Originally posted by: SystemAdmin

I am setting up a hot DR site for some critical applications. I need to know the best way to replicate information from the production site rootvg to the dr site rootvg. I want to have both sites up all the time.
#AIX-Forum

Originally posted by: unixgrl

The answer depends on what exactly you are trying to sync...and why rootvg?

Usually the most important stuff to sync to DR is the application data.

If you only need a few files in rootvg to be sync'd, use rdist via scp. If you want
the entire rootvg including patches, ODM, etc, that is much more complex since you probably don't want to have the same hostname.
#AIX-Forum

Originally posted by: SystemAdmin

I know I can do a mksysb to my NIM server and then push it out from NIM. That fixes up the hostname/IP issues. But this is not a process that I have time to do every day. So, I was thinking of doing the mksysb process periodically and then using rsync and scripts to keep files synched up in the interim.
Does this make sense?

If anyone else knows a better approach please clue me.
#AIX-Forum

Originally posted by: flodstrom

Depending on just how critical your applications are and to what length you want to go, you might want to look at the PowerHA (HACMP) option. It will provide you with "real" hot failover.

http://www.ibm.com/developerworks/aix/library/au-hacmpcheatsheet/index.html
#AIX-Forum

Originally posted by: SystemAdmin

I have the PowerHA solution with Global Mirrors. But that doesn't help me keep changes on the rootvg in sync. It's pretty easy to keep the SAN disk in sync. I want to feel relatively secure that my rootvg is pretty much the same between the production and dr nodes.
#AIX-Forum

Originally posted by: flodstrom

Well, I don't know what your general practises and policies are regarding changes, updates, etc. or what application(s) your server run? However, if I'm reading this correctly I would say that the short answer to your question/inquiry is probably "don't do it!".

It's all up to the administrator to make sure that the high availability farm is kept on the same OS levels, etc. At least I wouldn't trust any home made half automated solution that sync's the system disks of a HA farm, but that may also depend on how "expensive" downtime really is? In the end I would say that there's no room for being lazy here I'm afraid!

In case you do have important data on the system drive (which you shouldn't have?), then I would move it to the SAN.

Sorry for not being much of help, but I don't think there's any easy answer to this.
#AIX-Forum

Originally posted by: SystemAdmin

It's just not that simple. It is 2 sites with local PowerHA in each site and Global mirrors between the sites. The application drops a couple files in the rootvg and if they ge out of sync then the DR fails in hard to explain ways.
#AIX-Forum

Originally posted by: unixgrl

So it sounds like you have a nice DR setup yet your customer is dropping junk in rootvg. First, I would strongly encourage you to convince your customer to stop doing that. You stated that having those files out of sync breaks DR so it should behoove them to put those files in a disk group that is already globally mirrored.

If they don't stop dropping stuff in rootvg, then I will again suggest you use something like rsync to copy the files to DR. This will only work if the files are either all in the same directory or always the same names. Make a script to rsync what you need and put it in cron to run regularly.

I'll reiterate again that for best practices purposes they should stay out of rootvg. They shouldn't even have permission to do anything there except in the /tmp or /var/tmp.
#AIX-Forum