This may be better suited in the networking section, but it mainly focuses on xcp-ng.
So we’re in the midst of upgrading from vmware to xcp-ng, the main difference is that we’ve previously used individually managed hosts on the free license, whereas now we’re using xcp-ng with xen orchestra.
My plan was to have the hosts connected to trunk ports, and have the management interfaces on their own vlan, while allowing the VMs to be on the appropriate vlan. The issue I’m having is when xcp-ng is connected to a trunk port, connections drop and time out constantly and I can’t find any reason as to why. Hosts across the network will randomly drop out of XO, VMs whos adapters go through those interfaces are unable to pull update via apt.
I’m currently thinking a possible workaround is to keep all the machines to a native vlan, and use L3 ACLs on the switches to allow the hosts to get back to XO, but I can’t find any info on what port or protocols I’d need to allow in the ACL to permit this communication.
Can anyone suggest anything here? Ideally I’d like to solve the connection timeouts when using trunk ports but I’ve not been able to find anything on this either, so I’ll settle for know what to allow in the acl