Jump to content
Welcome to our new Citrix community!
  • 0

Management interface bonding

Paul Hutchins1709155747




I have a new host with 8.1 that is the start of a new pool.  It's got x4 10GbE ports that I am setting up as follows:


Eth0+Eth3 management interface plus VM external networks.

Eth1+Eth2 storage network bond.


The eth0+eth3 bond is set as LACP based on source mac address so that the tagged VM networks can be on this bond, but we are finding that we have had to leave the management VLAN as untagged, which seems to work, but it's an odd configuration.  Also finding that if I unplug either eth0 or eth3 I lose access to the management IP  (eth0 and eth3 go to different switches but both ports configured the same).


For eth1+eth2 this is just an active-active bond to two ports on different switches as well, both are untagged and configured on the storage network subnet.  I find that again during tests if I pull either eth1 or eth2 while running a ping to the IP configured on the bond it drops out and does not come back up even if I leave it for several minutes, the load balancing does not seem to kick in (only comes back up if I plug the cable back in again).  I was sent an article by Citrix support about switch stacking but as a workaround use active-passive on the bond settings.  I tried active-passive and the same behaviour occurs.


Any help greatly appreciated!

Link to comment

6 answers to this question

Recommended Posts

  • 0

Both untagged and tagged primary management bond VLANs should work under 8.1 (they had to be untagged under earlier versions of XS, not sure off-had before exactly when, bust sometime early 7.X I believe). No hard running then untagged, of course. Are the same NICs in use on all pool hosts and the numbers and PCI slot positions are also identical on all of them?



Link to comment
  • 0

Thanks, for the responses.  The switches are separate back to a core and at the moment its a new pool so I only have one host so far, the second one I am adding does match with cards in the same PCI slots but I have used the interface-rename command to designate eth0 eth1 etc to the order I want to use them in.


At this point we think it could be either a networking configuration or patching issue, someone from the networks team is going to go in and investigate with us to make sure something incorrect has not been done with those aspects of the setup.



Link to comment
  • 0

Thanks for the help, it was good to confirm I shouldn't really be experiencing an issue, I had someone go back in and check the patching while I had someone from the networking team also available.  We found the patching was not quite right, once resolved I tested disconnecting a NIC cable from each bond and did not experience any problems.  But one step forward, two steps back as I am having some problems removing stale iscsi entries from another host that is ready to join the pool!  I'll have to post a new question on this as I have been unable to resolve myself so far.

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Create New...