Replication settings for new environment

[tech_boc]

Hi,
I used to use DoubleTake (DT) for host based replication on ESX 3.5 host. Now I am on ESXi 5.0 where DT does not support host based replication (and guest based is an absolute nightmare).
I'd like to consider VEEAM Backup and Replication and don't mind waiting some time till full support for ESXi 5.0 will be implemented (as well as 10X replication perf improvements)
Few questions related to the task I am trying to achieve.

My enviro:
- ESXi 5.0 at location A (2 CPUs with 12 cores each) - I now have 8 VMs (1 to 2 vCPUs each) to replicate and this number will grow - they are Windows 2008 R2 Terminal Service systems with very little dynamic data stored on them
- ESXi 5.0 at location B (2 CPUs with 12 cores each)
- both ESXi systems have DAS SAS attached (large number of spindles)
- connectivity between location 30 Mbps (some 25% of it is used for Exchange DAG replication, SQL log shipping, and DFS file replication)

The goal is to replicate the VMs for DR. Point of time replicas are not required. RTO of 24 hours is acceptable. In case of DR I want to be able to fire up the replica VMs, change IPs and use them.

Questions:
1. I notice in eval guide two available installation modes for VEEAM: Virtual Appliance mode and Network mode. Which one would be more appropriate for my setup?
2. What performance impact should I see on ESXi hosts due to the replication?
3. Will I be able to achieve the RTO with my partially used 30 Mbps pipe between the location and the number of hosts growing to some 25 within next 3 years?
4. Any best practices I should keep in mind?

Huge thx

[Vitaliy S.]

1. I notice in eval guide two available installation modes for VEEAM: Virtual Appliance mode and Network mode. Which one would be more appropriate for my setup?

For the current version (v5) our recommendation is to deploy Veeam B&R server on the DR site/target host (pull replication) and use Network backup mode to replicate your VMs.

2. What performance impact should I see on ESXi hosts due to the replication?

When replication/backup job is running, the only notable impact on production ESXi host might be snaphost presence. Snapshot operation puts some load on datastore. Please note that If your storage array supports VAAI (hardware assisted snapshots), then this impact is really minimal.

3. Will I be able to achieve the RTO with my partially used 30 Mbps pipe between the location and the number of hosts growing to some 25 within next 3 years?

In case of replication jobs being used, RTO is really low. You simply need to fire up all the replicas, change their IP and that's it.

Please note that our upcoming version (v6) will do VM re-IP automatically, so it becomes even easier to fail-over in case of a disaster.

4. Any best practices I should keep in mind?

Please search this forums for existing topics, there is a lot of useful info > http://forums.veeam.com/search.php?st=p ... n+ESXi+WAN

[Gostev]

With ESXi target it would truly be best to wait until v6 instead of trying to architect anything now, since the difference between v5 and v6 replication is pretty dramatic. Especially since you don't mind waiting

[tech_boc]

Thx much for the answers. I'll wait till the v6 which I hope will be available fairly shortly.