If we do have the following setup
Node A is parent of Node B and Node C all have the minimal shared cell. And we had an dedicated cell between both pairs. Now Node B powercycles and resync to Node C.
Now it happens that Node A has a cell to Node B an there will clear that because it's not needed anymore. But sending the 6PCLEAR on the dedicated link fails on linklayer because there will be no ACK from Node B. In that case we end up in endless CLEAR Requests because sixtop_six2six_sendDone
always resets the sixtop statemachine. If the timeout would occure the cells would been removed then but the timer is only started in case we received an ACK.
can we do something like the following without beeing inconsistent with the draft?
, yes, we can. I will take this issue into my sprint and will be fixed in next release.
i guess this is solved by isn't it
With current code in the develop branch, sixtop will send on autonomous cell which won't have this problem.
to be fixed
The fix has been merged into develop branch.