WEIRD issue: an QNX6.3SP2 x86 device that has been running rocksolid for close to half a year now, suddenly no longer allows me to log in to the FTP server. An op I’ve done almost every day with total succes, since the very first day it was powered on.
Doing a ‘ftp 192.168.2.142’ from a windows command prompt, I’m getting the following response:
Connected to 192.168.2.142.
220 192.168.2.142 FTP server ready.
User (192.168.2.142:(none)): anpr
Connection closed by remote host.
The ‘connection closed’ message appears after like a minute.
The syslog on the device reports:
Apr 21 14:43:32 nto ftpd[446497-1]: connection from 192.168.1.10 to 192.168.2.142
Apr 21 14:43:32 nto ftpd[446497-1]: <— 220
Apr 21 14:43:32 nto ftpd[446497-1]: 192.168.2.142 FTP server ready.
Apr 21 14:48:32 nto ftpd[446497-1]: <— 421
Apr 21 14:48:32 nto ftpd[446497-1]: Timeout (300 seconds): closing control connection.
Apr 21 14:48:32 nto ftpd[446497-1]: User unknown timed out after 300 seconds
I’ve been poking around in the box, checking the ftp related config files, access rights, etc. Everything seems ok.
I think you may have answered your own question. You state that nothing has changed on the QNX target, but you don’t claim that nothing has changed outside of QNX.
This, combined with the fact that the syslog shows that ftpd starts, runs for exactly 5 minutes and then claims that the ftp client failed to respond, indicates that the probability favors a change external to QNX (could be the Windows box, the network itself - duplicate IP or MAC address perhaps?).
There is no indication that the QNX ftpd closed the connection in the syslog (I assume that when you ftp on Windows that it fails almost immediately - not after five minutes).
So either there is a hardware failure of some sort on the QNX target (which “changed” it), or there is some configuration/hardware problem outside of the QNX target entirely.