iTera Disaster Recovery and High Availabilty

[sashton]

Hi List,

We recently purchased iTera Echo2 for our disaster recovery and high availability solution. My big question is, if we have to failover to a new Enterprise server that has been mirrored with our PROD data, how do I get around not having them named the same name. For example, our i5 is called JDEPROD. The backup machine is called JDEPRODHA on the network. Anyone else setup with a similar scenario. Thanks.

 

[Tom_Davidson]

Steve,

We are also in the process of migrating to iTera. We are planning on using a combination of host files (Citrix side), and renaming the RDB (host side). We are still several months away from testing though.

 

[AS400Guru]

Steve,
Whenever you are ready to perform a roll over test, let the network guys change the dns entry to point to the new ha server, so in essence, your jdeprod dns entry will point to jdeproda, this will enable you to perform a roll over without changing the system name.
We are also in the process of testing our roll over procedure. Will keep you posted.

I was thinking that there are plenty of shops that run OW with iTera, would you guys be interested in sharing ideas/issues/resolution over the phone maybe every 15 days or how about starting a forum for OW shops with iTera.
Let me know.
Thanks.

 

[ekempter]

If anyone is interested, I will open an iTera forum in the third party apps section.

 

[Tom_Davidson]

Eric,

That would be great.

Tom

 

[Tom_Davidson]

AS400Guru,

We thought about that but it would take to long to propagate the name change through our global network, too long coming back also.

We are also considering just having the backup machine assume the IP address of the down machine. Then we won't have to change DNS at all and won't need host files either.

That would only leave changing the RDB name on the backup system to our primary system name.

Tom

 

[sashton]

The only issue I see with that scenario is by assuming the IP of the down machine, you have now lost your "primary" machine. The main idea of iTera is to have a role swap. Backup becomes primary, primary becomes backup. In the case of failover, this would be ok, but we want to come up with a concrete method of role swapping and ultimately if need be failover. A thrid party iTera section would be beneficial. Thanks!

 

[Tom_Davidson]

Assuming the machine has more than one IP address: one public and another private, so you can get to the machine when it comes back. What down side effects do you see?

Tom

 

[AS400Guru]

I have bounced various other sceanarios with iTera folks, and the one with changing the dns entry seems most workable.
The other option is to have a virtual server defined and have that work as the Enterprise server. During rollover, you toggle on/off the interface to that ip address. This will work if you have a brand new install of OneWorld or if have the time to go in and change all the mappings/entries within OW. I personally would prefer this solution.
As far as the changing the dns entry, it works both ways, in the sense that during say a rollover, the dns swaps the primary and backup, this would elimiate any changes to the OW configuration.
Hope this helps.

 

[sashton]

Yeah that sounds like the way to go. I think we are going to test it out on a small scale in a couple weeks and see what happens. I'll let you know what we find out.

 

[CGIRoman99]

List,

We are using Product M (but I would have to believe the ideas are all the same) along with something called LAM on our CISCO switches - that stands for Local Area Mobility - it tricks the switches and reroutes the IP address of your production machine to the HA machine when you are ready to do a switch/swap whatever you want to call it, never touching our DNS entries, this might be a more costly way to do it but it works like a dream. Not sure what type of switches you are using. We also had to change our RDB entries and CFGTCP option 12 before being able to do a successful switch. Hope that helps.

 

[AS400Guru]

Re: iTera Disaster Recovery and High Availabilty - Parallel Test

Anybody done a parallel test yet?. If yes, can you please post your results.
Thanks.