It's UDP, there's no connection, so it's not like an API call over http where you get back some error message that gives you a hint as to what is wrong. If the community doesn't exist, then the agent just drops the request. That request itself says 'hey there, I am a v2c request, here is the community name and an OID, please send me back the info on this one'. Some thing (a monitoring or management system, mostly) wants to get SNMP info from your host, so it sends a request to the host on port 161. SNMP is UDP, not TCP, so there's no persistent connection. The ESXi commands are just a slick interface to the SNMP agent, and if you really wanted you could write your own command-line snmp interface in bash with sed to update the config files and a command to turn on the daemon.

You can get SNMP not to respond to v2c requests (by setting no valid community names), but that is actually all you can do.ĮSXi has a nice little interface you can use to enable and configure the agent, but under the hood there you should understand that there is an SNMP agent that this interface is configuring and starting.

Given that the ESXi documentation talks about 'enabling v2c' any reasonable admin is going to say 'if I enabled it, surely I can disable it' but it turns out you actually can't, and some portion of the misunderstanding is due to how ESXi chose to write those docs. SNMP can be enabled or not (is the SNMP agent on the machine, configured, and listening?), and (if present, configured, and listening) respond to v1, v2c, or v3 requests - or not.īut there is no disabling v2c without shutting off the entire SNMP agent. Here's what is (likely) really going on there.

Esxi cannot start snmp