Cisco switch monitoring

[PFrog]

Hello,

I have Cisco network switches that I am wanting to monitor in a control logix processor (Cisco WS-C2960X-48FPS-L).

I want to monitor these so I know if we have a network "blip" (IT reset, network issue, etc.), or if I have a RIO unit that is powered down/faulted. I plan on ignoring the devices for alarm condition if I lose comm to the switches themselves.

I don't believe there is an AOP available for these, so I was looking for some help with coding if possible. These interruptions can be momentary (under 1 min in duration), so something quick would be good.

Thanks,

Keith

 

[Nova5]

You might be able to watch a few cards in a remote IO Rack that passes through that switch. if you lose all cards you can suspect a network fault. There also might be a Link status in the network card of your local rack. If Link status is Down then the switch went offline, but not all actions on a switch will take the link down.

use a GSV to pull data from the rack/cards. Class Name : Module, Instance Name : NameAssignedInIOTree, Attribute : EntryStatus. If your EntryStatus isn't 16384, there is a fault for the card.

 

[PFrog]

I am monitoring the RIO units with GSV commands, but would be unsure if the issue is related to them (fault/power down), or a network switch dropping out that they connect to.

 

[JeremyM]

You can't directly monitor these switches through EtherNet/IP since they don't implement any CIP objects. Stratix product lines (rebadged Cisco) do.

 

[PFrog]

Yes, that is the issue. I didn't know if I could ping the switch (for example) or something else like that with coding.

 

[Nova5]

As long as you have multiple racks and all drop it can work as a rough check. I don't think you can make a CLX Ping anything off hand.

 

[JeremyM]

You can’t and it doesn’t tell you much beyond whether an ICMP response made it back to the controller.

A couple good spoilers on doing a GSV on a “MODULE” object:
- It’s a misnomer. You’re actually interrogating a local Logix IO Map/Connection object Status attribute.
- This is good becuse what it tells you about the Class 1 (implicit) connection integrity is the key.
- From this you can derive whether the endpoint device is “connected” or “disconnected” to the originator (the PLC).
- It’s always faster and a better tool than pinging.

 

[PFrog]

No, I need to be above the RIO unit at the switch level. I need to monitor switch comms.

 

[PFrog]

You can’t. It’s also slow and doesn’t tell you much beyond whether an ICMP response made it back to the controller.

A couple good spoilers on doing a GSV on a “MODULE” object:
- It’s a misnomer. You’re actually interrogating a local Logix IO Map/Connection object.
- This is good becuse what it tells you about the Class 1 (implicit) connection integrity is the key.

Yes, that is why I was wondering if there was a faster way of doing it within code. I don't really need much in the way of feedback, just if it is responding. I can look to see if there is another way of accomplishing this. They make modules for network scanning, but none of them can integrate into control logix that I am aware of.

 

[Nova5]

No, I need to be above the RIO unit at the switch level. I need to monitor switch comms.

Stratix time.

 

[PFrog]

If I change them, I own them instead of our IT group... There is quite a bit of other stuff on them. Was trying to avoid that lol.

 

[Nova5]

Honestly.. thats how it SHOULD be. IT Should NOT have their mitts in Controls.

 

[JeremyM]

Then add a switch (properly configured Stratix or EIP Moxa/NTron) for IO and tell IT to pound sand when they want it back on their switch "because policy."

 

[JeremyM]

No, I need to be above the RIO unit at the switch level. I need to monitor switch comms.

Diagnostic feedback shouldn't inform how you respond to a connection loss in the first place.

 

[PFrog]

Then add a switch (properly configured Stratix or EIP Moxa/NTron) for IO and tell IT to pound sand when they want it back on their switch "because policy."

IT would be more than happy to turn those over to me lol. They handle quite a bit out on the production floor. They would be more than happy to point the finger at me if the production network drops out.

If it was just my equipment, no problem. I could look at running more cable, getting switches, protecting/maintaining them, but I was hopeful there would be an easier solution that I was not aware of.

I may be able to get a RIO unit in each switch panel and monitor those (GSV). On the same switch power, always on. If they drop, most likely the switch panel did versus being on the plant floor more exposed to failure or powered down.