Hi,
Currently I'm running a PVE node on a dedicated host at a host provider.
On that host I have a firewall VM to handle all traffic (including traffic of PVE itself).
Storage is ZFS based for the PVE server (boot and VM pools)
When I perform a backup of the firewall to my remote PBS server I get some weird behaviour.
The backup starts with making a snapshot, but during that process it apparently gets stuck somewhere and the VM is unreachable for about 30 minutes (so basically the whole node and all coneccted VM's that run over the Firewall VM).
After this the VM is accessible again and starts routing traffic, the problem is because of the network interruption, the backup times out and fails.
If I perform a local backup, or a backup to a network target not going over the firewall I have no interruptions in network traffic. (tested by performing a speed test and a file transfer during the backups, no drops in traffic)
I was wondering how the process of a backup of the VM actually works when backing up to PBS.
From my understanding, pve creates a snapshot through zfs of the VM and then transfers the VM snapshot to PBS.
What I don't get is why there is a network issue when performing the backup while the traffic is running over the VM in question.
The VM should be operational after the snapshot has been taken and before the snapshot is transferred to pbs, so there should not be a network issue occuring, unless I am missing something.
Anyone here have an idea about this?
Kr
Currently I'm running a PVE node on a dedicated host at a host provider.
On that host I have a firewall VM to handle all traffic (including traffic of PVE itself).
Storage is ZFS based for the PVE server (boot and VM pools)
When I perform a backup of the firewall to my remote PBS server I get some weird behaviour.
The backup starts with making a snapshot, but during that process it apparently gets stuck somewhere and the VM is unreachable for about 30 minutes (so basically the whole node and all coneccted VM's that run over the Firewall VM).
After this the VM is accessible again and starts routing traffic, the problem is because of the network interruption, the backup times out and fails.
If I perform a local backup, or a backup to a network target not going over the firewall I have no interruptions in network traffic. (tested by performing a speed test and a file transfer during the backups, no drops in traffic)
I was wondering how the process of a backup of the VM actually works when backing up to PBS.
From my understanding, pve creates a snapshot through zfs of the VM and then transfers the VM snapshot to PBS.
What I don't get is why there is a network issue when performing the backup while the traffic is running over the VM in question.
The VM should be operational after the snapshot has been taken and before the snapshot is transferred to pbs, so there should not be a network issue occuring, unless I am missing something.
Anyone here have an idea about this?
Kr