Disaster / Continuous Backups issue
Hopefully, someone would be able to help or shine some light on this.
I have two servers that are both connected to the network via eth0 and also connected to each other via eth2 & eth3 on a bonded network.
Server one has the Bonded IP of 192.168.10.1
Server two has the Bonded UP of 192.168.10.2
Both servers are on a pool of their own but in the same DC next to each other in the rack.
Within the pools, I have set them both up to use the Bonded network for migration.
When doing a Live Migration both servers do so using the Bonded Network.
When I run a disaster backup or Continuous that's when the issue starts.
Backing up a VM on Server two to Server one it does so via the migration network as it never shows up on my DC bandwidth graph which only monitors eth0. The backup was around 3GB so it would be easy to spot.
When backing up a VM from Server one to Server two it does so using the eth0 connection which then does so by connecting to the public IP which in result costs bandwidth.
So my main question is why does Server 2 > Server 1 backup via the Bonded and Server 1 > Server 2 backup via eth0?
Is there a log that shows the backup and also shows which network connection was used?
As far as I can see both settings are the same with both installs of XO
Both servers are running the same version of XCP-NG and XO.
I would like Live Migrations and Continuous backups to use the Bonded connection and Delta backups to use the eth0 connection since they go offsite
Thanks in advance for any help you can give.
olivierlambert Vates 🪐 Co-Founder🦸 CEO 🧑💼
@olivierlambert Thank you