UDP for Interface Counters

Need some help with interface counters. I’m in somewhat of a Catch-22. Here’s my Scenario, so please bear with me for a moment. We have 50+ remote locations. We have regional techs that maintain and connect our PCs and other peripherals at these locations. In their daily haste, there are a lot of interfaces that are left connected on the switch from the patch panel, but no devices at the wall jacks. Inevitably, I will get a call that the switch is full, where in reality it was just a case of them not cleaning up the data cables during moves or removal of equipment.

I’ve written a Perl script that goes out to the switches that they want to clean up, gather the counters from all of the interfaces, and report on the ones that have zero Bytes-IN (OID 1.3.6.1.2.1.2.2.1.10.x) AND zero Bytes-OUT (OID 1.3.6.1.2.1.2.2.1.16.x) which gets sent to me in an E-mail that I forward to them. This tells them which interfaces have not been used. I also report on the switch uptime, so they can determine if it’s been unused for long enough. It’s a nice report which gets e-mailed in a concise format. The trouble is that I have to run the report for them. I realize there are methods to get them this report so they can run it by themselves, but we’re trying to encourage Orion NPM for this type of information.

Here’s what they get now from the Perl Script in an E-mail:

Mon Aug 24 17:38:46 2009

Ports with no Usage for Brookfield sw-21-01

WS-C3560-48PS-S S/N:555555555 (10.21.2.1)

System Uptime: 392 days, 22:29:57.33

FastEthernet0/15 Bytes In: 0 Bytes Out: 0

FastEthernet0/24 Bytes In: 0 Bytes Out: 0

FastEthernet0/46 Bytes In: 0 Bytes Out: 0

Ports with no Usage for Brookfield sw-21-02

WS-C3560-48PS-S S/N: 555555556 (10.21.2.2)

System Uptime: 392 days, 22:30:00.58

FastEthernet0/2 Bytes In: 0 Bytes Out: 0

FastEthernet0/5 Bytes In: 0 Bytes Out: 0

FastEthernet0/6 Bytes In: 0 Bytes Out: 0

FastEthernet0/9 Bytes In: 0 Bytes Out: 0

FastEthernet0/32 Bytes In: 0 Bytes Out: 0

FastEthernet0/45 Bytes In: 0 Bytes Out: 0

The Catch-22 is that I can report on the counters in Orion, but that means that I have to monitor all of the interfaces on all of the switches (over 8,000 interfaces). If I monitor all of the interfaces, I’ll be gathering a whole lot of unnecessary clutter that I really am not interested in. Not to mention the up/down status every time someone shuts down a PC as well as the unused ports that will be “down” at all times.

I would like to create a new UDP that will gather the interface counter information once a day and then I can generate a report for them to run with ONLY the unused (zero count) interfaces. Or what would be even better is I would like to present the information on a “View” for them to look at that shows ONLY the unused interfaces on a switch with uptime information.

Any advice or help creating this information would be appreciated.

Jeff…

Find more posts tagged with

Accepted answers

All comments

bshopp

Couple things:

1- in 9.5 we have a new feature in which you can set user ports with a field called unplugged, so when they undock and go home for the evening, you do not get 5000 alerts for node down, they will just show as unplugged.

2- have you figured out the UnDP to pull what you want from the node?

jwbyers

We've only been on 9.5 for a few months, so I'm still getting up to speed on the new features. I was not aware of the one you suggested.

1- We have 9.5 SP2. I am going to have to experiment for a few days on the statistics I get. This may be enough information for me to pull a report. If I set all ports to unplugged and report on interfaces with zero bytes transferred in say 90 days, I may have enough to write a report. I am still concerned however, that we will be gathering a ton of statistics just for the sake of cleaning up our data wiring. Thanks for this suggestion.

2- I have the OIDs that I need, but I have not added the UnDP. I presume there has to be an interface in the databse to map the results to, which would require managign the switch with NPM. Prior to your suggestion, I was not interested in managing the interfaces. Again, I will have to experiment.

I realize there may not be a way for NPM to give me what I want if I don't manage the ports. But I would really like the information to show up in NPM. I may have to bite the bullet and gather the data anyway. Thank you for your quick reply and the unplugged tip. I will experiment for the next few weeks. Until then, I'll be interested in anyone else's experiences.

Jeff...

jwbyers

Quick Question.

1- I started to experiment. The only way so far that I found to achieve this is to choose the switch on the web console and select each interface, edit the settings and submit. Is there any way to do this in a "mass" configuration / selection? Selecting each interface in the web page will be a VERY painful experience. At about 15 second per interface, it'll take me more than 36 hours of straight typing to get through all of our interfaces. This is a best case scenario.

Thanks inadvance!

Jeff...

Nicki

If you not mind sharing your perl script, I would like to try some experimenting with this also. This could be a very helpful report, even if it were used as a link on the Orion Web Console. Thank you.

dblickle

Whoops mouse drift, didn't mean to suggest the non-answer. I mean to reply how useful this report would be to me and then mark it so I get notified of replies. If someone builds this report I want it!! :-)