VM Replication to DR site

[adilj82]

Hello All,

I have 5 virtual machines and recently i have setup DR site . I am backing up my VMs in LAN by veeam, now i want to make a replication of each server to the DR site. I have a 2MB data link to DR site. i have created the replication jobs but the execution is vvvvv slow. could anyone please guide me how to configure the replication jobs to execute faster and also i read something about WAN accelator, how i can make use of it.

BR

[Shestakov]

Hello Adel,
The recommended approach is to backup onsite first, than use backup copy job with WAN acceleration to copy the backups to DR site and to use those backups on DR site as a source for the replication job.
This way you don`t need to take a snapshot of the VMs twice, and don`t need to move the data through the network twice.

[Vitaliy S.]

If you don't want to use backup copy jobs, then you can use WAN accelerators for your replication jobs as well. Just deploy these components in each site. In this case the amount of data that is going to be transferred over the WAN link will be minimized.

[yavuz]

hi, you want, copy full backup file to dr site. after restore with veeam and latest create replication job from center to dr site. god luck

[Shestakov]

I believe, Yavuz meant manual replication seeding. It can also be automated by the way.
I would decide first if you want to use backup copy jobs + replication or replication only. Thanks!

[adilj82]

Shestakov ». how i can run backup copy job to DR site. do i have to create backup repository at DR site as well. for target repository which repository i have to select. ?

[Shestakov]

Yes, you need to add repository in DR site to VBR console. Then you choose it as a target for the backup copy job.
I`m also curious what is the size of the backup file? 2 MBps looks slow. Thanks!

[adilj82]

backup size varies machine to machine the largest i have is 350 GB. i have done the same way as u mentioned. the copy job is running to DR site. i ll let u know how much time it would take. the size of currently running job is 30GB. what speed do u recommend for data link>?

[Shestakov]

The more the better. I`m afraid it`s going to take at least 4.5 hours to copy 30GB over 2MBps (16Mbps) link. Next runs are going to take less since they are going to be incremental, but still. Just keep in mind there is a seeding for backup copy job to accelerate the process of the initial backup copying.
By the way, have you set WAN acceleration?
Thanks!

[adilj82]

I have done with copy job successfully. how i can create a replication job now to use this repository. any help would be appreciated.

[adilj82]

i have created the replication job with using the source as repository instead of live machine. but its again vvvv slow. i am using the repository created on the DR site to replicate the machine on DR site,still its vvvvv slow. what could be the reason.?

[foggy]

What are the bottleneck stats for the replication job? What target proxy server is used to write the VM data to datastore?

[Shestakov]

how i can create a replication job now to use this repository?
... i have created the replication job with using the source as repository instead of live machine. but its again vvvv slow.

Just checking, have you chosen From backup files as a source of the replication job?
If yes, you should have replication job run much faster than before. Otherwise, please provide the bottleneck statistics, as Foggy asked above.
Thanks!

[adilj82]

the primar bottleneck is Source and i have selectedthe source as backup files in repository created on DR site by copy job

[foggy]

What kind of repository is used as the source for this job? What kind of storage is behind it and how it is added to the Veeam B&R console?

[adilj82]

I performed a backup onsite with default repository for a single vm, from that repository i ran copy job to DR site creating one more repository in DR site. On DR site i have .vbm,vbk,vib files Now i am trying to run replication job using this DR repository with WAN accelerators. As shestpakov said i am using from backup files as a source to this job. the primary bottleneck showing me in job log is SOURCE.

[Shestakov]

I guess you don`t need WAN Acceleration for the replication job, since source files for the replication and replication target are on the same site. While WAN accelerators you add to the job may reside on the different sites. Try to run the replica job without WAN. Let us know if that helps.
Thanks!

[dwrice0]

Just reading through this. It doesn't appear the original poster mentioned having a 2nd WAN accelerator at the DR site to pair with the production site WAN accelerator.

[rborhara]

Also make sure disk are thin provisoned else they take forever to restore

[adilj82]

Great. It worked like a charm. it took only 15 minutes and i tested the machine by running on DR site. now the question is do i have to run all three jobs for incremental jobs? backup job to local then copy job and then replicate job ? please comment? BR

[Shestakov]

Just reading through this. It doesn't appear the original poster mentioned having a 2nd WAN accelerator at the DR site to pair with the production site WAN accelerator.

Not sure I`ve got your point. Mine was that WAN acceleration should be used when transferring data between sites, while replication from backup files happens on the same site.

[Shestakov]

Great. It worked like a charm. it took only 15 minutes and i tested the machine by running on DR site. now the question is do i have to run all three jobs for incremental jobs? backup job to local then copy job and then replicate job ? please comment? BR

Yes, since you need to take a snapshot to obtain changes, then you copy those changes offsite with backup copy job and apply them to the replica VM. This way you not only have an updated replica VM, but also backups on both sites.
Thanks!