Hi, my problem is solved.
My last 3 messages didn't make it to the group. So I'm writing again.
I'd like to thank you all who replied, but Dmitry solved it :)
I didn't know that there was a limit in /etc/xinetd.conf file before
reading Dmirty's message. The original file reads as
-----------
defaults
{
instances = 60
log_type = SYSLOG authpriv
log_on_success = HOST PID
log_on_failure = HOST
cps = 25 30
}
-----------

So, Firebird(CS) connections get limited with this setting. Anyone
who has more clients than 60 need to change this value to a higher
one.

I have no problems with Firebird classic server, as I had with
Firebird super server a few days ago. When I tried Firebird (SS), it
worked fine about 2 hours and but cpu usage got higher and higher.
After a restart of fbserver process, everything was fine about 2-3
hours again. I had a very hard day when this happened.

One more thing about firebird super server is, it's not possible to
monitor the connections(threads). So an administrator cannot
disconnect a specific thread, cannot see a which thread is using how
much cpu process.

When a connection thread goes crazy because of several possible
causes, you need to restart fbserver, disconnecting everyone. That's
not good if you have high number of clients.

I don't know if it's possible, but it can be very good if
administrators can monitor the threads that are connected (cpu and
memory consumption of a thread).

Thank you again and happy new year, happ merry christmas :)