If we remove the running processor card from NE40E, the router before switching to the second card is shutting down all interfaces. After a while, of course, they go up so redundancy works, but not without interruption.
Form the logs we can see that switchover is working ok but why the services are affected?
Mar 2 2016 11:43:34+01:00 HUAWEI%%01SRM/4/MPUABNORMALPULLOUT(l):MPU4 is pulled out abnormally.
Mar 2 2016 11:43:33+01:00 HUAWEI %%01SRM/4/BOARDRESET(l):MPU4 will be reset, the reason is Reset master because of hardware to slave.
Mar 2 2016 11:43:33+01:00 HUAWEI %%01SRM/4/BOOTMODE(l):Slot 5 has startup with Normal mode.
Mar 2 2016 11:43:33+01:00 HUAWEI%%01SRM/4/MASTERREGISTER(l):MasterMPU5 registered successfully, BarCode is NULL.
Mar 2 2016 12:40:21+01:00 HUAWEI %%01SRM/4/MPUSWITCH(l):MPU5 changed to slave MPU.
After checking the logs and diagnostic information we can observe that switchover mode used is nonstop-forwarding.
During the switchover, all the state information of interfaces will be reinitialized. For example, if the device has 100 interfaces in total, 30 of them are down, others are up.Then during the switchover, system will generate the logs that describe the state information of the interfaces. So we will observe some logs saying XXX interface is down and YYY interface is up. In this case, the interfaces are not really down, but describes the state information. We should make sure the switchover mode is NSR, or some service will be interrupted.As we checked, now the device’s switchover mode is NSF (Nonstop-forwarding), not NSR (Nonstop-routing). We need change the switchover mode form NSF to NSR.The interfaces are not really down, it is the normal behavior. As to the interrupt of some service, because the switchover mode is NSF not NSR, if we change it to NSR, then the service will not be interrupted.
So, the solution is to use NSR instead of NSF.