Comprehensive data protection for all workloads
Post Reply
sanman
Novice
Posts: 8
Liked: never
Joined: Sep 17, 2010 1:36 am

How does Veam 4 Replicates data?

Post by sanman »

Hi

We are planning on using Veeam 4.1.2 to replicate VMS on our ESX 3.5 hosts to our DR Site. I am currently setting up a test and wanted some documentation about how Veeam actually replicates the data. The Veeam 4 documentation talks about how Veeam replication does not use open snapshots for replication, (unlike other vendors), so how does this actually work? Since we are using VMWare 3.5 and not 4.0 does this mean that we would be using open snapshots on our ESX hosts? If the product does use snapshots then obviously this will impact on the Source ESX host diskspace and VM performance.

Can anyone help me with some information/documentation?

Many thanks

Dave
Alexey D.

Re: How does Veam 4 Replicates data?

Post by Alexey D. »

Hello Dave,

Snapshots are used in any image-level backup/replication solutions. Snapshots are created when you start backup/replication job, and are deleted after a job finishes.

Dataflow during replication process depends on job mode you are using: for more information, please refer to User Guide, page 15. Thanks!
sanman
Novice
Posts: 8
Liked: never
Joined: Sep 17, 2010 1:36 am

Re: How does Veam 4 Replicates data?

Post by sanman »

Thanks for your reply Alexey. So what does this bit in the "what's new in Version 4.1" pdf mean?

"Unlike competitive solutions that require replicated VMs to run with open snapshots (which prevent Live Migration and DRS)
Veeam Backup & Replication does not have this requirement and instead uses new vSphere ESX(i) 4 APIs to achieve fast replication
cycles."

Does this just mean that Veeam doesn't need the snapshot once the replication cycle has taken place? So for example, if i have a replica and i choose to sync that replica twice a day at say 12pm and midnight. Does this mean at these times every day a snapshot is created whilst the replication takes place then is deleted after?

I understand about the dataflow methods but there's no discussion about the overhead on the cpu/performance of the source VMDK and ESX host. Page 15 of the user guide just talks about impact on the WAN/network/storage. What sort of hit will the source ESX hosts take? If i have 10 heavily utilised VMs on an ESX host and I want to sync them consistently during the day, that's going to hit my ESX hosts quite hard with multiple snapshots being taken every few minutes. Isn't it?

Many thanks for your assistance

Dave

Dave
Alexey D.

Re: How does Veam 4 Replicates data?

Post by Alexey D. »

"Veeam does not keep open snapshots" means once a job finishes they are committed which prevents having VM running off snapshot.

Of course there is an impact on performance, it's a normal situation. You may want to schedule your jobs at non-working time to lower the load.
On the other hand, if you want to set a replication cycle to minutes, there will not be a huge overhead: job finishes fast due to little changes within a VM between cycles => snapshot has no time to grow => its deletion is very fast.
Gostev
Chief Product Officer
Posts: 31524
Liked: 6700 times
Joined: Jan 01, 2006 1:01 am
Location: Baar, Switzerland
Contact:

Re: How does Veam 4 Replicates data?

Post by Gostev »

There is no CPU/Performance/Network overhead on the source ESX host if you use direct SAN access processing mode. Datastore will have some overhead due to snapshot management operation and reading changed data, as Alexey explained.
Post Reply

Who is online

Users browsing this forum: Google [Bot], masahide.k and 98 guests