Bottleneck source

[dhayes16]

Hello. I know this is discussed at length on many different threads. So we have a new Dell BDR setup for a customer. Installed server 2016 and formatted the storage spaces drive as reFS. Note that they do not have a hardware raid controller (I know). Anyway they have 2 VM's backing up each night at 9pm. The total backup storage is 500gb and I am not seeing any major resource issues.
So we setup a backup copy job with a Veeam cloud connect partner. The backup copy job indicates that the throttle is source. That seems to indicate it is hardware related or perhaps disk subsystem.? I will say that they have a rather slow internet pipe at 50/5 (we are upgrading that in 2 weeks). But I would assume if the internet upload pipe was the issue the throttle would say "network". We have another customer (same exact specs except formatted with ntfs) with about the same sized VM and they say throttle network on their 100/10mbps pipe.

I am just wondering if I am interpreting this correctly.? Why would it say throttle "source" if the bdr itself it no where near loaded.

Thanks

[dhayes16]

Sorry All. Long day...The topic should be "Bottleneck Source" and not Throttle source.

[Gostev]

You are correct, normally this means your backup storage read performance is the bottleneck, which is very strange given their pipe indeed. I would have support review the debug logs to confirm based on performance records if it's indeed the backup storage that is very slow to read. May be some Storage Spaces setting is the culprit? Also, what is Dell BDR?

[dhayes16]

Thank you Gustev. Sorry to leave out important details. The dell BDR is basically the server onsite that is the backup target. Basically it is s dell T30 xeon server with 32 GB of ram with an ssd for the os and (2) 8tb 7200 RPM drives mirrored using storage spaces (software raid). Windows 2016 server standard with Refs. We are planing to replace it with a higher end server at the end of March but we are just testing it out at the moment.
I can dig a little into the storage spaces but performance monitor shows very little load on the server while the backup copy job is running. The disk queue length is somewhere between 0 and .05 on the storagespaces drive. Their network is 1gb. They are backing up 2 VM's and their daily backups finish in 4 minutes. The only bottleneck that I can imagine would be their internet pipe. I guess I am just curious why it would say bottleneck is source and not network.

As I said we have another location with the same specs for the backup target (dell bdr) and a slightly faster internet pipe and same settings in Veeam and it shows bottleneck network. The only difference is the storagespaces drive is formatted ntfs as opposed to reFS.
Thanks for the help. I will open a case if needed.

[Gostev]

Generally, in our testing we find ReFS a tad faster than NTFS, so I'd be surprised if the issue was connected with ReFS. More likely, something around Storage Spaces or server hardware.