OSPF LSA Issue ICX7150

[Vladimir_Kos]

Model - ICX7150-C12-2X10GR

Version - 08.0.95rT213

Each time a single port withing a vlan changes operational state, the VE initiates an LSA removal form the tree at the same time that it adds the LSA to the tree. This happens simultaneously and if there is an interface flapping due to a damaged cable, it creates excessive LSA churn that create random pauses in traffic.

[BenBeck]

Hey Vladimir, 

Is the VE flapping during this event? This issue is probably best tracked via an official case (see my signature). 

[Vladimir_Kos]

VE is stable during this event.

Now we have only Warranty for hardware replacement.

[BenBeck]

Could you share config if possible? If you are willing, it may be worth testing 10010g_cd8. Note: Many changes between 8.x and 10.x, so proceed with caution. 

[shane_owens_i88]

Ben, I work with Vladimir and I can provide a bit more context. We will have support renewed in a few days. We go through this every year with renewals taking to long to QA then get new quotes and by that time the original quote is expired and procurement needs things split. I'd love to say it's rare but it's been happening since the Brocade merger where they joined all FDOT agencies into a single company umbrella and there are 9 of us and we all operate independently.

Anyway to answer your questions the VE is stable the whole time. What triggers this is we have a device that goes bad for some reason the copper port starts flapping rapidly, sometime multiple times per second. Opther ports in the VLAN remain up and error free, but each time that port flaps we see and LSA for the subnet on the VE withdraw and then broadcast. We've noticed weird CLI delays while working on devices and tracked it down to this behavior. 

Also, we are aware of the differences between 8.x, 9.x and 10.x we run them all in production. 10.x has it's own issues with ICX7550-48f devices having the first port in each bank of 8 flap or completely shutdown at random. We noticed this on port 17 first, but after testing it happens on port 1,9,15,25,33, and 41, That is another case pending renewal of our service contract. We maintain close to 800 ICX devices in total just in the District Vladimir supports