All hosts in a cluster.. 1

[djr111]

Hi,

in /etc/opt/vmware/aam/FT_HOSTS file:

I have one host that is part of a 4 node cluster that does not have all the hostnames and ip's of the other cluster members.

my question is how do I fix it, do I just modify the file, or is there a vi client way to do it.

I guess the bigger question, how would this happen in the first place?

The host in question is complaining that HA has errors.

thanks

 

[TechyMcSe2k]

HA usually has errors when the Networks don't match. The names on each node network configurations must match exactly. Click on the Events button up top and it is usually more detailed for HA errors. Also, DNS has to be configured for each host within the VI client for each Host

________________________________________
Achieving a perception of high intelligence level can only be limited by your manipulation skills of the Google algorithm!

 

[djr111]

DNS configuration is exact on the four nodes, there is no errors in the Event log, it only goes back so far, so I'm not seeing anything there.

thanks

 

[TechyMcSe2k]

Have you verified the configuration of the Network Adapters/networks shown for that host in the VI Client GUI?

________________________________________
Achieving a perception of high intelligence level can only be limited by your manipulation skills of the Google algorithm!

 

[djr111]

Hi,

Yes, the network adapters on all four hosts coincide with each other in regards to how they are designed.


Thanks

 

[nhidalgo]

SSH in each host and ping the other's by FQDN. If they resolve from each host, then Disable HA and then reenable HA on the cluster. I had a similar problem and had to do this

 

[TheLad]

You can also add the hosts ip addresses and FQDN names to /etc/hosts using VI

--------------------------------------
"Insert funny comment in here!"
--------------------------------------

 

[djr111]

disabling and re enabling HA did the trick. Any one know why this error happens causing us to re enable HA.


Thanks

 

[nhidalgo]

i was told one host is the primary for the cluster. If this one has problems, it causes the whole cluster to show errors. I guess a call to vmware support could explain more

 

[Arisap]

HA actually uses 6 nodes as primary for HA regardless of the size of the cluster. I would not rely on DNS to resolve host names because what happens if DNS is unreachable? how is HA supposed to find everyone? Also, an HA event occurs if 15 seconds has elapsed without hearing from a host. Anything you can do to minimize the things that need to be done in 15 seconds the better. All your hosts should be in your /etc/hosts file.

For a peak into what HA is doing under the hoos check out the database table VPX_COMPUTE_RESOURCE