Lost networking in XenServer pool slave
I believe this started when I shutdown the pool master and slave to remove SAS HBAs to move to bare metal Debian Lenny 64-bit boxes, after discovering that with XenServer we can't attach VDIs greater than 2 TB to VMs (see this post). It used to work, but apparently was never supported. Anyway, after bringing the two XenServer pool members back up for reconfiguration, discovered that the slave had no networking.
Tried the various suggestions (such as this and this), what ended up working for me was the following on the slave:
xe pif-list
nothing...
xe host-list
note the uuid...
xe pf-scan host-uuid=[uuid from previous step]
Now, "xe pf-list" should show some interfaces, though "ifconfig" still produced nothing. Now we startup the XenServer console.
xsconsole
Place the server into maintenance mode, then configure management interface again.
At this point, I've got networking again and can ping the master. Then under resource pool configuration, select "Join a Resource Pool", provide master info and boom -- the slave shows up in XenCenter.
However, the slave is still just using eth0 as the management interface.
Now, in XenCenter, select the slave, go to "Network" tab, choose "Configure" and change from "Network 0" to "Bond 0+1". Accept the warning and XenCenter should now show "Bond0+1" as the Network and Management Interface.
Leave a comment