ignite and servicegaurd

[Dan Copeland]

Hi,

We're setting up a DR site and are going to srdf data there and recover os images via an ignite server at the DR site. In the production environment, we are using servicegaurd but in a DR situation everything will be standalone. My question is--has anyone experienced any problems bringing up a server that was mc/sg but can't be at the new location (i.e. diff networks)?

I know the volume group import will be tricky, but I can get around that. I just haven't had a chance to test the ignite process yet.

tia

[melvyn burnard]

Have you thought to look at using MetroCluster, or even Continental Cluster, to ease this process?

If these are not an option, you woul dneed to take into account how clients connect to the application, e.g. hostname/ip addressing, also VERY thorough documentation, ande testing. Also remember you would need the correct scripts to change the personality of thr R2 pairs on the EMC etc. Essentially you are going to have all you need for Metro or Continental Cluster

[Santosh Nair_1]

Oddly enough, we're going to be doing exactly the same setup. Unfortunately, we haven't gotten to the point where I can test anything yet. I don't anticipate any problems. In a disaster situation, your DNS would have to be updated to point to the IP address of the DR machine/DR application rather than the production IPs, but other than that, it should be fairly straight forward. But like I said, we haven't tested it yet.

-Santosh

[Dan Copeland]

Metro and Continental aren't feasible:

metro needs syncronous writes (I believe)-- our dr site is 500 miles away (too much latency)

and here is what an HP employee said about continental -- "The mechanism using TimeFinder and SRDF is not supported by HP???s Continental Clusters capability nor recommended by Oracle for the protection of the Oracle E-Business Suite. HP has no plan to support with Continental Clusters the TimeFinder/SRDF data replication scheme"

These two options would be nice though -- I guess I forgot to mention we use emc's not xp's

[melvyn burnard]

Well on Metro you can set either sync or async, but this is a moot point for you.
As for CC and EMC/SRDF, I must confess I need reminding what TimeFinder does. I think it does the BCV (business copy volumes).
As for Oracle, I am surprised as this is the dtabase we currently have a toolkit to use with the Oracle Standby Database.
You may want to peruse some of the documents at:
http://docs.hp.com/hpux/ha/#ContinentalClusters