SNMP polling on Proxy and Trap Receiver on Server

[skipzoid]

I'm not sure if this is a bug or I'm trying to do someting that is not possible,

I have a server and several proxys on different networks, all running v5.4.9 we use SNMP for polling for values and it all works very well feeding data back to the server.

We did have a seperate trap receiver running on a windows box that simply received the trap and translated it via the MIB. It was setup by the support company and we weren't allowed to touch it. They've now gone and we're taking control in house.

So in essence as it was, all the devices were being polled from one snmp collector and then sending their traps somewhere else which gave us two places to look for data. it was an odd setup but it did work. theres something like 2000 devices across multiple networks dotted around the world. We're now trying to bring it all together into a single zabbix eco system.

What I'm trying to do is avoid reconfiguring lots of devices with a new trap destination. We've setup a second IP in the zabbix server in place of the windows box and taken that offline to receive the traps and its logging them in a local file as it should but its only deals with the traps for the devices 'the' server is directly monitoring. it ignores all the traps coming in from all the other devices if those devices are being monitored by one of the proxies.

It seems strange that the server knows the device and so should be able to match the trap to the device. is there away to allow the server to process the incoming trap even through its monitoring is being done by a remote proxy?

[cyber]

Well that's the point, that server does not know the device (you have to separate here polling for data and data processing/calculations)... its monitoring is trusted to a proxy and server does not care about incoming data for that one. It expects data coming from proxy.. I guess it would be a question to support... Or a feature request to be able to use dedicated and separated snmp trap receiver...

[skipzoid]

Well that's the point, that server does not know the device ...

Actually it does!, otherwise it wouldnt know what to do with the data gathered by the proxy.

its going to take hours into weeks to reconfigure 2000+ devices with a new trap destination. also the proxy's are mostly on networks where their local IP is DHCP assigned so it could mean the proxy's IP could change and that would invalidate and re-cause the problem of traps not being received all over again. I can see why the support company originally set it up this way. Also speaking to someone else they use nagios and that can have a proxy on one machine gathering data and the trap receiver on another.

solar winds aparently can do this, so can logic monitor.

seems it should be something that can be an option or even just does it, also it doesnt add much overhead to the server if any as the trap receiver is running anyway and it's the server that responds to the scripts and generates the emails/sms etc.

[tim.mooney]

its going to take hours into weeks to reconfigure 2000+ devices with a new trap destination.

I'm sympathetic to what you're trying to do, and I agree it would be nice if it was possible to do what you're asking about. I agree with Cyber that you should at least log something in the Zabbix bug tracker, to see if that gets any traction.

At least at 5.0.x (and presumably at the 5.4.9 that you're at), the proxy setting is set at the client level, not at the individual monitor level. I suspect that makes it an "all or nothing" setting. Centralized trap reception even when you're using the proxies for polling does make sense, though, so it's worth asking about.

Regarding changing 2000+ devices... I guess I look at it like the problem is elsewhere. I'm not trying to be rude, but if you've gotten to 2000+ devices but you don't have any way to manage them, that's not a monitoring problem.

I do hope you find a simpler solution to the problem than touching 2000 devices individually.

[skipzoid]

I'm sympathetic to what you're trying to do, and I agree it would be nice if it was possible to do what you're asking about. I agree with Cyber that you should at least log something in the Zabbix bug tracker, to see if that gets any traction.

I've done that.

Regarding changing 2000+ devices... I guess I look at it like the problem is elsewhere. I'm not trying to be rude, but if you've gotten to 2000+ devices but you don't have any way to manage them, that's not a monitoring problem.

I do hope you find a simpler solution to the problem than touching 2000 devices individually.

There is a way to manage them, its the individual utils and web consoles the devices offer. there just isnt a centralised select all and update. I'm not aware of any such system that is cross platform, cross manufacture and architecture and technology that would do that.

[cyber]

Actually it does!, otherwise it wouldnt know what to do with the data gathered by the proxy.

Thats why I also said "you have to separate here polling for data and data processing/calculations".. Yes, it knows, what to do with data coming from proxies, but from polling side, if host is assigned to proxy, then it ignores those traps coming to server. Would your proxies also have trap receiver in them, they would be accepted.

To distribute to multiple different platforms you could use ansible. Not exactly out of the box thing, but if you do it once, you have it ready for future changes.. And you will be happy about it.. They do have a lot of built in functions for different devices. https://docs.ansible.com/ansible/2.9...k_modules.html