Poor performance after enabling encryption?

[ChrisL]

We have a brand new installation of Veeam B+R v10 with brand new SAN storage for VMs and brand new QNAP for backup storage. First run of the backup jobs went very well, reporting 350+ MB/s and backing up 2.8TB in about 2 hours. Very pleased!

Realised we hadn't enabled encryption, so added this, accepting that it'll run a new Active Full at next run, all fine, plenty of space on the new storage.

But this time, the performance is really poor, showing 15-20MB/s and running for 15 hours so far and only about 80% complete.

I can't imagine encryption makes that much difference, it never has before, so something else must be going on, but each link in the chain is showing no excessive load or bottlenecks, nothing is really running at anything other than tickover. Seems that Veeam has just decided to run super slow. Any suggestions greatfully recieved.

[Egor Yakovlev]

Hi Chris,

For sure encryption should not cause that much of a difference with performance. I suspect transport mode has changed or bottleneck moved.
Please check:
- session History view, what were old bottleneck counters at and what are new ones?
- also what was Transport mode in old sessions and what is on new ones? (session statistics window - select VM in the list and check "Using Proxy XXX for Disk YYY [Transport Mode].

/Thanks!

[ChrisL]

Got this sorted now, essentially by restarting the job but then running them one at a time.

For context for anyone else getting the same - I had enabled encryption on a couple of different jobs and then kicked them all of at the same time. In my experience, veeam has always been happy with running multiple jobs so didn't have any concerns. It seems though that when Veeam starts a job, it works out how much bandwidth it has available to it and sets the jobs to run at that rate. Since it was shared beween a couple of jobs at the time it started, it decided to run at a fairly low rate of 15-20MB/s and then stuck with that rate for the remainder of the job. By stopping the job (and dealing with it 'failing' and tidying up the mess afterwards) and then restarting just that one job on at a time by itself, it detected that it had plenty of bandwidth and ran at a nice healty 400+ MB/s and completed in a couple of hours.

It would be handy if Veeam could monitor the available bandwidth during a job and adjust the rate accordingly. It feels a little limiting that it works out an initial rate at the time it runs and then sticks with that rate regardles of whether the available bandwidth later changes.

[Gostev]

Actually, Veeam will always use all bandwidth available at any given time - unless you specifically tell us not to by, creating the corresponding Network Traffic rule. Otherwise, we will always push as much data into the network as it will accept. We don't do stuff like "monitor the available bandwidth during a job and adjust the rate accordingly" in principle.