1 Would .Cancel() work if I execute it on another thread (the watchdog
one) ? I can live with 'wait till fb 2.5 finall' since I`m really eager to
update anyway :)
2 2.1 Mon tables don`t seem to work every time (I`ve had about 10% success
in "delete from mon$statements" on such long-lasting operations)
3 The network problem happens this way :
3.1 The database server and the work station are working perfectly
3.2 The internet (and vpn connection) fails, but the network interfaces
of the station/server are up so the network stack does not reset
3.3 Is a bit windows specific - the default syn/ack timeout of a
working
connection is about two hours
It`s happened a few times and presents two problems - the server
continues executing a doomed query and the client does not switch to the
offline database because it thinks the connection is alive.
I know N1 is a serverside issue, so I reconfigured the server to a lower
timeout (after 2-3 two-hour lockups (wasntme) I finally got it ...), but
N2 is a bit recursive ... If the client reconnects before the server has
killed the hung connection - it starts waiting on the records it has
locked by itself and therefore hangs again ...
> This property is ignored by provider.
>
> If it's a network timeout the network stack is responsible for
> handling it, if on server (like server is overloaded) then you have to
> cancel command yourself (kill connection, monitoring tables (FB2.1+),
> .Cancel() (FB2.5+)), there's no timeout for command in Firebird. :(
>
--
Sanity is a sin!
------------------------------------------------------------------------------
_______________________________________________
Firebird-net-provider mailing list
Firebird-net-provider@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/firebird-net-provider
