snmp monitoring of Virtual Hosts via Physical Node? [message #5693] |
Sun, 27 August 2006 16:33 |
tchipman
Messages: 28 Registered: June 2006
|
Junior Member |
|
|
Possibly an absurd question: Is there any "straightfoward" way to be able to poll the physical node via SNMP to get some characteristics about the VirtualHosts running therein?
ie, I would love to be able to log via Cacti/SNMP ...
-CPU utilization / Load average, Memory & Network utilization for each / all virtual nodes running .. simply by polling the physical node itself.
I suspect most of this info is available via the /proc tree of the physical node .. but I'm not sure how to bridge this info over to SNMP Pollable content. (Possibly would require using the net-snmp perl module for example?)
I suppose I could always run net-snmp inside each virtual node and poll info directly, but that seems somewhat absurd.
Any comments are appreciated...
--Tim
|
|
|
Re: snmp monitoring of Virtual Hosts via Physical Node? [message #5713 is a reply to message #5693] |
Mon, 28 August 2006 19:34 |
|
curx
Messages: 739 Registered: February 2006 Location: Nürnberg, Germany
|
Senior Member |
|
|
Hi,
thoughts:
install the a snmp daemon! on the hwnode, (like that one http://net-snmp.sourceforge.net/ or a snmpd, provided by your distro),
and write a (sh|perl|php|or_other_lang)script, which print the information out of the ve's /proc or other info src (i.e. with vzctl exec <veid> 'grep venet0 /proc/net/dev' - for the veid's network-traffic )
When walk through MIB, the script is executed and print out the information.
You can read more in the snmpd.conf under "Executables/scripts"
[Updated on: Mon, 28 August 2006 19:39] Report message to a moderator
|
|
|