04-08-2011 01:01 PM
Hi,
I want to configure a UDLD l in link between asr 9000 and ME 3400, but i thint that it is not supported in ASR 9000, then i am configuring ethernet OAM,
in two sides.
When i disconnected TX fiber of 3400, the asr 9000 sends OAM message and the ME 3400 put the port in err disable status, thats OK,
but in the opposite way it doesnt work, the 3400 dont send packets when y quit the tx fiber in 9000 side, i read in 3400 documentation that 3400 dont send Link Fault OAM BPDUs.
Then how i can do it? i mean, that the port in asr 9000 shut down in one way comunication case?
thanks
04-15-2011 09:03 AM
Hi Adrian
I have the same problem with those two equipments. Unidirectional fiber cuts occurs some times on my network, and I need a mechanism that would trigger flex link and LAG protocols on those situations.
Hopefully somebody would help us both on this topic
Sent from Cisco Technical Support iPhone App
04-19-2011 07:56 AM
Hi Ricardo & Adrian,
the ASR9000 does not support UDLD and there is currently no plan to implement it.
Starting from IOS-XR 4.0 Uni-direction Link Fault detection (UDLF) is supported , which is a feature of 802.3ah OAM.
However, the ME3400 can process the Link-Fault packets, but cannot generate it, which explains the outcome of Adrian's test.
Looking for an alternative for failure detection, you might check for BFD
Hope that helps,
Tim
04-19-2011 08:28 AM
Thanks Tim:
The thing is that I am using ME3400 & ME3400E as pure Layer 2 mode (Carrier Ethernet UNI device), so BFD wont work there
- BFD is not supported on Layer 2 interfaces, pseudowires, static routes, or port channels.
- Although you can configure BFD interface commands on a Layer 2 port, BFD sessions do not operate on the interface unless it is configured as a Layer 3 interface (no switchport) and assigned an IP address
I've been trying with common Autosense=ON on both links member of a LAG (work+stby) group (ME3400 with common LAG, then 2 ASR9K with MC-LAG). In that case unidirectional failure does shut down the whole port and LAG goes to the protection link (around 80ms of service disruption which is ok). Nevertheless when I reconnect the primary link again I have a service disruption of 2 or 3 seconds when LAG goes back to its initial state
I will try the same with Flex link to see what happens there.
Any other idea to get a minimum service disruption with LAG and Felxlink when unidirectional failures occur? As TE FRR and OSPF Fast Convergence can do that work into MPLS domain I suppose it should be something as good as that into Layer 2 (access) domain
Regards
Ricardo
04-20-2011 08:49 AM
Hi Ricardo,
I would not look for any other solution. On top of my head I do not have
any further idea.
In my opinion, it needs to be determined, why there is a 2-3 second
outage, when the primary link recovers. I do not have an explanation for
this right now. The default behavior is, that there is no revert to the
primary link and the secondary link should stay active.
Once this is sorted out, this seems to me like the most practical solution.
For a small check, you may post the related config and mention the used
IOS(-XR) releases.
If you require a deeper analysis, it might be worth to open a TAC case
for it.
Regards,
Tim
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide