At 10:13 AM 16/11/2007, Adam wrote:

>--- In firebird-support@yahoogroups.com, "Stephen Boyd" <sboydlns@...>
>wrote:
>>
>> I am running a Windows 2003 server with FB 1.5.3 classic. I have a
>> fb_inet_server process that appears the be dead and I believe that it
>> is causing the OAT to be "stuck".
>>
>> What are the ramifications of simply killing the process?
>>
>
>Firstly, be careful, there will always be at least one
>fb_inet_server.exe running under Windows when the server is running.
>Don't mistake that for a connection.
>
>If you are sure it is a connection, then killing the process will
>cause any uncommitted transactions within that connection to be
>abandoned. This will potentially leave a lot of garbage in your tables
>if that connection had a lot of outstanding work to do.

Another thing to keep in mind on Windows is NOT to allow the Guardian to be installed for fb_inet_server. In an ideal world, one would hope the Guardian would just focus its attention on the listener process and ignore the user processes. Unfortunately, it appears to be not the case...Guardian will faithfully attempt to restart any fb_inet_server process that has died. Then, you have these "ghost" instances of fb_inet_server that refuse to go away...and they never time out, either, because they appear to be stuck in a kind of null state that's neither active nor idle.

My advice is always to run Classic on Win as a service and set the restart option on the service itself if you find you have a problem with the listener processing crashing. FWIW, I've never heard of a case of that happening...

./heLen