When a SIGHUP is sent to the nntpcached (v1.0), it sometimes results in a crash due to an incorrect cwd. The reload logic changes directory to the config area to reload configuration files and fails to reset the directory to the cache area. When the reload code calls expire(FALSE), which in turn calls hisPrune, it attempts to open the cache.history file and fails. Nov 5 16:44:12 boris /usr/local/sbin/nntpcached[20509]: nntpcache.c:831:Interrupted system call: caught SIGHUP, restarting... Nov 5 16:44:12 boris /usr/local/sbin/nntpcached[20509]: nntpcache.c:675: cwd now /usr/local/etc/nntpcache Nov 5 16:44:12 boris /usr/local/sbin/nntpcached[20509]: history.c:272:No such file or directory: fatal error: couldn't open/stat cache.history It would also be nice if the openlog call used the basename of the nntpcached program rather than the full invocation path. This is on a Sparc Solaris 2.5 system, BTW. Mark
