[Check_mk (english)] Specifying parameters of snmpwalk

Antti Mäkelä Antti.Makela at vintor.fi
Tue Apr 17 18:40:00 CEST 2012


Hi,

  I've set up a few checks that are used for polling ESXi 5.0 servers with SNMP.

  The thing is, directly from their releasenotes (see below), the ESXi SNMP daemon is not exactly stable.

https://www.vmware.com/support/vsphere5/doc/vsphere-esx-vcenter-server-50-release-notes.html

----
The snmpwalk command returns an error message when you run it without using -t option
When the snmpwalk command is run without using the -t and -r options for polling SNMP data, the VMware embedded SNMP agent does not show complete data and displays the error message, No response from host.

Workaround: When you run the snmpwalk command, use the -t option to specify the timeout interval and the -r option to set the number of retries. For example: snmpwalk -m all -c public -v1 host-name -r 2 -t 10 variable-name.
----

  As such, every now and then the SNMP poll fails. This creates errors like "CRIT - SNMP Error on X.X.X.X, execution time 10.1 sec", when the ESXi simply refuses to give proper answer. 

  We have tried increasing the number of checks (max_check_attempts) to 3, but I think that in this case the issue should really be handled as part of the check command itself. Is there some way for me to pass custom parameters for the snmpwalk command (in this case the retry and timeout options)?

  Thanks.

- Antti


More information about the checkmk-en mailing list