Simples Interface Down Problem

Bluejohnny · Post by **Bluejohnny** » Wed Dec 30, 2009 9:43 am

Hello,

How can I detect a interface down problem on a switch? I have my Thresholds up and running and cacti tells me when the equipment is down, but what about a single interface? Can I do it with Threshold plugin?

Post by **TheWitness** » Wed Dec 30, 2009 12:42 pm

I have a nice plugin for that. However, if you don't have it, which you don't, you need to graph the ifOperStatus and when it goes to 0, you can send an alarm. However, having an unsolicited message (aka snmptrap/snmpinform) is the better way to go, unless you don't have configuration access, then graphing the ifOperStatus will allow you to thold on the value.

TheWitness

Bluejohnny · Post by **Bluejohnny** » Thu Dec 31, 2009 9:33 am

Letâ€™s see,

If I understood you well, I have 3 possibilities to solve the problem, and I have configuration access on my switch.

1- Install a Plugin. I think this is the simpler solutionâ€¦ can you tell me what plugin you use for this?
2- Monitor the ifOperStatus. How can I do that ? What king of graph should I run?
3- SnmpTrap. Ok, for this possibility, I should have a syslog running on my server and some plugins to manage this, correct ?

Post by **TheWitness** » Thu Dec 31, 2009 10:30 am

1- Install a Plugin. I think this is the simpler solutionâ€¦ can you tell me what plugin you use for this?
2- Monitor the ifOperStatus. How can I do that ? What king of graph should I run?
3- SnmpTrap. Ok, for this possibility, I should have a syslog running on my server and some plugins to manage this, correct ?

1. Working on releasing that as OSS. It's not a priority at the moment
2. Create a Graph Template and Associate with the Interface Data Query, then install THold and trigger on that.
3. Yes, check under the Plugin Announcements forum. There is a great syslog+snmptrap plugin there.

TheWitness

Howie · Post by **Howie** » Thu Dec 31, 2009 12:11 pm

It's also worth noting that absolute zero traffic is pretty rare on a working interface. You might be OK to just set a minimum threshold of 0 on your interfaces.

Usually STP, OSPF, CDP etc etc generate *some* traffic every polling cycle on a connected interface.

(there was a problem with getting Cacti to parse the ifOperStatus output from snmpwalk. I don't know if it's still a problem, but potentially it makes graphing that a little harder)

Bluejohnny · Post by **Bluejohnny** » Wed Jan 06, 2010 7:08 am

1. Working on releasing that as OSS. It's not a priority at the moment - OK!
2. Create a Graph Template and Associate with the Interface Data Query, then install THold and trigger on that. I will try this option but I have some doubts, it's the first time i will make one, thold is up and running for quite some time.
3. Yes, check under the Plugin Announcements forum. There is a great syslog+snmptrap plugin there.For the moment I will try the second option

1 - Created a data Template. It's ok ?
2 - Created a Graph template. It's ok ?
3 - If the two previous steps are ok... shoud I create a new a new DataQueries or use an existing one? I tried to create a new on, but I have problems with the xml file.

Can you help ?

Bluejohnny · Post by **Bluejohnny** » Thu Jan 07, 2010 5:55 pm

Bump! Some help please.

Post by **gandalf** » Fri Jan 08, 2010 4:19 am

I'd start looking out for the XML file that was not located properly
R

Simples Interface Down Problem

Simples Interface Down Problem

Who is online