10GE LAN link flapping on customer router due to SNCP switch
【Problem Summary】Huawei DWDM 10GE LAN link from Hyderabad to Banaglore flapping on customer router.
【Problem Details】Below is the diagram shown end to end connectivity details :-
http://www.thunder-link.com/blog/wp-...0505145411.jpg
The end points which is connected to the router is on the LSX cards. You can refer to the PDF for end to end connectvity details.When there is fiber cut in any ring, there is ODUk SNCP ring switch. Below alarm flow to the router which causes router to flap :-
[IMG]file:///C:/Users/THUNDE~1/AppData/Local/Temp/Rar$EXa0.000/image/ba64e874e75947a89e03fb5af2eecc78.jlr[/IMG]http://www.thunder-link.com/blog/wp-...0505145424.jpg
Although the ring is within 50ms, this will still cause the router to flap. So customer asked us how to avoid this flapping.
Below two causes has been obeserved in this router flapping :-
1. Whenever there is ODUk SNCP ring switch in any ring, it will cause the SNCP switch in the other rings due to SNCP cascade effects so to avoid that we should set hold off time in the ODUk SNCP protection group in each SNCP protection group. but in this case we can`t ensure the switching within 50ms. Total switching time will be :-
Hold-off Time in ODUk SNCP Protection Group +<= 50ms(Means whatever will be th exact switching time).
This will also cause the router flapping. This point only help to avoid the SNCP cascade effects i.e before switching will happen in the other rings it will wait for 100ms.
2. Principles of Router
（1）Router includes physical layer and protocol layer
（2）Condition which can trigger physical layer down: R_LOS/R_LOF/AIS/RDI;
（3）Condition which can trigger protocol layer down:
Whenever physical layer is down, protocol layer will be down;
When physical layer is up but the hello protocol packets between Routers are lost continuously;
（4）Hold off time: the waiting time when physical layer detects the condition which can trigger itself down , suggest setting hold off time as 200ms;
（5）Hold up time: the waiting time when physical layer cannot detect the condition which triggered itself down ,
【Resolution Summary】 Set the hold off time in Router and Hold off time in each ODUk SNCP protection group.
【Resolution Details】
We have to set the hold off time in each SNCP protection group of 100ms, so this will make the actual switching time more than 100ms.Actually switching time will be:-
Hold-off Time in ODUk SNCP Protection Group +<= 50ms(Means whatever will be th exact switching time).
And need to set the hold off time in router should be greater than the above time. To avoid this flapping.
Required to set the hold off time in ODUk SNCP protection group and in router as well to avoid flapping on OSN 1800.
