My first thought is that you have to disable TCP Checksum
offloading. There should be plenty of documentation on how to do this from a
simple Google search.
From: xen-devel-bounces@xxxxxxxxxxxxxxxxxxx
[mailto:xen-devel-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf Of Gian Paolo
Buono
Sent: Tuesday, October 05, 2010 11:23 AM
To: xen-devel@xxxxxxxxxxxxxxxxxxx
Subject: [Xen-devel] [VLAN tagging on XenServer 5.6 with Dual Port
10GbEFCoE]
we have a big problem, my hardware are two host
XenServer 5.6 server IBM x3850 with Dual Port 10GbE FCoE (qle8152) and switch
Cisco nexus 5020
For the CNA driver I followed this instructions:
After reboot, I have the new ethX (FCoE) and everything
seems to work fine.
But we have a big problem only if we use the nic FCoE
(10GbE) with the vlan trunking.
(after configuring the switch and create the vlan on both
XenServer)
- If I create 2 VMs on the same VLAN and on the same host,
they can talk to each other (ping and ssh and all protocols work correctly)
- If I create 2 VMs on the same VLAN on different hosts they
cannot talk to each other (ping are OK but ssh and all protocols TCP do not work
properly)
We have tested on RedHat5.5 with xensource (on the same
hardware) and the VM (with vlan trunking) are ok.
But we want to use XenServer 5.6 and buy the license to use
the support because we have buy many hardware.
We opened a case to qlogic
Citrix has also opened a problem