Hi,
I recently configured situation similar to what is explained in the user manual. But would like a bit more advice on it.
My problem now is that I have network equipment plugged in on a switch. Weekly there is a power shutdown for testing purposes. For NODE_DOWN events I send SMS notification to my team. So in those cases, I receive quite a bit of those notifications for the switch and the equipment connected to it. When reading the manual I thought that the example for the situation is fitting perfectly my case, if the switch goes down, don't send any of the notifications for equipment plugged in to this switch. However in most of the cases, the NODE_DOWN event for the switch arrives after a lot of NODE_DOWN for the equipment. Is there anything that I can try to do to avoid this. Maybe increase the frequency of the checks for the switch?
Thanks,
Hi!
Situations will not help this. Source of this problem is that node behind the switch/router can be polled before switch or router that is down, so NetXMS still consider it up and so mark node as down before marking router as down. This problem will be addressed in upcoming 1.2.5 release (issue https://www.radensolutions.com/chiliproject/issues/180 (https://www.radensolutions.com/chiliproject/issues/180)).
Best regards,
Victor
Hi,
I saw that 1.2.5 has been released earlier this month and that the problem of this thread has been addressed. Is it done automatically or are there any specific things to configure to have this working?
Thanks,
Hi!
This is done automatically. The only requirement is that NetXMS server must know full route to the nodes, so intermediate routers must be accessible and provide routing information.
Best regards,
Victor
Cool,
I'll try that in a couple of days. That means that node downs won't be reported if the switches that the device connects to is down? What if the switch comes back up
All my routers and switches are Cisco 2960S and 4507, and NetXMS has access to those switches via SNMP, this setup should be fine right?
Hi!
Yes, it should work. You should get SYS_NODE_UNREACHABLE instead of SYS_NODE_DOWN, and node should go to UNKNOWN state instead of CRITICAL. It then become NORMAL again when connectivity restored.
Best regards,
Victor
Hi,
I upgraded my installation with version 1.2.5 yesterday on my customer site and luckily it's the day that they also schedule some power testing so I was able to check this functionality. However I didn't seem to see any difference with before. I still got all my NODE_DOWN events, and I didn't see any SYS_NODE_UNREACHABLE.
Thought that SYS_NODE_UNREACHABLE is actually in the list of events, so version wise I think I'm good.
Anything I could do to get this to work?
Thanks,
Hi!
Is it possible to send me diagram of how nodes and management server are connected? You can send it to my e-mail to avoid posting on the forum.
Best regards,
Victor
Hi Victor,
Did you get my email? I sent you a diagram of how my network is arranged.
Thanks,
Hi!
Yes, I got it. Sorry for not responding, I didn't have time yet to work on it. I hope I'll do that in a next few days.
Best regards,
Victor
Ok,
Thanks a lot.