When icecc-scheduler retries for "failed to accept an incoming connection", it sets timer value based on the time_offset_table table.
But when attemp count is the maximum, there is a bug in code in CompileServer::updateInConnectivity function in icecream/scheduler/compileserver.cpp file.
The limit checking shuold be before that next timeout value calculation:
so from current code:
and by the way, is there a way to scheduler to notice the icecc daemon clients appearing again and making the retry and not waiting this timer to expire?
When icecc-scheduler retries for "failed to accept an incoming connection", it sets timer value based on the time_offset_table table. But when attemp count is the maximum, there is a bug in code in CompileServer::updateInConnectivity function in icecream/scheduler/compileserver.cpp file. The limit checking shuold be before that next timeout value calculation: so from current code:
to:
and by the way, is there a way to scheduler to notice the icecc daemon clients appearing again and making the retry and not waiting this timer to expire?