hi,

I just read from several network issues so I wanted to post my current experience:

 

I wanted to use unraid with trunk-lacp (supported by my switch) which was running fine for years with proxmox and openvswitch

 

I had following configuration active, which kicked my network conenction regularly:

 

The System is a HP microserver Gen8 with the integrated network ports

 

1.) bond with eth0/eth1 (802.3ad)

2.) automatic ip on the bridge (I would not even need an IP on the bridge itself) & manual DNS (IPv4 only)

3.) vlan support enabled

4.) 2 VLANS (server-NET & client-net) where each vlan had an IP, I configured only for the server-net the gateway

 

Because it was also already mentioned: I tried the configuration for Docker wit ipvlan & macvlan as well, but no effect. - so I set the configuration back to macvlan

 

With this configuration I had network issues:

1.) SSH was kicking my client after some commands - sometimes 6, sometime 10 commands before connection break

2.) webinterface response was multiple times gone & Terminal from WebGUI closed automatically

3.) The dockers were the whole time online (at least it was shown by the webinterface)

 

 

The issue looks gone after I configured the switch from trunk to access, which makes it impossible to use the VLANs and made me reconfigure my docker-installations

 

any possible explanaitions ?

I was playing around a bit more and to me it really looks like the issue is connected to the topic of VLANs - bond is not a problem, but the VLAN configuration is causing the connection errors. Due to too much playing around now some of my dockers are "dead" or "removal in progress". 

Edited October 16, 20232 yr by syntactic-mobilization3954