Date: Thu, 26 Aug 2004 13:15:24 -0700 From: Joshua Tinnin <krinklyfig@spymac.com> To: freebsd-questions@freebsd.org Subject: Re: crontab question involving cvsup Message-ID: <200408261315.24610.krinklyfig@spymac.com> In-Reply-To: <200408260228.47205.kstewart@owt.com> References: <200408260007.26659.krinklyfig@spymac.com> <200408260109.12229.krinklyfig@spymac.com> <200408260228.47205.kstewart@owt.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thursday 26 August 2004 02:28 am, kstewart <kstewart@owt.com> wrote: > On Thursday 26 August 2004 01:09 am, Joshua Tinnin wrote: > > On Thursday 26 August 2004 12:42 am, epilogue > > <epilogue@allstream.net> > > > > wrote: > > > On Thu, 26 Aug 2004 00:07:26 -0700 > > > > > > Joshua Tinnin <krinklyfig@spymac.com> wrote: > > > > OK, I have searched the archives, and I can't find that my > > > > question has been answered previously, but please forgive me if > > > > that's incorrect. > > > > > > > > I'm using (or rather trying to use) cron to update my ports > > > > tree daily. I've tried several different combinations without > > > > success, and lately this is what I have in my crontab file: > > > > > > > > /usr/local/bin/cvsup -g -L 2 > > > > /home/krinklyfig/supfiles/ports-supfile && > > > > /usr/local/bin/portindex && /usr/local/sbin/portsdb -u > > > > > > > > It runs as root once a day. What appears to be happening is > > > > that the cvsup is happening, but portindex is not, and because > > > > of the latter portsdb -u doesn't either. The reason I know > > > > cvsup is working is because portindex indicates that the ports > > > > tree has been updated if I run it manually later, but running > > > > portversion before manually running portindex will not indicate > > > > any changes. The cron log doesn't show anything but the > > > > commands being executed. So, my question is: is this type of > > > > command valid, or should each command be separate? Or is it not > > > > working for some other reason? > > > > > > man cron gives: > > > > > > crontab [-u user] file > > > > > > 'file' being the important part, methinks. ;) > > > > I'm not sure what you mean ... If you're wondering, I'm using the > > main crontab file (/etc/crontab), as right now there's no need for > > me to use multiple ones. > > > > > what you might want to do, is simply write a shell script and > > > feed that into your crontab. in case you're not sure how to make > > > a script, it is very simple and google will return many > > > tutorials. > > > > I had considered this, and eventually would like to do so, as I'd > > like to add the output of fastest_cvsup to the server listed in the > > supfile. > > > > > in a nutshell, you put the commands you want into a file, make > > > that file executable (chmod), and away you go. > > > > > > the first line of a shell script has an obligatory format and > > > invokes the shell that will be used. > > > > > > #!/bin/sh << the leading # is required > > > /usr/local/bin/cvsup -L 2 /foo/path/to/your/ports-supfile; > > > # comments are allowed > > > portindex; > > > exit > > > > Is verbosity of -L 2 allowed in a script without output? IOW, > > should that first line be: > > > > /usr/local/bin/cvsup -L 2 /path/to/supfile > /dev/null 2>&1; > > > > Or does it matter if the output has nowhere to go? > > > > > note: you might also prefer to end commands with && rather than ; > > > > > > i'm new to scripting myself so please forgive my feeble > > > explanation. > > > > I'm pretty new to scripting as well. Does ; allow the next line to > > run, even if the previous one didn't, as opposed to && which would > > only allow the next line to run if the previous one was successful? > > (This would be similar to how a one-line command works outside a > > script.) > > > > > about the commands which you are planning to include, why the > > > 'portsdb -u'? is that not doing essentially the same work as > > > 'portindex' ? > > > > > > (http://www.freshports.org/sysutils/portindex/) > > > > No, portindex doesn't update the database. The command portsdb -U > > generates an INDEX, which is what portindex does (although > > portindex does it faster), while portsdb -u generates the INDEX.db > > from the ports INDEX file. However, the database is generated > > automatically if need be when it's looked up, so it's not necessary > > - man portsdb mentions this - but I like to have my ducks in a row, > > so to speak ;) > > > > > anyhow. hope this helps. > > > > Yes, it has me thinking I should probably start testing out a > > script, but it will include a bit more than just cvsup'ping, > > updating the INDEX and database. Thanks. > > This is my cron job. I don't test for completion because of the > multiple commands I execute. It doesn't fail very often. It also > include the usage of portindex and portsdb. > > ruby# m uports > #! /bin/sh > export > PATH=/sbin:/bin:/usr/sbin:/usr/bin:/usr/local/sbin:/usr/local/bin:/us >r/X11R6/bin:/root/bin cd /root/cvsup > cvsup -g -L 2 ports-supfile 2>&1 | tee /var/log/build/ports_cvsup.log > > cd /var/log/build > > # Now convert the log to html` > cvsuplog < ports_cvsup.log > ports-`date "+%Y%m%d-%H%M"`.html > > # Now update the index pages. > cd /usr/ports > # > # make bzip2 backup and save 4 old ones for the days when make index > # is broken > # > rm INDEX.3.bz2 > mv INDEX.2.bz2 INDEX.3.bz2 > mv INDEX.1.bz2 INDEX.2.bz2 > mv INDEX.0.bz2 INDEX.1.bz2 > bzip2 -c INDEX > INDEX.0.bz2 > # > # get new INDEX > #make index 2>&1 | tee /var/log/build/make-index-`date > "+%Y%m%d-%H%M"`.log portindex 2>&1 | tee > /var/log/build/make-index-`date "+%Y%m%d-%H%M"`.log # > #fetch www.freebsd.org/ports/INDEX > #chmod 644 INDEX > portsdb -u Excellent. Thanks so much for sending this, as this is just the sort of thing I need. I'll work on it a bit tonight to customize it for my system. I'm glad you posted this, as in searching the archives I noticed you've posted it before, but it's changed since the last time. - jt
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200408261315.24610.krinklyfig>