Yes, I'm familiar with syslogd. Yes, I'm familiar with newsyslog. On FreeBSD. Yes, I'm familiar with logrotate. On Linux. Yes, I'm aware that logrotate is available as an add-on package from sunfreeware.com. For Solaris. However, having used syslogd+(newsyslog|logrotate), and having used multilog (and Dan Bernstein's other tools), I find multilog to be far superior to syslogd, so I use it whenever possible. In the case of nntpcached, if my feature request is summarily discarded, I intend simply to hack log.h to my liking. On the second point, nntpcached is NOT dying unexpectedly on me. For the entire 323,685 seconds that I've been running nntpcached, its uptime has been rock-solid. For all I know, the nntpcached process will run for the next ten years without a hiccup. However, my level of comfort is increased by knowing that should an unlikely, unforeseen event happen, and nntpcached should go kaput, daemontools will automatically restart it, and I'll receive an e-mail notice the next day rather than a 3am call from a user. Another thing that I like about using daemontools is documentation for junior SysAdmins. I can write something like Usenet news server start svc -u /service/news stop svc -d /service/news bounce svc -t /service/news reconfig svc -h /service/news check status svstat /service/news log file location /var/log/news/current Pretty much all of the add-on services on our different boxes (FreeBSD, Linux, Solaris) are controlled consistently, will start automatically at system boot time (via svscan), and will restart automatically upon unexpected death. To each his own, I guess. --- Adam Montague <AMontague@worldwithoutwire.com> wrote: > > man newsyslog > > there's no need for another program to mess with log files. > and why is nntpcached unexpectedly dying on you? > I'd be looking to solve the problem instead of having another program > restart it for you if it breaks. There's no reason nntpcache should > just stop working for no reason, ours has been up non-stop since it was > installed with no issues. > > Adam Montague > Systems Administrator > amontague@WorldWithoutWire.com > phone : 519.725.0606 > toll-free : 877.725.0606 > fax : 519.725.1430 > > > > -----Original Message----- > From: James Smithe [mailto:nntpcache@yahoo.com] > Sent: October 16, 2001 2:26 PM > To: nntpcache-users@suburbia.net > Subject: NNTPC: Feature request: stdout instead of syslog > > > > > Hello, > > I'm using nntpcache-2.4.0b2 from the FreeBSD ports > collection on FreeBSD 4.4-RELEASE. > > I'm currently running nntpcached with the "-n" > switch (do not detach from tty) so that I can > control nntpcached using Dan Bernstein's > daemontools package: > > http://cr.yp.to/daemontools.html > > which will automatically restart nntpcached > if it dies unexpectedly. > > It would be nice if nntpcached had a switch (or > perhaps a directive in the configuration file) > which would instruct it to send all messages to > stdout instead of to syslogd. The reason for > this is that it would permit log messages to > be captured by Dan Bernstein's multilog program > (part of the daemontools package) > > http://cr.yp.to/daemontools/multilog.html > > multilog lets you specify (amongst many other > things) how large a logfile is allowed to grow > and how many old logfiles should be kept when > multilog auto-rotates the logfile. > > > Thanks! __________________________________________________ Do You Yahoo!? Make a great connection at Yahoo! Personals. http://personals.yahoo.com
