Yesterday I wrote:

Unfortunately, I do not have time to begin debugging until some time next week. At this point it is not clear at all why the child process that is supposed to take care of refreshing the data dies.

After one more look at the modification date of /home/risops/bin/riswhoisd I suddenly saw the light and realised what might cause the premature death of the child process. Will spare you all details, bottom line is that calling free() multiple times for the same const char* is not a good idea :) I've protected the code from doing that, now riswhoisd again seems to do the right thing upon receiving a HUP signal. The final check will be in ~4 hours, when the daemon should have processed a new set of BGP table dumps. -- Rene