On Mon, 29 Apr 2013 15:39:55 +0200 Man Gro <mangr...@gmail.com> wrote:
> Hi everybody, > I'm running Firebird 2.0 using inetd on a dedicated machine (FreeBSD) > for several years. This morning the database suddenly stopped > responding and then suddenly started responding. > > File /var/log/messages was full of these messages > inetd[856]: /usr/local/bin/fb_inet_server[84685]: exited, status 1 > > I also find another one related to firebird > inetd[856]: /usr/local/bin/fb_inet_server[84569]: exited, signal 14 > > Almost all of them were "exited, status 1". > > I searched the web and couldn't find what that means. Any advice what > "exited, status 1" means? "status 1" most probably means the exit code of the process. Unfortunately, the FB classic server code seems to just use code 1 for everything gone wrong. Googling yields that signal 14 is SIGALRM on FreeBSD (as required by POSIX). The signal(7) manual page says that the default action for a process that received this signal and not handled it is to terminate. The code of FB 2.5 does really uses the alarm(2) syscall and installs a handler to trap SIGALRM. It might turn out that the code misbehaves under certain circumstances so that the signal is delivered at the moment where no handler is installed. So... Try to look to the syslog logs -- is there anything related to FB logged by the inetd? Try passing the "-d" command-line option to the fb_inet_server in the inetd configuration file and see if this improves logging in any way.
