Proxy selection is going wrong? Or am i wrong?

[denit]

Dear readers,

We have 3 running jobs.
2 jobs are actually doing something.
1 job is waiting for a proxy and is at 0%.

When i look at the 2 tasks, we see proxy 07 being used. This proxy 07 is active on 4 disks simultaneously.

We are wondering why the waiting job is not kicking off because we have 6 other proxies waiting.
WHen we stop the job and restart it, it again says "waiting for proxy".

By now we were thinking, could it be seriously trying to go through proxy 07 while there are 6 proxies doing nothing?
So we edit the job, deselect proxy 07 and leave the other proxies selected and start the job again.
It kicks into action and starts backing up a VM.


Now maybe i am missing something here, but i assumed that if i have 7 proxies assigned to all jobs, veeam would select the next available proxy if it cannot use one that is busy.

It now feels like some jobs are determined to use a certain proxy and wait for it, even when it takes hours.



We are constantly learning new things, but now we are thinking of splitting up all the jobs and spread them over different proxies manually. But that seems a bit weird as this should go automatically right?

We were amazed at what we discovered.

[tsightler]

There is a lot of logic in deciding what proxy to use and it will be difficult to determine the exact cause without knowing a little more about your setup, however, one example that can cause the behavior you are seeing is if one proxy is able to access the VM disks via hotadd while the other 6 proxies can only access via network mode. Because network mode can be much slower and cause significantly more network traffic, if Veeam determines that any proxy has access to a VM via another mode (Direct SAN, hotadd, BfSS), it will wait for that proxy to become available rather than use network mode on the remaining proxies. That's just one example that can describe why you might see this.

[denit]

Thanks for the reply!

For now we have rolled back our "automatic" setting on the transport mode for all the proxies.
All the proxies are on "network mode" again and all our jobs finish in time again.

Now i understand the benefits of "automatic", but in our current situation that seemed to cause trouble.
This may be caused by the fact that not all our VMWare clusters have proxies.
And our VMWare environment is spread over multiple datacenters. Our bandwith, between the datacenters, however is not a bottleneck.

The story for optimization continues, but for now we are happy that at least our jobs run well.

We have already contacted some Veeam officials some time ago and we will continue to do so.
Thanks again.
Sincerely,

Stefan.