Hello, I get this kind of events:
Link lost from Cisco 3550 to 2950:
1) Loop-back detected on FastEthernet0/1. The port is forced to linkdown (14h.35.37)
2) Interface FastEthernet0/1, changed state to administratively down (14h.35.39)
3) Line protocol on Interface FastEthernet0/1, changed state to down (14h.35.40)
- STP is ON
- UDLD is ON (agressive mode)
- Errdisable recovery cause udld is ON
What are these loop back on FX100 link ?
How do I get rid of the link loss ?
Already did that and asked provider to re-certify the fiber which they did and no problem was there.
Problem occure sometimes after 4-5 days without any problem.
Thanks a lot for your help... GREATLY appreciated.
are you sure you need UDLD agressive mode?
I think the problem might be caused by some line outage (maybe one-way) in the provider network. UDLD detects a problem and errrdisables the port then.
You have to enable it manually then.
At the beginning I did not have UDLD ON, and now I put it ON an it had greatly improve the stability of the network.
No link loss for 4-5 days, before without UDLD, links were lost after few hours...
Agressive mode might not be needed, but if I understood well, agressive mode means that UDLD tries to bring back up a link which has been down on one side, which is not the case when not agressive...
I do not get from where this loop back is detected?
IMHO, UDLD itself doesn't improve the link stability. It detects one-way connection problems, etc.
Agressive mode doesn't bring the line back when errdisabeld.
"Port shutdown by UDLD remains disabled until it is manually reenabled, or until errdisable timeout expires (if configured)."
Read http://www.cisco.com/warp/public/473/77.pdf for details.
This is what I wanted to mean:
[i]In aggressive mode, if the link state of the port was determined to be bi-directional and then the UDLD information times out while the link on the port is still up, UDLD will try to re-establish the state of the port. If not successful, the port is put into errdisable state. [/i]
I have the "Errdisable recovery cause UDLD" command "ON" on the switches, so these kind of error are recovered automatically now when it happens...
The strange thing is now the sequence of event I have in my first post...
Many Thanks for your help... it is a bad experience to see links desapearing from the monitoring screen !!!
THANKS IN ADVANCE
Depending of which version your running in boxes...it can be a bug.
Fixed in 12.1(13)EA1.
Were having some alike problems and we fixed it with upgrading the switches IOS and changed all wiring (Cat5) cables. I still think its were bad cables that created the problem. But thats a tip if you cant find the problem.
Bug, might be the reason yes, so strange behavior !
So the version I have is: 12.1 (11) EA1 (fc1)....
So I might try this.... Good God... cool...
Now I will look for that IOS Upgrade....(no idea where but I guess that Cisco's site will help...)...
I'll let you know...
Forgot to say that these link loss are only on FX100 link (Uplink of the 2950 and port of the 3550 which are the core switch... but loopback are detected only on 3550... never saw that error on 2950...)...
This would go against the bug idea... as a bug should bug either on copper or on fiber ?!...
But any way I will update the IOS version... ;)
We have several 3550 and I've seen this behavior with two
of them. After a couple shutdowns on one I finally found a fiber jumper
that was not seated properly. The other incident was also related to
a bad fiber. Haven't seen the problems since.
Thanks a lot for your information.
I will check that too, but since I turned off the keepalive on the FX100 ports, I did not see that, only 4 days I did that but since, no link loss, but I had once 6 days without problem and then on the seven'th one, link were lost again...
Let's hope and thanks fot the tip, I'll check that tomorrow !