What would cause this behavior? All my jobs do it. When processing they'll be doing something like 20MB/s, but then they'll just sit there for minutes doing nothing, then pick back up.
Happens a lot on the backup copy job, seems (but I may be wrong) to happen less on the backup itself. Both methods of storage are iSCSI stores (virtualized Veeam installs), but they're handling heavier loads no problem.
Hi, William. Only actual data processing (moving) is reflected on throughput graph, meanwhile, preliminary/preparation steps aren’t shown there. I believe, this must be the reason of why you see such gaps on the corresponding graphic. Thanks.
On my jobs, this represents times when the job is "waiting for infrastructure availability" or something like that. Which for me usually means that there are 1 or more VMs that are being processed by a different job and thus are locked and this job cannot process them until the other job is finished with them. Say a replication job is processing VM-A, then my backup job can't process VM-A until the replication job is finished with it. So the backup job sits and waits and has a gap.
It also happens during all the prep/unprep times. One noticeable gap is during replication jobs. When all the VM data is replicated, there is a gap while the job applies retention to the replica VM (removes a snapshot), removes the snapshot of the production VM (which does not happen in parallel), chooses a new VM, prepares the replica VM (removes a snapshot), prepares the production VM (snapshot), and then (finally) starts moving data. This gap can be eliminated most of the time if you have parallel processing enabled.
