LoadBalancer throws Divide By Zero Exception

I wanted to adjust the Max Message Size of my Photon Server. After doing so the Server can't be started anymore. It won't start even after returning it to the initial value. After starting the LoadBalancer it only runs for a few seconds and then stops. The Photon-LoadBalancing Logs simply say

CService::OnException() - Exception: 0x000000014007B42D - INT_DIVIDE_BY_ZERO

I can't seem to find the option to enable callstack collection, so I can't provide a more detailed error message. If anyone could point me to where i can do so, that would be much appreciated.

Best Answer

  • chvetsov
    chvetsov mod
    Answer ✓

    hey, @jKnepel

    it looks like you do not have at all listeners. that is why it may fail in such weird way.

    Could you send your config?


    best,

    ilya

Answers

  • hi, @jKnepel

    that is something new. Please provide as logs from deloyg/log and deploy/bin_Win64/log folders


    best,

    ilya

  • The deploy/win_64/log files. Most recent attempt at starting the server was at 2022-11-09 15:42:51. The deploy/log files were last changed long before the restart and haven't been updated since.

    Is there a way to activate the callstack collection, so I can post a more detailed error message? Thanks in advance.

  • The deploy/bin_Win64/log files. The deploy/log files were last updated long before the restart and consequent crash, so they shouldn't matter much. Is there a way to enable Callstack collection to provide a more detailed error message, as the stack is shown empty. Thanks in advance.

  • chvetsov
    chvetsov mod
    Answer ✓

    hey, @jKnepel

    it looks like you do not have at all listeners. that is why it may fail in such weird way.

    Could you send your config?


    best,

    ilya

  • Hi chvetsov,

    after your hint I looked closer at the server config again and looks like the PerPeerTransmitRatePeriodMilliseconds was set to a much higher number than before the last server restart. When I returned it to its normal state the server started completely normal again. Division with the high PerPeerTransmitRatePeriodMilliseconds most likely resulted in some value becoming 0, thus causing the INT_DIVIDE_BY_ZERO issue.

    Thanks for the help, the issue is now fixed.

    best,

    Julian

  • cool, thank you

    best,

    ilya