Subject | RE: [firebird-support] internal gds software consistency check (cannot start thread) |
---|---|
Author | Alan McDonald |
Post date | 2006-02-26T23:19:44Z |
> Hi Group,have you checked task manager processes? I've seen this on a machine where a
>
> Firebird (1.5.0.4306) Superserver crashed on one of our production
> Windows 2003 Server machines over the weekend, well sort of. It
> crashed to the point where connections were refused but the process
> did not die so guardian did not restarted the process. I am unsure
> whether existing connections were affected at this stage. This morning
> we manually restarted the Firebird service and it is now working fine.
>
> The following line appears 19 times (with different timestamps and two
> or three different databases) in the Firebird log:
>
> FBSERVER (Server) Sat Feb 25 23:48:47 2006
> Database: C:\MyDB.fdb
> internal gds software consistency check (cannot start thread)
>
> After these errors, the following message appears for every attempted
> connection until we restarted the process this morning (hundreds of
> messages):
>
> FBSERVER (Client) Sun Feb 26 00:01:36 2006
> INET/inet_error: connect errno = 10061
>
> I know that SS spawns a different thread for each connection, and I
> know that 10061 is the winsock code for 'connection refused' (which
> makes sense if the server could not spawn a new thread for that
> connection). This is the first such incident in nearly two years of
> operation so I am looking for ideas as to what may have caused this.
> Could it be a resource issue or is it more likely to be a known issue
> with that version of Firebird? Is there anything else I can check? I
> am in the process of confirming, but I highly doubt any changes were
> made to the hardware or software over the weekend.
>
> Thanks
>
> Adam
malicious script is cycle starting sockets
Alan