- Posts: 5
- Liked: never
- Joined: Sep 03, 2009 2:32 am
- Full Name: jon bishop
I have some questions regarding using Veeam to replicate large volumes of data over WAN to a DR site.
We have 20 VM’s totalling around 3 TB in data. 10 of these VM’s are quite critical to the running of the business. Any replicated data from these VM’s at the DR site can be no more than 4 hours old. The other 10 VM’s are still critical but the replicated data from these VM’s can be as old as up to 1 day.
So my questions are as follows.....
1. If we are backing up 3TB of data every day, what would be a good guess at the total size of the incrementals combined? For example, I currently backup an 830 GB machine and the incrementals are consistently at 30 GB even when little to no work has been done. Can I assume somewhere around 5% of the total size of the VM?
2. How does Veeam recommend we calculate the size of the incremental? Are there tools available or is it simply a case of trial and error ?
3. Can you point me towards any case studies?
So the whole idea is to replicate 20 VM’s consisting of 3 TB of data over a WAN link to a DR site for Fail Over and Fail Back. I am open to any ideas and recommendations you may have. Also, we are using Vsphere 4 and veeam backup and replication 4.0. All guests are capable of using the changed block tracking feature.
Any suggestions and ideas would be greatly appreciated.
- Posts: 67
- Liked: never
- Joined: Feb 18, 2009 10:05 pm
Another good idea, having all vm coming from the same template is very effective.
- SVP, Product Management
- Posts: 28667
- Liked: 5184 times
- Joined: Jan 01, 2006 1:01 am
- Location: Baar, Switzerland
To achieve this, you should add new virtual disk on the source VM, move the swap file to this disk in the OS settings, and exclude this disk from replication. Now, in Veeam Backup 4.0, there is a new option to prevent automatically dropping excluded disks from replica configuration on the replica side. So you just manually create VMDK file for swap on target with the same name as on source (or just copy source VMDK once to replica VM), and now if replica is started, it will see the disk and will use it for swap. This way you get the exact copy of source VM on target, with swap enabled and Windows not complaining and suffering due to the disabled swap - without having to replicate actual swap file data!
- Posts: 50
- Liked: 2 times
- Joined: Feb 10, 2011 7:27 pm
I've got a number of vm's that have volumes for log/temp type files. It would be nice if I could replicate the presence of these volumes without the content.
Then, in a DR situation, I can spin up the replica without having to allocate/present/configure storage on the vm.