Zhone Shelves Report Non-Standard Interface Operational Status

Tags: #<Tag:0x00007fd88abed5b0> #<Tag:0x00007fd88abed4e8> #<Tag:0x00007fd88abed420> #<Tag:0x00007fd88abed358>

When polling Zhone MX shelves we are getting a number of interfaces reporting as down with an interface operational status of ‘0’. These show up under the Down port count under the summary screen but when you click on the down ports in the status screen the filter does not show ports with a ‘0’ status making tracking these down a bit annoying. Please adjust the logic behind the polling to not report these as down for counts and such. If the interface reports a 0 status, do we even need to show it in the interface list? For these zhone shelves specifically it can be several hundred interfaces which will never have traffic.

Specifically on the zhone shelves - these interfaces are added to the iftable in a dynamic fashion. When you provison a port, when you add a vlan bridge to the port, when you enable remote management to an endpoint. I’ll attach a few examples - theres a lot of extra goo in the snmp tables due to the Zhone ZMS product using snmp to monitor/control the shelves.

Attached is a graphic detailing the problem

Hi @Dan_Fusselman
Unfortunately, we follow the RFC way of speaking SNMP using IF-MIB. This is a standard. If Zhone does not comply, only 2 solutions :

  • They correct their code
  • Somebody writes code in LibreMNS to circumvent their wrong implementation

In case solution 2 is chosen, then you are probably the only one interested in it, as you are the 1st one to complain about it, and probably the 1st one with this device in LibreNMS. Which gives you a lot more motivation (and test ressources) to do it than anybody else.
The way to do it would be for you to start a PullRequest in GitHub, so we can see your code and help you write something “maintainable” that does not break anything else in LibreNMS (if only it is possible).
Bye

Understood, for the scope of this request I would like for the down ports section to show all non-up ports the way the rest of the counters figure it out. I can handle disabled alerting on ports as they pop up, its much easier if I can actually see them to do so.

Or, conversely, don’t count ports that aren’t ‘down’ as being ‘down’ =)

@PipoCanaja Do you think this is something that can be done?

@Dan_Fusselman I cannot tell just by reading this conversation. Any change could be done, but it has to be checked deeply to ensure it does not break something else.

As already said, if you are ready to go for it, you’ll find help on github, and the automatic tests (regression) will help avoid breaking things.