Suspicious times when run remove-unknown

1

Hi guys,

i did a remove-unknown to make sure DB and files are in sync. I have 2 server running and i check the time and consumption of both backups.

|Computername|Abbilder|Dateien|Alle|
|SERVER230|14.28 GB|10.11 TB|10.12 TB|
|SERVER231|29.49 GB|2.63 TB|2.66 TB|

Server230 has 4 times more than server231 on files in the backup space. But remove-unknown behaves the other way.

2022-06-26 13:34:43: Removing unknown for client “SERVER230”
2022-06-26 13:41:32: Removing unknown for client “SERVER231”
2022-06-26 14:34:26: Removing unknown for client “XXXX”

Server230 updates takes ~7 minutes to run through while server231 takes 53 minutes!?

Last backup run on Server230: Indexing of “F” done. 253 filesystem lookups 1845036 db lookups and 39 db updates

Last backup run on Server231: Indexing of “F” done. 288 filesystem lookups 716076 db lookups and 205 db updates

So how is it possible that remove-unknown runs must faster on server230 data entrys? Even Server230 has 4 times the backup size and 2,5 times on files as server231?

Any hints?

2

Shot in the dark here, but are server 230 and 231 identical in the hardware department? Perhaps one has a slower CPU, or has a higher load? Are the storages local and with identical disks/raid card/raid configuration setups?

3

It is a 5 Node Cluster with a 10 GBit 110 TB CEPH Storage, so they are running on identical hardware in the cluster.
Urbackup VM is Linux in the same Cluster on a different Node with Urbackup Database in the CEPH Storage and Urbackup Storage is on a 60 TByte NAS with 1 GBit connection.
Urbackup VM has 24 Gbyte memory and 4 CPU’s, Windows VM’s (230 & 231) have also 24 GByte memory and 4 CPU’s. Every Node has 128 Gbyte memory and 16 Xeon CPU’s. Write speed in the CEPH storage is ~500MB/s, on NAS is near the 1 GBit max.

I see urbackup only use 1 CPU when it do the remove-unkown, so i don’t think its a harware problem.?