notconnect vs notconnect:

[stonent01]

Recently I've seen some switches show ports with a status of "notconnect:"  What does the colon at the end signify?

I don't have it documented where all I've seen it but I just now saw it on a 6 stack of WS-C3650-48PD running 16.12.06. About 85% of our switches are CAT3K series running CAT3K_CAA-UNIVERSALK9 and running that exact IOS, so I don't have a lot of IOS diversity to compare it to, and we've trickled in a few CAT9K from some projects, but they're all running 17.xx and I don't think I've seen it there.

Is this a bug that makes it do that, or does that : at the end have any significance?

I asked a few other people in my network circle and none recall seeing it before.

[balaji.bandi]

I have not seen that too before even on 3650 / 3750 / 3850 / 92/9300 so on

Looks to me like some Bug (the coder forgot some leftovers?)

show interface status | in notconnect (what is the outcome with out

[Joseph W. Doherty]

"Looks to me like some Bug (the coder forgot some leftovers?)"

That would be my guest too assuming this is consistent on same platforms with identical .bin, i.e., just "sloppy" .bin building.

[stonent01]

---

@balaji.bandi wrote:

I have not seen that too before even on 3650 / 3750 / 3850 / 92/9300 so on

Looks to me like some Bug (the coder forgot some leftovers?)

show interface status | in notconnect (what is the outcome with out

---

That just shows all non connected ports including the one with the : at the end. On that switch stack it's just that one port like that.  But I did see some switches with multiple ports with the : at the end.  Here's a snip of it with the : in it.

[Max Jobs]

Would you please run "Show version" command, so we can have a better view.

[mradams33]

Good chance it's actually "notconnect: TDR running". You can see the whole output in a "show int g1/0/4" as it's truncated in the show int status. Saw this today too, and it's because there was a test running on the port.

[Richard Burts]

@mradams33 offers an excellent observation. 

[Joseph W. Doherty]

---

@Richard Burts wrote:

@mradams33 offers an excellent observation. 

---

Indeed.

Not mentioned in my prior reply.  I did a project (years ago) running Python scripts, interacting using terminal/SSH access, to analyze running routers and switches for standards compliance.  What I found, not all IOSs (just then?) would respond exactly the same to exactly the same "show x" command.  When parsing output, would encounter minor differences, similar to one device might have embedded colons where another did not, etc.  I recall (?) some of these differences applied to different platforms running the same IOS version, although not the same IOS binary.

[mradams33]

Yep. That's been the most frustrating part of building automation tools using screen scraping. The outputs from show commands are not uniform across the platforms so I've had to put in several ifs to cover all of the different platforms we've had over the years for each show command. That's where tools like DNAC and Ansible come in handy. They have standardized output and handle the difficult part for you.

[Joseph W. Doherty]

Unfortunately my referenced project predated DNAC and Ansible, etc., but definitely better ways now available vs. screen scraping.  Good you mentioned.