Very strange beahviour on netbalancer.
this evening, after the weekly pppoe connection operator drop, ZS 3.3.1 was unable to change the netbalancer to UP status for this interface. Pppoe is connected, as I can see in the log, but still found down by ZS in the netbalancer and considered as Failed
Just clicked on the “save” button of the netbalancer, and the connection is immediatly put in the UP status..
I’ve immediately tested a manual forced failure by disconnecting the DSL phone line of the pppoe modem. ZS netbalancer rightly change to FAULT again on pppoe. Plugin on the phone line again, and after one minute, the netbalancer pppoe was online again, perfectly.
So, netbalancer seems now to be KO on operator-side connection drop, and OK on end-user-side phone line OFF/ON test.. This is not logic !
This is very frustrating, as the netbalancer runs well from many years, and after testing 3.3.2, in which version it is totaly broken, now it seems to be broken in previous releases too.. May be the 3.3.2 corrupts the database net balancer configuration in some way ?
I will try if I have any time to restore a databse backup.
I need also to test net balancer a full week with the rock-solid old ZS 3.2.1