Path: utzoo!attcan!telly!eci386!clewis From: clewis@eci386.uucp (Chris Lewis) Newsgroups: comp.unix.questions Subject: Re: crontab update Message-ID: <1990Mar23.224743.28401@eci386.uucp> Date: 23 Mar 90 22:47:43 GMT References: <1990Mar1.195750.25818@eng.umd.edu> <1990Mar9.234707.7782@eci386.uucp> <3025@auspex.auspex.com> <1990Mar14.003245.2932@llustig.uucp> Reply-To: clewis@eci386.UUCP (Chris Lewis) Distribution: na Organization: Elegant Communications Inc., Toronto, Canada Lines: 41 In article <1990Mar14.003245.2932@llustig.uucp> david@llustig.UUCP (David Schachter) writes: > Someone wrote: > >>The *only* documented and reliable ways of getting cron to recognize > >>a new crontab is to use the crontab command. > > Or become root, kill cron, change/add/delete crontab stuff, and re-invoke > cron. Violent, but effective. Not, perhaps, documented, but reliable. Quite. Not. What if something's scheduled to run during the time that cron's dead? Especially if you have to consult the vi manual... ;-) Even su root, change/add/delete, kill cron, /etc/cron may miss *something*. There's still a window of vulnerability, albeit short, but why not do it right and not have to worry about it? I had some code using the kill approach. Worked fine until it missed something rather critical.... Got my buns toasted ... Hell, for automating in scripts, the crontab -l > /tmp/$$ ed - /tmp/$$ <