Hi,

can anyone help me with the following problem(s):

Running a Firebird SuperServer (1.0.3) on Linux (SuSe 8.2), this
server crashed once and has been restarted by the guardian.

At the time of the crash, there was no activity on the database though
there were open connections from several Java applications (pooled
connections). Everyone knowing these applications was shure, that
there was absolutely no activity (inserts, updates, selects, or other
jobs) on these (pooled) connections at the time of the crash.

The problem was discovered, as one of the applications tried to use
one of the pooled connections and crashed due to the restart of the
firebird server.

Does anyone know a reason for such a crash? The log doesn't say much
about that. Here's an exceprt from the firebird's log:


...
SAM (Client) Tue Apr 27 03:11:05 2004
/opt/interbase/bin/ibguard: bin/ibserver normal shutdown.


SAM (Client) Tue Apr 27 03:11:07 2004
/opt/interbase/bin/ibguard: guardian starting bin/ibserver


SAM (Server) Wed Apr 28 07:41:32 2004
Fatal lock manager error: invalid lock id (0), errno: 22

SAM (Client) Wed Apr 28 07:41:32 2004
/opt/interbase/bin/ibguard: bin/ibserver terminated abnormally (1)


SAM (Client) Wed Apr 28 07:41:32 2004
/opt/interbase/bin/ibguard: guardian starting bin/ibserver

... No more entries


Within the same Log, I found some entries of the form
INET/inet_error: read errno = 104
or
INET/inet_error: send errno = 32
Does anyone know the reason for these errors? Can it be a termination
of any (Java) client without disconnecting the database (what I
believe, but noone can confirm)?

Thanks

Roger