Hi!
Many databases (like Firebird) have some time limitation for the clients.
These cases possible with client:
a.) It is working, but network connection is temp. down.
b.) The client is crashing
c.) The client is on, but it is not working for N minutes.
When the server experience that client is not speaking in N minutes, the handle and all resources will be freed.
I want to know about this case in PG.
1.) Do I need to PING the server periodically N minutes with "select time()" query to keep the handle live?
2.) What happened, if client crashed, or died? WHen the resources (temp tables, cursors, etc) freed?
3.) Can I configure this per databases, or only per server?
4.) If I use PGDAC or ZEOS to access PG are them make this PING automatically?
5.) What happens if a client died but it's query (long) is running? Is query cancelled by the server, or wait for finish?
6.) If a long query running the client can got timeout because it is cannot work?
Thanks for your help:
dd
