We added a new host to our Management Cluster like we have dont many times without fail however this last time was troublesome as we have run our of IP addresses in the IP pool.
As NSX stated we still have 5 left we had to release the IPs via the API and then we could add the vxlan vibs successfully however now in this cluster we have an error stating that the Cluster is already configured for VXLAN and we are unable to use any of the resolve functions as when this is selected the same error appear as in the screenshot below.
Some website said it may be something in the cluster that is causing a conflict but apart from that there is very little published about this error.
Creating a new Cluster and moving the hosts to the new cluster would more than likely fix the issue but because we run VIO changing to a different cluster would cause all sorts of issues with VIO is suspect.
By the way all the Hosts and VMs in that Cluster still work fine.
We are using NSX 6.3.1 vSphere 6.0U3
Any ideas ?
0.- not clear on if after adding ESXi host to the custer the error happens or after prepare to VXLAN
1.- verify please check host.event log file and paste what is in the logs here
2.- check vmkernel log as well
3.- check if VTEP ip pool has enough ip addresses
Firstly do check NTP/DNS of the new host which is added to the cluster.
Hi
I have confiemed the there is more than enough vxlan IP addresses however I did need to free some using the API as there were not being release when removing other hosts.
I have posted the esxcli.log and notice an error. I cannot see to find the host event error log that you ask
I have also looked at the vmkernel.log and cannot see anything that stands out.
Hello Again.
All DNS and NTP info are correct and is exactly the same as every other host that gets added to the cluster.
I believe that the vibs have been installed as the hosts works perfectly fine and as the error is cluster based the error is not particular to any given host.
I can ping the vcenter from the host and reverse DNS is working too.
This looks promsing...The vmkwarning.log
Maybe this but if all the hostsin the management cluster require a reboot this is going to be major downtime as we have VIO and NSX in the management cluster and VIO especially takes about 30minutes to shutdown and start up
Looks like VIbs are loaded to the host . If feasible try rebooting the host once and confirm the status.
can you please be so kind to post this logs as well?
/var/log/esxupdate.log
thanks
VTEP Pool shows not assigned any IP's
ESXi servers are rebooted but still having same issue
Did you have any luck solving this? we had to change the VXLAN VLANID on a fully configured and are now stuck for one host in one cluster...... It seems a bit overkilling to destroy and rebuild from scratch....