ELRP ingress port = egress port

[eyeV]

Hi everybody!

I configured standalone ELRP-client on my SummitStack (two x480). So, after a couple of days I saw strange log messages...

ELRP.Report: Slot-1: [CLI:Default:2] LOOP DETECTED : 35590 transmited, 23 received, ingress slot:port (1:7) egress slot:port (1:7)ELRP.Report: Slot-1: [CLI:Default:2] LOOP DETECTED : 35550 transmited, 5 received, ingress slot:port (1:4) egress slot:port (1:4)It happens periodically, but only on stack. I have nine standalone x480 switches and ELRP works correctly on it. I've never seen this situation on it at any rate.

Can anybody suggest me something?

[davidj_cogliane]

We see this with some older Cisco phones when switches get rebooted. After some research we found the Cisco phones are designed with a closed circuit that can loop back edp packets before the phone is powered up. We are currently forced to go through the switches after reboots and enable the ports that are disabled due to these "phantom loops".

Has anyone seen ports reported as elrp disabled but actually be active?

[davidj_cogliane]

Wow,

This bring back bad memories.

In the interest of providing complete resolution.

We discovered that these older Cisco phones have a physical loop that opens when they receive power. The behavior we saw appeared to be the result of the data link coming up before power could get there to open the loop. As a result the elrp frames made it back to the switch and disabled the port.

[Drew_C]

If there is a known issue, it's one that I don't recall seeing. Just because I don't recall it doesn't mean there isn't an issue though.
Instead of disabling/re-enabling, you can try the

restart port

command since that's a little quicker (note it won't reset inline-power, just the link).

[davidj_cogliane]

Last time I saw this we disabled the port and re-enabled it to get it to clear out of the elrp. I have also seen the ports because stuck in the disabled state. After hours today I plan to try disabling the port and re-enabling it.

I guess I was trying to find out if this is a known reported bug.

Thanks,