Scheduling multiple backup copy jobs

[Dsmelser]

I have four backup jobs grouped by function:
-FileServers (3 VMs)
-DatabaseServers (3 VMs)
-MailServers (3 VMs)
-Misc.Servers (10 VMs)
I also have four backup copy jobs to copy the data offsite (using the WAN accelerator).
The backup copy jobs are configured for daily backups.

All the backup copy jobs start at the same time and veeam processes the VMs in round robin order (one VM from FileServer backup copy job, then one VM from the DatabaseServer, etc.).

Is there anyway to get all the VMs from one backup copy job to complete before it starts processing the VMs from another copy job?

[foggy]

David, what are the maximum numbers of concurrent tasks specified for the source and target repositories?

[Dsmelser]

Four and Four and I have parallel VM and virtual disk processing turned off.

I'm using the WAN accelerator. The WAN accelerator limits things to one concurrent task. When I checked the status early this morning, one job was idle and three were working (one at 10%, one at 27%, and one at 75% percent complete). Looking further, only one VM was actively being transmitted all the other VMs had a status of complete or queued.

When veeam finished with VM#1 in the Fileserver backup copy job, it then started sending the a VM from one of the other jobs instead of sending VM#2 in the FileServer backup copy job.

[veremin]

Hi, David, wouldn't it be easier in this case to put all backup job within one backup copy job and process them one by one in required order? Thanks.

[foggy]

Also setting synchronization intervals for backup copy jobs to start with short periods between them would help.

[Dsmelser]

I do not want to put all the backups into a single backup copy job:
- I want to keep the backups file sizes small enough to fit onto portable media for seeding/rotated media purposes.
- I want options for different backup frequency/retention policies.

I have staggered the synchronization intervals with five minute gaps. I will report tomorrow if this had any effect.

[Dsmelser]

Staggering the synchronization intervals does not help.

[foggy]

Basically, synchronization interval start time does not handle the job start itself but it is the time when the job begins to look for the new restore point to process. If there is a new restore point already available at that moment, the job will start processing it immediately. So if you schedule your regular backup jobs in the way where new restore points are always available at the beginning of the backup copy job sync interval, corresponding tasks will be queued one by one and all the VMs in the given job will be processed before the tasks from the next job.

However, the more reliable approach would be to deploy additional WAN accelerators on the source side.