Tony van der Hoff wrote: > From time to time, folloing appears in my syslog. Always in pairs, > always 10 mins apart: > > Feb 16 13:44:41 tony-lx inetd[1870]: tftp/udp: bind: Address already in use > Feb 16 13:54:41 tony-lx inetd[1870]: tftp/udp: bind: Address already in use > > Any suggestions as to what I can do to track down the cause of these > messages?
First, look in /etc/inetd.conf for a line configuring tftp. You might find a line there that looks similar to this: tftp dgram udp wait root /usr/sbin/in.tftpd /usr/sbin/in.tftpd -s /var/lib/tftpboot Then additionally you have probably installed one of the available tftpd packages in Debian and it started a daemon standalone. dpkg -l | grep tftp For example you may have installed tftpd-hpa. In which case it is already running a daemon. $ ps -e | grep tftp 2154 ? 00:00:00 in.tftpd In this case I think you have already been running it standalone so I would continue with it that way. The inetd is reporting the error but it is the "extra" one. The standalone is running first and the inetd one can't bind to the port. The inetd keeps retrying periodically but it fails since the standalone one is continuing just fine. Comment that tftp line out of the /etc/inetd.conf and tell inetd to re-read the file with a SIGHUP. That should keep things in a consistent state and stop the messages to the syslog file. After editing /etc/inetd.conf the inetd needs to be signaled to read the updated file by sending it a SIGHUP signal. There are a variety of methods to do this. Old timers would 'ps -ef | grep inetd' and then 'kill -HUP 32768' the proccess id. Most today would run the init script 'service openbsd-inetd reload'. A few would know about the very spiffy Debian update-inetd command (designed for package postinst scripts) and would use it to do both at the same time. update-inetd --comment-chars '#' --disable tftp Bob
signature.asc
Description: Digital signature
