Question around a secondary copy.

[Kitkat]

Hi all,
I am backing up a NAS to a local repository (which seems to work fine)-9PM. Then within the same job I have a secondary repository for offsite going to DR over WAN-11PM. (2 hours is enough for the promary job to finish)
The issue is my secondary copy stopped and started from the beginning when the primary copy started at 9pm. That is not normal, right? The secondary copy should continue within the scheduled window and just 'add/queue' the new data found from the next day’s routine daily primary backup to the secondary?

Any help/feedback will be appreciated.
Thanks.

[Gostev]

Hello! I'm pretty sure it's normal for the copy job to stop, because a primary job needs to actively update the source data of a copy job. However, it should not start from the beginning after this. What makes you think it started from the very beginning, as opposed to continue processing only the missing files? Thanks!

[Kitkat]

Looking at the progress of the job when it first started, it was at 25%. Since the 'restart' the progress was noted at 18% and when I dive into the job summary all the data counter details look like they started from 0. Let me see what happens tonight...

[soncscy]

I think what Gostev maybe means is that just because the progress for that specific __job run__ starts from 0, it doesn't mean the content it starts with is "0"

Like, to (poorly) diagram it out:

Original run
Progress visible: 0||{}||||||||100
Files total: 100
Files Processed:1-25

Interrupted, new run:
Progress visible: 0{}||||||||||100
Files total: 100
Files Processed: 26-???

[Gostev]

@Dima P. please comment on the progress display when transfer is resumed.

[Dima P.]

Primary job can 'interrupt' backup copy process, in such case backup copy will be restarted as soon as possible. Backup copy progress will be recalculated based on new data created by the primary job.

[Kitkat]

Veeam confirmed it. The copy needs to complete by the time the primary copy re-runs, if not, it will stop/start again. For the jobs are will never finish within time I need to either disable the primary job or create a temp clone while the primary is disabled.

[Gostev]

Assuming you understood each other correctly, this is not a correct statement by support: there's no requirement for the copy to fully complete by the time the primary copy re-run. While the copy jobs will indeed be stopped, it will resume later from where it stopped versus starting again from scratch. I re-confirmed this directly with the devs.

However, the copy job statistics will indeed be confusing in such case, because to show the correct statistics after interruption would require lots of metadata processing on both source and target side before correct statistics can be displayed, which the job does not do for optimization purposes (as in some cases this can take longer than the actual remaining data transfer).