
AIC-AdminAuthor Commented:
Thank you everyone for the replies. I am researching the VEEAM option more as that may be a viable and cost effective solution.

To my understanding VMWare SRM requires two sets of hardware for fail over (i.e. Primary and secondary Hosts runnign vCenter, SRM, and SAN's). I didnt think that SRM would work within the same datacenter if all of the blades in my single chassis are being used for production systems.

The H Chassis has an upper and lower backplane as well as mutlilpe FC and Ethernet NIC's, PSU's, etc. While yea mechanically someone could surely make the device fail is a single psu, nic, backplane, blade, etc failed there are others to take over. I agree two SAN's is over-kill but I'm being told by my CIO its 2 SANs or nothing.
andyalderSaggar maker's framemakerCommented:
You need to re-educate your CIO; the DS3500 has the same level of redundancy as the bladecenter, every part is duplicated, i.e. no SPOF. So does just about every other SAN storage box on the market. Syncronization software is a SOPF of course, if the code is written badly it could kill both boxes just like an unexperienced person using vCenter. I'd bet he'll be happy for you to connect the whole thing to a single power feed.

By the way, what are you going to put in the other server slots in the BladeCenter? It's pretty power inneficient having one mainly empty.
AIC-AdminAuthor Commented:
We decided to go with the H Chassis for its ability to support multiple external SAN's instead of an S Chassis so the remaining slots will nto be filled.

We will be connecting the Chassis to two different UPS's. We also have a backup generator that powers our data center withing 10 seconds of a power failure (including A/C) so power is accounted for.

I am hoping to convince them that two SAN's is not necessary but so far I ahve been unsuccessful.
andyalderSaggar maker's framemakerCommented:
10 empty slots in your bladeCenter? I think you'd be better off with four off 1U servers plus external SAN and LAN switches.

I haven't done the power calculation for the IBM enclosure but for both the Dell and HP solutions it comes out at less than half full = better to buy discrete units as far as BTUs+watts are concerned. Plus you're not tied in to a single manufacturer.
andyalderSaggar maker's framemakerCommented:
Well, I gave you my opinion so maybe you could give me something in return. Is the accoustic baffle that IBM sells to quieten down the back of the BladeCenter any good or does it restrict the airflow through the rack too much?
Dual SANs for this reason is nuts, IMO. Get an Isilon, LeftHand, or other scale-out SAN/NAS that can be operated in an N+1 (or N+X) resiliency model. Sync/Async replication then becomes a non-issue within your rack.

Of course, you'll connect to storage and data over multiple switches (with redundant uplinks, right?), so that takes care of N+1 for the interconnects.

Don't waste your company's money on a chassis: that itself is an SPOF (backplanes can fail). Get N+1 (or better yet, N+2) dual-socket 1U or 2U hosts. That takes care of the host-level redundancy.

Finally, enable and configure HA and--for the VMs that are really, truly critical--FT.

WIth that setup, only the datacenter environment (all that stuff sitting in a single rack) remains as the non-redundant element. This gives you better overall utilization in the same way a RAID5 array makes better use of raw space than RAID1/0.

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
AIC-AdminAuthor Commented:
I do not own the IBM equipment yet so I honestly do not know anything about the accoustic baffle that you are mentioning.

I'm not overly concerned about the cooling requirements in our data center. I will be gathering some white papers on the SAN redundancy and I am beginning to think the H Chassis really isn't the correct solution like you and Millardjk are saying with all the empty slots.

Thank you for your detailed suggestions and input on this. I am going to research all of the items you listed further. I am really thinking that the blade center will not be my final solution any more.

Thank you a million everyone!
AIC-AdminAuthor Commented:
Thank you everyone! I have lots of homework to do now :)
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today

From novice to tech pro — start learning today.