VMware Cloud Community
aiea96701
Enthusiast
Enthusiast
Jump to solution

Site/SAN to Site/SAN continuous Fault Tolerance

We have a customer who is seeking the following desired results.

1. He has several Web Servers and a SQL Server that he houses a ton of jpg's (over 2gb) and that grow approx 500g/day

2. He wants to place his Jpgs on a Dell EqualLogic SAN at his site

3. He wants to have a failover site that will replicate the data on his current SAN at his site as quickly as he can get

here is the question: If we setup an EqualLogic SAN at his offsite location and implement SAN replication with Site Recovery Manager (bearing in mind that his offsite location is in another state), will he be able to replicate his data/JPGs every 15-30min? AND, in the event his site has a disaster, can the remote site be placed in production within minutes so that his operation is not adversely affected?

If so, how would this be done? Does it require Enterprise Plus licensing to encorporate FAILOVER operations?

Reply
0 Kudos
1 Solution

Accepted Solutions
AndrewBarnes
Contributor
Contributor
Jump to solution

First of all, a disclosure. I work for Neverfail.

There are a couple of other things to consider

1) Don't forget about "failback", how will operations be moved back to the original site?

2) Neverfail has in-band, de-dupe based WAN acceleration that could help with the bandwidth limitations.

3) Are any applications on physical servers involved. SRM will only failover virtual machines.

Let me know if you want to discuss further.

View solution in original post

Reply
0 Kudos
10 Replies
idle-jam
Immortal
Immortal
Jump to solution

It is possible if:

1) we need to know the bandwidth between both site. it is shared, or it is dedicated for equallogic replication only.

2) what is the RTO (recovery time objective) the duration that the VM will be up to serve user again

3) what is the RPO (recovery point objective) how much data lost can he tolerate when the VM is up

Any version of vsphere would do, of course version with DRS with enable auto placement of the virtual machines. SRM could be thrown in to automate the whole process.


iDLE-jAM | VCP 2, VCP 3 & VCP 4

If you found this or any other answer useful please consider the use of the Helpful or correct buttons to award points

Reply
0 Kudos
sketchy00
Hot Shot
Hot Shot
Jump to solution

Sounds like you have several questions here. I think it’s best to break this down into two specific categories:

1. SAN to SAN replication.

2. Fail-over to remote site with SRM.

I can’t speak much on #2, because I do not run SRM yet. So I'll leave that to the SRM experts. I do run SAN to SAN replication with EqualLogic arrays. There are lots of things to consider. The frequency that you are able to replicate at is a function of changed/added data (not in file size, but by block), and your effective throughput to the remote location (which is a result of bandwidth, minus degradation due to packet loss and latency).

Best to get the replication part figured out first, since you need that for any successful SRM implementation. I posted a 5 part series in deploying replication with EqualLogic arrays. Here are the links if you are curious

http://itforme.wordpress.com/2010/04/30/replication-with-an-equallogic-san-part-1/

http://itforme.wordpress.com/2010/05/13/replication-with-an-equallogic-san-part-2/

http://itforme.wordpress.com/2010/05/31/replication-with-an-equallogic-san-part-3/

http://itforme.wordpress.com/2010/07/24/replication-with-an-equallogic-san-part-4/

http://itforme.wordpress.com/2010/10/31/replication-with-an-equallogic-san-part-5/

aiea96701
Enthusiast
Enthusiast
Jump to solution

Thanks for your reply.

RTO.............he prefers no interruption in service but probably will accept 15min

RPO............he prefers to have the failover site up and running "ASAP". So, the soonest possible.

............. what I believe he "desires" may not be practical. So, what I believe I need to do is come up with a scenario that provides the best alternative. So, I believe in that case, I can setup the parameters to come to (a) the best outome or (b) the next best outcome, etc.

Reply
0 Kudos
aiea96701
Enthusiast
Enthusiast
Jump to solution

thanks. I'll review the links.

Reply
0 Kudos
idle-jam
Immortal
Immortal
Jump to solution

Yes you're right. Do look out solution like Neverfail, Doubletake, or XOsoft if he wants instant failover and etc.


iDLE-jAM | VCP 2, VCP 3 & VCP 4

If you found this or any other answer useful please consider the use of the Helpful or correct buttons to award points

ChrisDearden
Expert
Expert
Jump to solution

putting 2 and 2 together of his budget is Equalogic sized , I wonder if the bandwidth required to be able to replicate the ammount of changed data is going to be out of budget for those expectations?

just to confirm as I think there was a type in the 1st post -is that 2Gb of image date or 2TB ? with 500Gb a day growth ? it doesn't quite add up.

@chrisdearden

If this post has been useful , please consider awarding points. @chrisdearden http://jfvi.co.uk http://vsoup.net
Reply
0 Kudos
aiea96701
Enthusiast
Enthusiast
Jump to solution

currently has 2Terabytes of data and he claims it grows at 500gb a day but I believe he also flushes data on a regular basis to keep the size to about 2 TB

Reply
0 Kudos
AndrewBarnes
Contributor
Contributor
Jump to solution

First of all, a disclosure. I work for Neverfail.

There are a couple of other things to consider

1) Don't forget about "failback", how will operations be moved back to the original site?

2) Neverfail has in-band, de-dupe based WAN acceleration that could help with the bandwidth limitations.

3) Are any applications on physical servers involved. SRM will only failover virtual machines.

Let me know if you want to discuss further.

Reply
0 Kudos
aiea96701
Enthusiast
Enthusiast
Jump to solution

did some checking on NeverFail. Sounds like "the" solution. I've placed a call and will have a rep contact me. Thanks for putting me on the right path.

Reply
0 Kudos
AndrewBarnes
Contributor
Contributor
Jump to solution

Glad to be of help

Reply
0 Kudos