Backup very slow

[paolone]

Hello

I've a PVE and a PBS new . I've a job from PVE to PBS .
In a day I've added a NFS storage to PBS for a second destination and the same day a disk was with a smart error with 8 sectors. Replaced and resilver disk.
I've added a Job in PVE to make backup in NFS and first Job it lasted 4 hours (the first full).
After that Backup is very slow:

Before: Total running time: 7min 56s Total size: 2.689 TiB
After: Total running time: 1h 57min 33s Total size: 2.689 TiB

produzione is the datastore with local disks in PBS
nas02 is the datastore mounted via NFS in PBS

In the PVE log

Before At the end of the first VM 101

101: 2024-06-28 21:01:00 INFO: Waiting for server to finish backup validation...
101: 2024-06-28 21:01:02 INFO: backup was done incrementally, reused 740.72 GiB (98%)
101: 2024-06-28 21:01:02 INFO: transferred 9.37 GiB in 56 seconds (171.3 MiB/s)
101: 2024-06-28 21:01:02 INFO: adding notes to backup
101: 2024-06-28 21:01:02 INFO: Finished Backup of VM 101 (00:01:01)

And in PBS



After

101: 2024-06-29 21:12:57 INFO: Waiting for server to finish backup validation...
101: 2024-06-29 21:13:00 INFO: backup is sparse: 477.49 GiB (63%) total zero data
101: 2024-06-29 21:13:00 INFO: backup was done incrementally, reused 740.90 GiB (98%)
101: 2024-06-29 21:13:00 INFO: transferred 750.00 GiB in 774 seconds (992.2 MiB/s)
101: 2024-06-29 21:13:00 INFO: adding notes to backup
101: 2024-06-29 21:13:00 INFO: Finished Backup of VM 101 (00:12:58)

If you can see: verification start after VM backup and not at the end of the job like in the day before. Verification is very slow and I think slows the backup.

Can you help me to understand problem ?

My best regards

 

[paolone]

Hello

solved the problem: was the dirty-bitmaps. If you have 2 different backup job, every backup delete the dirty-bitmap and for every job restart the dirty-bitmap creation. Now I've 1 job and 1 sync job