Errors in PhotonServer RC9
Options
Sergiy
✭
Hello
We have following errors in PhotonServer log:
Server is 3.0 RC9, client - Flash RC4
Please see attached logs:
Following is from Photon server log after shutdown command, perhaps it will help to identify the problem:
We have following errors in PhotonServer log:
CTCPPolicySocketServer: connection timeout CThreadPool::Process - Exception - CTCPPeer::HandleIncomingCommand() - Unexpected! No CLR application: 2looks like these errors are preventing everything else to be logged. The server itself works without any problem.
Server is 3.0 RC9, client - Flash RC4
Please see attached logs:
Following is from Photon server log after shutdown command, perhaps it will help to identify the problem:
5676: 10:01:24.214 - Service is stopped... 5676: 10:01:24.214 - Server shutting down... 5676: 10:01:24.214 - Notifying CLR applications of shutdown... 5676: 10:01:24.229 - Shutting down TCP Outbound Connection Manager... 5676: 10:01:24.229 - Shutting down ENet Outbound Connection Manager... 5676: 10:01:24.229 - Disconnecting all outbound peers... 5676: 10:01:24.229 - Shutting down ENet host... 5676: 10:01:24.229 - Disconnecting all peers... 5676: 10:01:24.229 - Shutting down socket servers... 5676: 10:01:24.229 - Shutting down ENet thread pool... 5676: 10:01:24.245 - Shutting down TCP inactivity timers... 5676: 10:01:24.245 - WaitingForShutdownToComplete: TCP:0.0.0.0:9991 5676: 10:01:34.245 - OnShutdownFailedToComplete: TCP:0.0.0.0:9991 5676: 10:01:34.245 - 5676: 10:01:34.245 - WARNING: *************************************************************** 5676: 10:01:34.245 - WARNING: Socket servers failed to shut down in time... 5676: 10:01:34.245 - WARNING: Socket servers forcibly shut down... 5676: 10:01:34.245 - WARNING: You should expect peers to leak in this situation 5676: 10:01:34.245 - WARNING: *************************************************************** 5676: 10:01:34.245 - 5676: 10:01:34.245 - Shutting down CLR applications... 5676: 10:01:34.245 - Shutting down business logic thread pool... 4500: 10:01:34.261 - CManagedHost::OnDomainUnload() - 3 5676: 10:01:34.276 - Shutting down I/O thread pool... 5676: 10:01:34.276 - Flushing ENET buffer allocator... 5676: 10:01:34.276 - Flushing TCP buffer allocator... 5676: 10:01:34.276 - Flushing TCP buffer handle allocator... 5676: 10:01:34.276 - Flushing stream socket allocator... 5676: 10:01:34.276 - Sockets were active during flush... 5676: 10:01:34.276 - Flushing inbound datagram socket allocator... 5676: 10:01:34.276 - Flushing outbound datagram socket allocator... 5676: 10:01:34.276 - Destroying ENet thread pool 5676: 10:01:34.276 - Destroying Business logic thread pool 5676: 10:01:34.276 - Destroying servers 5676: 10:01:34.276 - Destroying TCP Outbound Connection Manager 5676: 10:01:34.276 - Destroying resolver 5676: 10:01:34.276 - Destroying TCP Proxy Connection Manager 5676: 10:01:34.276 - Destroying TCP inactivity timer 5676: 10:01:34.276 - Destroying ENet host 5676: 10:01:34.276 - Destroying ENet host 5676: 10:01:34.276 - Destroying CLR dispatcher 5676: 10:01:34.276 - Destroying CLR applications 5676: 10:01:35.979 - Destroying I/O thread pool 5676: 10:01:35.979 - Destroying ENet buffer allocator 5676: 10:01:35.979 - Destroying TCP buffer allocator 5676: 10:01:35.979 - Destroying TCP socket allocator 5676: 10:01:35.979 - Destroying ENet inbound socket allocator 5676: 10:01:35.979 - Destroying ENet outbound socket allocator 5676: 10:01:35.979 - Destroying performance counters 5676: 10:01:35.979 - 5676: 10:01:35.979 - WARNING: *************************************************************** 5676: 10:01:35.979 - WARNING: CPhotonPeer leaks detected 5676: 10:01:35.979 - WARNING: CTCPPeer leaks detected 5676: 10:01:35.979 - WARNING: *************************************************************** 5676: 10:01:35.979 - 5676: 10:01:35.979 - Shutdown complete...
0
Comments
-
This was new for us but relatively easy to track.
The error "CTCPPeer::HandleIncomingCommand() - Unexpected!" happens when a client sends operations before a initialization for the application is done.
In this case, the Flash API is most likely the cause: Not all outgoing operations are cleared when a connection is lost and when it is connected again, operations might be sent before this init.
This thread has a similar issue: viewtopic.php?f=5&t=1675
We will fix the Flash library asap. We will post the update in the blog and as tweet.0 -
Thanks. We'll look forward for update0
