Error message:
Just curious if you guys have worked your way through reading SNMP on these devices when they’re stacked without having them time out. Whether I use v2 or v3, if these switches are stacked, I cannot get them to read the standard 60 second polls without timing out constantly. Anyone have settings they’ve tried that have worked decent?
Output of “cmk --debug -vvn hostname”: (If it is a problem with checks or plugins)
No plug-in error, just timeout errors.
I have seen Extreme devices perform extremely (pun intended) poorly on several occasions.
The only “solution” was to troubleshoot SNMP step by step and find sensible timeouts and intervals in Checkmk.
Ugh, I appreciate that, Robin. I started going through that but didn’t know if anyone found a set of settings that worked decent. I know it’s probably all relative, but I’d imagine that a server with decent stats can use similar settings to what others have found for some initial testing. I’ll have to mess around with it and see what I find. Probably do some videos to go through the delays like what’s shown.
Extreme SNMP sucks. The Avaya SNMP worked great and I can still poll 5-stack switches in time.
The problem with “decent settings that worked” is, that for every misbehaving Extreme device, I have seen one, that performs perfectly well.
So it really is about finding the sweet spot for every setup.
SNMP monitoring is a pain. We have several thousands of SNMP devices and it just takes a whole lot of time and troubleshooting. I might write a blog post about this as its an extensive subject
But I recommend running tcpdump on the server that does the monitoring and target your device, you will se in real-time (unless you use v3) how fast your device is responding.
Sometimes its checkmk’s fault - You have a bad written snmp check that takes time to decide it should “not do anything” - or I crashed (we had one plugin that was poorly written so it was targeting all SNMP devices where it should not, and it crashed)
SNMP is UDB based, if there is a firewall between your server and switch this can affect polling as well if links are over capacity
There are also a lot of SNMP settings in checkmk to play around with. I suggest creating a separate site and find an Extreme switch you can play around with to try to find good numbers.
We create labels from our CMDB/IPAM where we have model and manufacturer of devices so we can target specific snmp rules to a model or manufacturer, or even device type. T
Yeah, it’s a good idea for a test site, I appreciate the ideas, everyone. A single site I can test with and screw around with settings would be a good technique. We can take 2x5320 stacked switches with a newly setup monitoring server doing nothing, poll it via SNMP v2 and have it not get back to us within that 60 second window, even when putting the monitoring server on the same subnet. So, no delays in hops through routers and the like. I’ll see what I can do with some of these SNMP settings and hopefully I can make a strategy for other Extreme switches. Other vendors work fine, but we love Extreme otherwise
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed. Contact an admin if you think this should be re-opened.
