2.9.17 - PPPoE users falling off one by one slowly

Finally went into the deep end and threw 2.9.17 back into production again as 1 of our PPPoE servers (the rest are ‘safe’ on 2.8.28 until 2.9 evolves…)

Anyway, thanks for the routing fix. It appears OSPF works great this time.

BUT

PPPoE just seems ‘weird’ and something doesnt stack. It accepts connections fine and took 240 users last night for todays production test. All appears OK so far except the fact that this server drops another user approx every 10mins. Logs report ‘peer is not responding’. The three other 2.8.28’s are picking them up and holding them just fine. Configs between the three 2.8.28’s and the one 2.9.17 are totally identical.
It ‘looks’ like the TCP MSS thing isnt working. There are no entries under ip firewall mangle as there is on the 2.8 boxes…

We try to stick with the saying ‘if it aint broken dont fix it’. …but we need the BGP functionality on 2.9 that is not found in 2.8…

The question is to go back to 2.8.28 and stay there or pursue 2.9.x ?

Idea’s, anyone ?

You should send a supout file to support@mikrotik.com

John

this is not a bug and nothing is broken since that version (2.8 is very old btw). this is a behaviour change and is not related to pppoe