5 out of 7 agents fail incremental backup

[jarzi]

VBR server 11a
Linux agents 5.0.1.4493
RHEL 7.9 reposerver

I have 7 agents behind a firewall running under oracle VM. The network is 10G from end-to-end. Problem is that only 2/7 can take incremental backups, other five just wait for an hour and fail with "read: End of file". Active full backup works fine. All the agents are in the same network and use same firewall rules.

I've tried different jobs, repo with immutability and without but can't figure out what's going on.

There are about 60 other agents that work without problems so the veeam server side should be ok.

Case ID 05091885

[jarzi]

Also, if I stop the job gracefully, it still sits there until the hour is full and the gives up.