hostipaddrsdiffer: cmd remove failed...

[seangar]

I get this error on my 29th server in the farm of 30 (last one dag-nabit) when trying to get HA online for that host. Only the console NIC and IP are runing on this box right now.

hostipaddrsdiffer: cmd remove failed: ft_gethostbyname and hostname -i return different adresses: 10.xxx.yyy.113 and 10.xxx.yyy.115

I assume this error is telling me I have a conflict some where. Our DNS servers are returning the correct info. At first I did have a few duplicate IP issues so this is where it might have stemmed from, however that has been all cleared up. I reformatted the remaing 8 servers in our farm with correct IP info.

Any info appreciated

S

[jhicks]

What version of VC are you running?

2.0.x only supports 16 hosts per HA cluster. 2.5 supports 32, but I'm wondering if they still recommend you don't approach the limit.

[seangar]

VC 2.0.2 and all hosts 3.0.2. There are 3 clusters in this farm. this particular cluster has 6 hosts in it. I dont think this is the issue...

I did find this post : "Feb 29, 2008 2:21 AM by boabie"

This might indicate I have an issue with the hosts file in the ESX server...

S

[admin]

aam caches dns info in the FT_HOSTS file. For VC 2.5 this file has been moved to /etc/opt/vmware/aam. Remove the stale entries from that file and try reconfigure the host for HA.

Ah, I see you're using VC 2.0.2. In that case the FT_HOSTS file should be directly under /etc

[seangar]

Not running 2.5.

[admin]

looks in /etc/FT_HOSTS

[seangar]

I dont see this etc/FT_HOSTS using "ls" and "ls |more". I am connected via DRAC card on this dell so size of window is funky.

I did have the wrong name in the hosts file and corrected that (nano hosts), then rebooted. Still same error, but the edit above did hold so the hosts file is correct now.

[kjb007]

If you've fixed resolution, right click on hosts and reconfigure for ha. If HA did not install, it could be why the /etc/FT_HOSTS does not exist.

-KjB

[seangar]

Does FT_HOSTS get created when the ESX install is performed or when HA is called on in VC ?

I am trying to figure out why this ESX hosts is missing the file, thus render a solution.

S

[admin]

It gets installed when HA is configured. You might try look at other hosts in the cluster to see if a stale entry for the new host is in their FT_HOSTS files.

[seangar]

I checked every ESX host in the farm and in each cluster. Not 1 has an entry in either the hosts file or the FT_HOSTS file for that particular server. Would the VC server itself need host entires to help it along (does VC look at the windows hosts file) ?

TIA

S

[kjb007]

HA relies heavily on name resolution, whether it be through the hosts file, on windows for VC and /etc/hosts for esx, or DNS. You have to be able to resolve the shortname and the fully qualified name. If you don't have DNS entries, add it into the hosts file, then reconfigure HA.

-KJB

[seangar]

I do know about DNS (and its importance), how without it - your hosed... non-issue. DNS is correct, triple checked.

There is someything else go-in on with this box. I will try to updtae the hosts file on all the ESX boxes in the cluster and VC box. See if that works.

Any other suggestions are welcome.

TIA - S

[kjb007]

Is this server multi-homed? Do you have more than one SC address? If you type hostname, and then do an nslookup on that name, is it the IP address you have in the problem server's /etc/hosts file?

-KjB

[seangar]

NSlookup I tired yesterday as well as double checking all DNS entires in NetID.

DNS resolves correctly. Not multihomed. We have 29 other servers running just fine in this farm.

Its a buger for sure.

[ac57846]

I had this with a customer site a while go.

I presume that you've done the "rule of four": Forward, Backward, short & long for DNS from every ESX server in this cluster for the dag-nabit last host.

I'd also go for a walk for half a hour, that way the DNS cache's will expire & your hosts will no longer remember out of date or incorrect IP addresses.

Al.

[seangar]

Forward, backward, short long were all verified.

I reformatted the host a few minutes ago, works fine now. I will award pts to the most replies in this case for sticking with it.

Thanks to all

Sean

[admin]

I suspect the host's IP address in DNS doesn't match the value in /etc/hosts. In the original post, I assume that the hostname -i IP address was the correct one, is that right? In that case, if there is no /etc/FT_HOSTS file, then I am confused.

If you really wanted to get to the bottom of it, you could have run

> strace ./ft_gethostbyname <hostname>

and compare the search path for the name resolution against that of

> strace hostname -i

and see where they diverge. They should be indentical if you have no FT_HOSTS file.

[kjb007]

All of the FT_ are part of the aam which is used for vmware HA. It should not exist until you turn on or configure HA.

-KjB

[VMGenie]

Had the same error, this was my fix

1. Run "hostname -v -f " this will show full details on the host. To correct simply run "hostname server.domain.com"

2. Run "vi /etc/hosts". Press "i" to fix whaterver may be incorrect

3. Run "service network restart" and then "service mgmt-vmware restart"

4. Re-enable HA on the cluster\host

[Tengineer]

Its easy.

1.Put the host into maintainence mode.

2.remove the host from vcenter.

3. add it again, but this time, instead of adding the host with its IP-address, use the fully qualified name. xxx.something.com

you cant add the host with the IP address.!!!