Hello Dmitry,

thank you for your quick and qualified response.

> This ticket mentions the bug in the database validation code, i.e. it
> can be caused by the GFIX activity only. Is it really your case?

No, GFIX was not involved. The message was produced when client
connection have been closed. After looking at the source code at

http://fisheye1.atlassian.com/browse/firebird/firebird2/src/jrd/cch.cpp?r=1.129.2.16#l3772

and with our explanation it became clear to me that the observed
message probably has just

* the message code and
* the line of code where it is reported

in common with Core-1961.

> BTW, do you use SuperServer or Classic? If the former, then a switch to
> Classic might help to prevent this error from happening.

Yes, we already use the classic server and for more stability we
explicitely configured the per_source parameter in the xinetd
configuration of the classic server (and increased in this way the
value opposed the default value of the centos box).

Greetings

Andreas