Suggestions for Datastore Rep VS Individual VM Rep

[msidnam]

Hello Everyone,
I have a ticket in to support for some suggestions as well, but i'd like to get some of your input to see what everyones doing and what has/hasn't worked for them. A little background on our main network and what i would like to do:

We are in the process of virtualizing all of our servers, We've done a good chunk but we have several to go.

We have a mix of 2003/2008 and Linux (Ubuntu) servers and one solaris server.

As we were virtualizing I was creating replication jobs based on the servers and grouping them up. Ie, Linux machines with mysql, apache and tomcat would go in one group, windows 2008 with sas (since it can only be done via network) would be another, windows 2003/2008 with parallel and so on.

We have developer who is creating an ubuntu server for everything. MySQl, apache, Tomcat, Liferay, etc. Since i already have them grouped, when i add a new server i find myself adding the new server individually to test. if it works i then delete the job and the files, then put it in the main group of like server. I find this is a lot of work, especially if the initial job takes a couple hours, i sometimes forget to delete the test job and add it to the main group.

Since we have the room, i am thinking of creating a datastore of just linux machines and copy the whole store nightly. because these servers will need to be up 24/7 if one goes down i need a quick way to get them back. starting a replicated server would do the trick, however we recently moved the the replicated envirnment to our colo which is on a different subnet. This issue will probably be solved by our current bex backup.

If i go the route of not creating a datastore for just the linux machine and just add the datastore that all servers are in, would the VSS and enable vmware tools quiescense play nice together with a mix of servers like that? I've been reading the forums that some users had issues if both options were selected.

My last question is, depedning on what i do what happens to the current replicas that i have? If i decide to move to a full linux datastore and repicate the full store do i lose my 14 snapshots since now they are in seperate areas?

[Gostev]

Hi Mike,

I would use VM Folders to organize the jobs. This won't require any infrastructure changes (unless you are already using them for something else).
You will not lose your existing replicas unless you actually delete them. Keep the old jobs for a while, just disable them.

Thanks.

[msidnam]

Gostev,
My apologies. I cant beleive i overlooked that. I do have the VM's pretty much organized using the folders in vcenter. a couple minor tweaks and i think that should help alot. once i add a new VM to that folder it should replicate. Any way to start a job on just one VM?

Thank you again. I will test in the next few days.

[Gostev]

You are welcome! There is no way to start the job on just one VM in the current version. Thanks!

[msidnam]

In Advanced setting should i change storage from Optimized for Local to WAN since i am not going over our WAN links to our colo?

[Vitaliy S.]

In Advanced setting should i change storage from Optimized for Local to WAN since i am not going over our WAN links to our colo?

Hmm...are you going over WAN link or not (didn't get)? If not, then stay with Optimised for Local

[msidnam]

We are going over 100mbps links from AT&T. Sorry, i see my post. it should have said I am going, instead of I am not going.

[Vitaliy S.]

In this case I would recommend to go with WAN target.