We are considering this product as an alternative to standard Array based replication products and understand that it is fully supported on VMware and many people are using it successfully in this environment.
Anyone have any first hand accounts of how its working for them? Any gotchas or limitations?
Our plan is to have our production ESX site online at all times, if a disaster strikes bring up the ESX hosts at the DR site with the replicated luns.
I like the fact that it keeps snapshots and allows you to roll forward/back in case you have data corruption as well as huge compression over the WAN. (6-8x)
Any feedback would be appreciated.
I'd be interested in your experiences thus far. We have been talking with RP/EMC for the past few months about this. Slideware and demos are great, but before I take this rather costly plunge, I'm interested in real world results.
What are you doing with the copies once replicated? Are you backing them up or just keeping them on standby for a disaster? Did RP eliminate your need for traditional backup clients on each of the VM's?
Not sure how far along you are with the product..but I do have more questions.
We are also looking to implement RecoverPoint, with the SANtap product, for our VMware VI3 environment. We have successfully tested this in our lab, but EMC has yet to provide us with any customer references for us to speak with. If anyone is doing this today, I'd really like to speak with you about your experiences and compare notes.
\- Chris Graham
The main benefit to RecoverPoint, as I see it, is its scalability and compatibility in environments with a variety of SAN Vendors. That's its sweet spot. EMC loves to push it because no one else has anything really like it with the same levels of automation (Replication Manager) and so forth.
But I'm not really convinced that in say a regular EMC CLARiiON only environment that it would really offer you anything of value over, say, SANCopy. You can still use Replication Manager with SANCopy if thats your thing, or from a VMware level just replicate straight-up... In which cases I have successfully done this many a time with pre-registered remote site VM's and so forth.
Honestly SQL does SQL replication best and Oracle does oracle replication best etc.. The main reasons you choose instead a standard replication means across all products are very similar to the reasons for choosing RecoverPoint as a standard in a heterogenous SAN environment.
As for the snapshot thing I think its uses are a bit over-hyped. You can still do a more limited number of snapshots with something like SANCopy and how many do you really need to address potential data corruption?
Another attempt at recovering this thread.
According to my reading on-line, RP can only be used if you have Cisco MDS FC switch (because of SANTap) or are using RDM volumes and the RP agent on the guest, is this true?
How does it compare to http://www.yysystems.com/ ?
There are two methods or types of splitters - Host and Fabric. Host splitters would be used on the guests (RDM in necessary for this). There are two types of Fabric splitters out there - Cisco/SANtap and Brocade/FAP. To date - I've heard of quite a few successfull RecoverPoint installations for VMware ESX environments.
Actually how is using RecoverPoint for replicating VMware instances any different (in planning requirements, strategies, dependencies or issues) than it is for replicating any other content?
What specific considerations should come into consideration in this context?
Any feedback would be most welcome...
I have a question EMC have just installed Recover Point, Is it the case that the DR ESX Servers should not be able to connect
to the Replicated LUNS during normal operation or Just in a Fallover Situation, Reason I ask is that although our DR ESX servers
see the Luns they can't connect during normal Recover Point Replication. I can't even add other Storage the system just Times out.
EMC UK apear to be using us as a test site so I an not confident in there setup.
Using the Cisco San Tap with CX-40
  If you have any Doucments That can help me with the setup that would be grate.
Had to use Advanced Options DiskMaskLuns with below string to to mask out replicated LUNS all is Ok now.
I hope the the new Vmware DR Site Manager Product will make all theese Advanced Options Edits Point and Click when Released
and make DR Fallover easier.
Hi all, interesting thread.
I am interested in Countinuous Data Protection
I can see no way i could ever use RDMs for each vm, so i would have to use either a cisco or broccade spliter solution. I would like to have 100's of snaps (basically i want to replace traditional backups with this kind of solution)- i would have to keep 90 days worth --- is this posible with this kind of technology ? i guess i would then need to mirror the Journal.
What kind of visability do you get in the application, i am guessing that if you use the cisco solution you only get visability at the LUN level (is this correct).
You are correct - you would have to use the Cisco SANTap or Brocade FAP solutions to do replication at the ESX lun level. The Journals are not replicated (They are the working area for the replication prcesses and where the "Snaps" are held at either end). There are 3 vulume types (Repository, Journal and replication volumes). The repository is specific to a RPA site, journals are per consistancy group and replication volumes are specific to hosts. Depending on how many snapshots you will need - that will dictate the sizing of the jvols. You can do local and remote replication with RcoverPoint.
VMware Communities Moderator
I have been doing a RecoverPoint rollout at my company, working closely with EMC to get it all set up and online.
To answer your question, you cannot access the TARGET LUNs while replication is in progress, this is done by design. You must pause the replication to allow the lLUNs to become visible.
This is kind of the problem we are having. Even though RecoverPoint has a 100 Snapshots you can roll back to, #1 you need to make sure all possible snapshots are added to your zones, since it randomly picks slots during snapshots.
My problem is when we actually failover to the other side (from production to DR) per EMC we have to recreate all of the VM's from the .VMX files per failover. This is a bit of a lengthly process to go through every time you want to fail over to DR. It seems that when you choose to go to a particiliar point in time, you get a RESIGNATURED LUN on the other side so even though the DR side can see the new LUN (Named something like Snapshot 03 LUN 0) you must recreate ALL of your vm pointers each time you fail over and fail back.
EMC has said that the new VMWare Site Recovery Manager should fix this process but there is no solid release date yet (Q2 or Q3 of 2008). Anyhow, I highly recommend reading the 330 page RP Admin manual as well to see how quick you can hose your data.....scary! hehe.....Take care....
I have EMC CX3-20 SAN's on each side, and 2 ESX server farms, and considering Recoverpoint
Would it be possible to speak with you for a few minutes regarding your experiences?
Thanks and regards,