SNMP strange behaviour

[mustafa073]

Hello.

I'm a complete noob regarding the vSphere Solution. I

managed to install and configure a host with the vCenter software and

another host with the ESX 4 server software. Everything looks fine, and

works except the SNMP monitoring part.

- I have confirmed that the ports TCP 161 and UDP 162 are alowed through the firewall and that the snmpd daemon is running.

I'm

experiencing a strange behaviour when performing the snmpwalk command

over the ESX host, from a remote host (the monitoring software). If the vSphere Client is running (logged on the

vCenter) the snmpwalk command over the ESX host succeeds. If the vSphere client is not

running I get a timeout whenever I execute the snmpwalk.

Does anybody know what I'm missing?

I appreciate any help you can give me.

Thanks.

[AndreTheGiant]

To check if is a ESX firewall problem try to disable temporally the firewall with this command:

/etc/init.d/iptables stop

But check also if there is a firewall on your monitor host.

Andre

[mustafa073]

- It's not the firewall (neither on the host, or the monitoring tool). If it was the firewall, then the snmpwalk command would never work.

- Also, the thing of only working when the vSphere Client is logged in is bogus - It was a coincidence that all the attempts worked when the client was logged in.

- The snmpwalk command works once while trying several times, indicating that there is something wrong on the host -The ESX 3.5 hosts don't seem to have this problem.

[mustafa073]

Does anybody know is there anything else, other than the firewall, that can (occasionally) block the SNMP requests?

I'm out of ideas, so I appreciate any help you can give me.

Thanks!

[DSTAVERT]

Test snmpwalk from the ESX host to verify everything is set up correctly.

snmpwalk -m all -c public -v1 localhost

Have a look at http://kb.vmware.com/kb/1013758 (although it is for 3.5 may have some relevance)

Also try restarting the services

service mgmt-vmware restart