Date: Wed, 27 Nov 1996 16:17:50 -0600 (CST) From: Joe Greco <jgreco@brasil.moneng.mei.com> To: verdell@novalink.com Cc: freebsd-isp@FreeBSD.ORG Subject: Re: INN Message-ID: <199611272217.QAA19821@brasil.moneng.mei.com> In-Reply-To: <329C6EAB.2D1E@novalink.com> from "Verdell Hicks" at Nov 27, 96 04:39:07 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> Good Afternoon, > Using INN and wanted to ask a few questions. My history file > has > really gotten screwed up and I can't seem to sync things up again. I > brought the entire sytem down ran makeactive and makehistory and still > when I connect to my news feed I still get a lot of can't symlink this > to that. I've been through the faq and have tried the > > ctlinnd renumber '' > makehistory -buv > ctlinnd renumber '' > > and it still did not work. First my expire file reads > > /remember/:5 > > ## Keep for 1-10 days, allow Expires headers to work. > ## for now, expire everything quickly....2 days > *:A:1:3:4 > ## For alt binaries, expire them in 1 day > alt.binaries.*:A:1:2:2 > > ## Some particular groups stay forever. > #dc.dining*:A:never:never:never > #uunet*:A:never:never:never > > Yet it doesn't seem to expire the articles when my hard drive gets down > to 40%. It stops right there. I don't seem to get any error messages > concerning expire. Anyhow, in order to sync things up again I read that > you can simply reformat your news spool and let INN recreate all of the > directories again as the news comes in. What other files will I need to > delete in order to keep this process from blowing up in my face ie. > should I after formatting rerun makeactive and makehistory? should I > actually delete my old history file and let INN start that over also > (will it?). Does anyone have any pointers for me before I reformat and > totally screw things up more? Three options: 1) Assume you have some files that are not in the history for whatever reason. % cd /news % find . -type f -name '[0-9]*' -mtime +10 -print | xargs rm % ctlinnd renumber "" The +10 is the number of days old a file must be to be forcibly removed. Run this from time to time and you will catch articles without the inconvenience of makehistory. 2) Assume your spool is messed up but your history is fine. Start over by removing all articles. % cd /news % find . -type f -name '[0-9]*' -print | xargs rm % ctlinnd renumber "" This way you do not incur directory recreation penalties. Both find based solutions do not handle symlinks. This may be a problem. 3) Assume your spool is messed up but your history is fine. Blow the spool away. % stopnews % su # umount /news # newfs <newfsoptions> </dev/rnewsfsdevice> # mount /news # mkdir -p /news/out.going /news/in.coming/bad /news/in.coming/tmp /news/news/archive # chown -Rf news.news /news; chmod -Rf 775 /news # exit 4) If you wish to blow away your history, % stopnews % cp /dev/null /usr/local/news/history You generally do not want to do this unless you have a compelling reason to believe that you need to (very corrupt history, etc). Okay, that was three options plus a history hint. I can't count. ... JG
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199611272217.QAA19821>