Host Manamgent not responding

[JimKnopf99]

Hi,

i have an issue with one of our hosts. First, it was disconnected from vCenter and not pingable from the network as well.

I tried to restart the management network and test it. During the test i recive the error "Resolving hostname failed"

Ok

I open a ssh session an make a nslookup and ping and vmkping to other servers and all works well. So there is no DNS issue.

I check the logfiles and see one error in the  syslog

"RoutingInfo: Unable to restore VMkernel default gateway (IP Adress): Unable to set VMkernel gateway Address. Please verify your IP settings and try again"

But the IP Settings are OK. There where no changes befor the issue comes up.

I also done a services.sh restart

There are one issue that i saw: Connect to localhost failes: connection failure

Any ideas how to restore the default gateway?

But esxcfg-route show me the right gateway!

Frank

[a_nut_in]

Were there any storage or any other changes? Can you connect to the host using a vSphere client to the host directly and using the IP address?

Would help if the vmkernel logs or vm-support bundle could be shared. But the easiest way to get back the host would be a reboot.

[JimKnopf99]

No,

there are no  changes.

I am not able to conntact with the vSphere client. I can open a SSH Session.

There are running vm´s on the host. Therefor it is not easy to reboot the host.

Frank

[jdptechnc]

From where are you testing whether the host is pingable or able to be connected to from vSphere Client?  If you're trying it from the vCenter Server desktop, try it from somewhere else - the fact that vCenter Server dropped the connection makes the network communication in general between vC and the host suspect.

Your gateway is probably OK if you didn't change anything there.  I see that error message sometimes when I restart services, but it resolves.

Are your hosts and vC server on the same network?

Are you able to connect to host with https?

[JimKnopf99]

Hi,

thanks for your help. I reboot the host. But I could not connect to the host either.

Only a ssh connection is possible. I am not sure if https works. I move the vm's to other hosts.

Tomorrow I will take a closer look to the log files.

The host was pingable from multiple host. Also Dns work fine.

The host itself is also able to ping different machines on different networks.

Frank

Am 21.02.2013 um 18:03 schrieb jdptechnc <communities-emailer@vmware.com<mailto:communities-emailer@vmware.com>>:

VMware Communities<http://communities.vmware.com/index.jspa>

Host Manamgent not responding

reply from jdptechnc<http://communities.vmware.com/people/jdptechnc> in VMware ESXi 5 - View the full discussion<http://communities.vmware.com/message/2199199#2199199

[a_nut_in]

Hey Frank,

Here's the deal. If even after the host reboot you are unable to connect to the host, there are just two possibilities, the port 443 might be blocked (ssh uses port 22) or the hostd service might be crashing.

Can you try the following command and share the output?

less /var/log/vmware/hostd.log or less /var/log/hostd.log depending on the version or ESX you are running.

Also try a telnet hostname of ESX 443 to check if this is a ports issue.

So if the hostname is localhost

telnet localhost 443

Regards

a

[a_nut_in]

If you are moving all VM's off this host, try renaming the /etc/mware/hostd/vmInventory.xml and pools.xml and reboot once. This will help if there is a VM on the host causing this issue

Regards

a

[JimKnopf99]

Let me check this all tomorrow. Thanks until know for sharing your thoughts

I will give you an update.

Frank

Am 21.02.2013 um 19:30 schrieb afunnynut <communities-emailer@vmware.com<mailto:communities-emailer@vmware.com>>:

VMware Communities<http://communities.vmware.com/index.jspa>

Host Manamgent not responding

reply from afunnynut<http://communities.vmware.com/people/a_nut_in> in VMware ESXi 5 - View the full discussion<http://communities.vmware.com/message/2199256#2199256

[JimKnopf99]

Hi,

the server has problems resolving his hostname. After getting rid of that issue the host works as aspected.

Thanks for your help.

Frank