Proper Veeam Configuration when using WAN Acceleration

[theether]

I just got the off the phone with Veeam Tech Support who was utterly useless in answering my questions. Geee.. Thanks guys..

Hopefully some of you can help.

I have a slow link between two sites and am using Veeam 5 to replicate 1 VM. I got successful replications before I installed WAN Accelerators. The issue I have is that I expected to get a nice boost in acceleration when the WA's were installed and frankly they haven't sped things up at all. I'm using Cisco WAAS with 4.3 code for those that are curious. I think my issue is either with the way I have my Veeam Replication job is configured or with the WAAS units themselves not recognizing accelerate-able traffic or patterns. At this point I'm not sure. I asked Veeam support (Level 1 and Level 2) what the appropriate setup for Veeam when replicating over a slow link with WAN Acceleration would be. They couldn't tell me.

So here's my questions. With WAN Acceleration should I configure Veeam to:

1. Do inline deduplication? ( I have only one VM that I'm replicating, not sure this helps much for only one)
2. Do Compression? (I would think I'd want the Wan Accel's to do that and set it to NONE in the Veeam Job)
3. Should I Optimize for Local, WAN or LAN Target?
4. Should I enable changed block tracking?

Thanks very much for your help!

Ether..

[Vitaliy S.]

1. Do inline deduplication? ( I have only one VM that I'm replicating, not sure this helps much for only one)
2. Do Compression? (I would think I'd want the Wan Accel's to do that and set it to NONE in the Veeam Job)

Doesn't matter. These settings are only applied to replica rollbacks (VRB files), because an original, unchanged .vmdk should be presented to a target host.

3. Should I Optimize for Local, WAN or LAN Target?

WAN target would be preferable.

4. Should I enable changed block tracking?

Yes, of course.

On top of that, could you tell me what is your replication target? Do you use full blown ESX host or ESXi? Where is Veeam Backup server located?

[theether]

Thanks for the reply! The Hosts are full ESX 4.0 classic. The Veeam Server is a Virtual Machine and is on the Source side. Do I need to setup a policy in my WAAS units to accelerate traffic on port 2500?

[Gostev]

Might be better to direct this question to WAAS technical support or forums. We have no knowledge of how WAAS should be configured. Although we do have some WAAS users here, so may be they will be able to answer your question right here.

[tsightler]

What are you replicating to? Is it a ESXi host? If so, the problem is that Cisco WAAS by default has a policy that only perform TFO optimizations on port 902, since that's typically used for VMware remote control and other agent communications that aren't very useful to accelerate. You need to either configure a policy exception for the replica hosts, or modify the default VMware-VMconsole policy to use full acceleration (TFO, LZO, DRE).

This shouldn't be an issue if it's full ESX since with that mode, Veeam pushes and agent that uses port 2500 (or some port around there). You should be able to use the WAAS console to easily identify the traffic and see what policy and optimizations are being applied to the data stream.

We use WAAS replicating about 500GB of data from our EU facility over an E1 (2Mb) and we typically see anywhere from 60-95% improvement (60-70% for services like Exchange, 90-99% improvement for servers with less change like file and application servers) so with a little tweaking WAAS can add some serious improvement.

[Gostev]

Here we go, just as I posted

[theether]

10.30.1.44 is my Veeam Box. It's a Virtual Machine on the source side. I've run the show stat command on the remote waas unit.

Veeam is setup to use network mode, with no compression (thinking the WAAS would do the magic).

RemoteCiscoWaas#show stat con | include 30.1.44
248287 10.30.1.44:54964 10.30.2.62:902 1c:c17fcc T 00.0%
248345 10.30.1.44:54975 10.30.2.62:902 1c:c17fcc T 00.0%
248386 10.30.1.44:55029 10.30.2.62:902 1c:c17fcc T 00.0%

I[m not seeing any appreciable acceleration it appears that TFO only is what is being triggered. TDL is not happening which is why I'm not seeing what I should.. I'm going to modify the Vmware policy to enable full acceleration and see how it goes.. THANKS!!

[theether]

It works now! Just like you said, I had to set the VMware-VMconsole policy to full acceleration. Thanks tsightler!

RemoteCiscoWaas#show stat con | incl 902
196026 10.30.1.44:49533 10.30.2.62:902 1c:c17fcc TDL 99.2%
196038 10.30.1.44:49542 10.30.2.62:902 1c:c17fcc TDL 98.6%
196050 10.30.1.44:49552 10.30.2.62:902 1c:c17fcc TDL 84.6%