Topic: Rupture of the user connection in 2 hours
Kind time of days!
There is an application written on C#, DBMS Firebird 2.5.7, SuperServer.
After two-hour silence of the client connected to server Firebird, there is a connection rupture. Under silence I mean - the user does nothing in the program within two hours. After at attempt to fulfill any action (for example to write down the information or to save), there is an error about connection rupture: Error reading data from the connection.
From this that dug out: understood in KEEPALIVE, it works if to look on the intercepted packets in Wareshark, KeepAlive happens each two minutes from the server, parameters changed, it does not influence error presence.
It is interesting that rupture is stable exactly in two hours. On interception of packets... In two hours the packet from the server about connection acknowledgement as I think, tag ACK comes, on what the client to it throws an answer packet. Then the CLIENT the first sends a packet with tags ACK and FIN on port 3050 (the 192.168.132.79-client, the 192.168.100.97-server.):
192.168.132.79 192.168.100.97 64388 -> 3050 [ACK] Seq=405 Ack=469 Win=65024 Len=0
192.168.132.79 192.168.100.97 64388 -> 3050 [FIN, ACK] Seq=405 Ack=469 Win=65024 Len=0
And if the port 3050 from the client as I understand is used, it Firebird itself breaks off connection, not OS.
Server OS: Windows Server 2008 R2 Standart. OS of the client: any, on win 7, 8, XP quits the given error.
Prompt that at me not so if there are questions-willingly I will answer. Weighs that after two-hour idle time it is necessary the client . The server empty packets go to the networks everyone 30 on the server Ping th for check of availability of the server, all packets successfully reach and are returned.
I can system operation , help, a week killed on the decision of this question...