Copy Job sometimes fails

[geets]

Hello,

we have a problem at one a customers site. Sometimes Copy Jobs are failing.
Now I suspect they are failing because due to some overlapping with other jobs (copy or backup) the take VERY long to finish and thus blocking each other.

I am trying to figure out the bottleneck here. In my investigation I came about an interesting metric that I need help to understand:

21.04.2020 01:01:53 :: Hard disk 1 (147,3 GB) 5,7 GB read at 84 MB/s
21.04.2020 01:05:24 :: Hard disk 2 (784,2 GB) 14,8 GB read at 3 MB/s

Its from the statistics of a Copy Job. Those two lines up there tell me that a the Copy Job processed the second Hard disk 20 times slower than Hard disk 1.

So here's my question:
What could be reasons for that? The slower processed Hard disk is larger than the other and I think I see that pattern on other Jobs too..

Help would be appreciated. I can provide more data if necessary.

Thanks for your input in advance
Alex

[Vitaliy S.]

Hi Alex,

Before we start investigating it can you clarify one thing for me: are you using VM copy, backup copy, backup or replication jobs? If this is a pure VM copy, then why aren't you using replication?

Thanks!

[geets]

Hi,

of course, I see I was a bit unclear on that.
I am talking about backup jobs and backup copy jobs.

I will provide some more detail on the setup:
Generally, we have backup jobs that run dayly, weekly and monthly.
and we have copy jobs that dayly, weekly and monthly.

There are two Hyper-V Hosts. All VMs on these HV-Hosts are backed up to a single dedicated storage repository. We call it BackupRepo.
Most of the time the backup jobs are processed at a rate of 30mb/s to 150mb/s. With a single exception where it was 6 mb/s: (again large hard disk -> slow processing?)
[these lines are from a backup job:]
21.04.2020 22:09:08 :: Hard disk 1 (1,1 TB) 16,2 GB read at 6 MB/s [CBT][fast clone]
21.04.2020 22:09:10 :: Hard disk 2 (350,0 GB) 45,9 GB read at 19 MB/s [CBT][fast clone]

Then there are copy jobs that copy to another repo: a QNAP-NAS.
The copy jobs are generally slower than the primary backups. Between 4MB/s and 10MB/s

I am aware that this is a complex situation that's fairly hard to debug from remote but I hope someone can hint me in the right direction.

[foggy]

It is not necessarily the disk size. Can be also the source storage load and its random I/O capabilities as the changes for the particular disk might be scattered along the source backup file.