While this issue has been discussed at length both in the communities and in knowledge base articles (ex. http://kb.vmware.com/kb/1003409), I cannot find a single KB article that lists every step I would perform to fix the issue and the order in which I would perform them.
There are two different kinds of ESX to vCenter connectivity problems that I would like to discuss:
- Initially adding an ESX host
- Reconnecting a disconnected ESX host
In my experience, the first problem is almost always caused by a DNS or network connectivity issue. To solve this, first try to add the ESX host by IP address instead of FQDN. If this works, ensure the ESX host and vCenter Server have the appropriate DNS servers, that they can resolve through them (i.e. nslookup) to other hosts, and most importantly that they can resolve to each other. If/When DNS is working properly, try to connect between the ESX host and vCenter server via ping and SSH.
In the case of a disconnected ESX host, the second problem listed above, simply restarting the management services fixes the issue most of the time:
# service mgmt-vmware restart # service vmware-vpxa restart
The important thing to note is that after restarting the management services you may need to wait several minutes in order to confirm whether or not the issue is resolved (i.e. the host reconnects to vCenter). If this fails, in some rare cases closing out of the VI session and establishing a new connection resolves the issue. If the issue is still not resolved, disconnect the ESX host from vCenter and then manually remove the vmware-vpxa rpm from the host:
# rpm -e `rpm -qa | grep vpxa`
Finally, reconnect the ESX host. As a last-ditch effort, if the host is still listed as disconnected or remains “In Progress” during the “Add Host” function, restart the vCenter Server service.
While the KB articles suggest a variety of other options, the above steps have always resolved the issue for me.
© 2010, Steve Flanders. All rights reserved.
