<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN"> <HTML> <HEAD> <META NAME="GENERATOR" CONTENT="SGML-Tools 1.0.9"> <TITLE>ISP-Hookup-HOWTO: News</TITLE> <LINK HREF="ISP-Hookup-HOWTO-6.html" REL=next> <LINK HREF="ISP-Hookup-HOWTO-4.html" REL=previous> <LINK HREF="ISP-Hookup-HOWTO.html#toc5" REL=contents> </HEAD> <BODY> <A HREF="ISP-Hookup-HOWTO-6.html">Next</A> <A HREF="ISP-Hookup-HOWTO-4.html">Previous</A> <A HREF="ISP-Hookup-HOWTO.html#toc5">Contents</A> <HR> <H2><A NAME="s5">5. News</A></H2> <H2><A NAME="ss5.1">5.1 How do I set up an online news-reader?</A> </H2> <P> <!-- news-reader --> As long as PPP is active, it will be possible to read news <B>online</B>. There are lots of available programs, two simple alternatives being <!-- rtin --> rtin and <!-- trn --> trn. <P>To start reading news, the only thing required in terms of configuration in most cases is to set NNTPSERVER (usually once and for all in the file <!-- .profile --> .profile): <BLOCKQUOTE><CODE> <PRE> export NNTPSERVER=news.acme.xz </PRE> </CODE></BLOCKQUOTE> <P>To get the <I>From</I>-address correct in postings, some programs <I>may</I> require: <BLOCKQUOTE><CODE> <PRE> export NNTP_INEWS_DOMAIN=acme.xz </PRE> </CODE></BLOCKQUOTE> <!-- NNTP_INEWS_DOMAIN --> <P> <H2><A NAME="ss5.2">5.2 How do I set up an offline news-reader?</A> </H2> <P> <!-- news-reader --> To be able to read news while offline and thus reduce phone bills and give greater flexibility, one must set up a local news-spool of one sort or the other. This requires some configuration, and there will also be a certain amount of disk space involved. After initial setup, things should run more or less by themselves, with only some attention needed from time to time. <P>Two different solutions will be described here. <P> <H2><A NAME="ss5.3">5.3 How do I set up C News?</A> </H2> <P>The solution described here is based on the news-server <B>C News</B> and the NNTP protocol. C News was originally targeted towards another sort of configuration, but is flexible enough to handle our situation too. One might also use the more recent <B>INN</B> news server but it might require a bit more in terms of resources. Either way, be careful <B>not</B> to install both; they don't live together easily. <P>It is crucial that all maintenance of news is done while logged in as user <CODE>news</CODE>, and that all configuration files is placed in <!-- /usr/lib/news --> /usr/lib/news. One way of handling this is, while logged in as <CODE>root</CODE> to write <CODE>su news; cd</CODE>. <P>The most important files in the configuration are: <P> <UL> <LI> <!-- active --> active is an overview over active newsgroups. It is updated as required by the command <!-- addgroup --> addgroup, e.g. <CODE>addgroup comp.os.linux.networking y</CODE>.</LI> <LI> <!-- organization --> organization should simply contain whatever you want in the <I>Organization:</I> header field, in our case:</LI> </UL> <BLOCKQUOTE><CODE> <PRE> Dirk Gently's Holistic Detective Agency </PRE> </CODE></BLOCKQUOTE> <P> <UL> <LI> <!-- mailname --> mailname should in our case be set to <CODE>acme.xz</CODE>.</LI> <LI> <!-- whoami --> whoami is set to the name of your <CODE>site</CODE> in the Path: thread. In a setup as described here, using <!-- newsx --> newsx, this name will never leave the machine, so you can set this to whatever you like as long as you are pretty sure it is unique. In this case <CODE>roderick</CODE>.</LI> <LI> The file <!-- sys --> sys controls fetching and further distribution of news. We will assume the ISP in our case adds <CODE>acme.xz</CODE> to the Path, and that this is the only news source we have. The example given really tells that we will accept everything that arrives, and that we will only post news to <CODE>acme.xz</CODE> that it hasn't seen before, and is originally posted at our own site. In this simplified setup we assume that the all groups will come from a single source. <CODE>/all</CODE> specifies the distribution, and <B>must</B> be included. The letter <CODE>F</CODE> says that (pointers to) outgoing news articles will be collected in a file.</LI> </UL> <BLOCKQUOTE><CODE> <PRE> ME:all/all:: acme/acme.xz:all,!junk/all:FL: </PRE> </CODE></BLOCKQUOTE> <P> <UL> <LI> A subdirectory for the outgoing news must be created, in our case:</LI> </UL> <BLOCKQUOTE><CODE> <PRE> mkdir /var/spool/news/out.going/acme </PRE> </CODE></BLOCKQUOTE> <P> <UL> <LI> <!-- mailpaths --> mailpaths controls posting in moderated groups, although this task may usually be left to the ISP.</LI> </UL> <P>C News needs a certain degree of daily maintenance, but this can be specified once and for all via the command <CODE>crontab -e</CODE> <!-- crontab --> issued as user <CODE>news</CODE>. A suggested setup follows; it can be tuned as required: <BLOCKQUOTE><CODE> <PRE> # maintain incoming and outgoing batches 10,40 * * * * /usr/lib/newsbin/input/newsrun # expire C News, once a day 30 0 * * * /usr/lib/newsbin/expire/doexpire # monitor and report if needed 00 2 * * sat /usr/lib/newsbin/maint/addmissing 40 3 * * * /usr/lib/newsbin/maint/newswatch 50 3 * * * /usr/lib/newsbin/maint/newsdaily </PRE> </CODE></BLOCKQUOTE> <!-- newsrun --> newsrun moves articles in and out (twice every hour), <!-- doexpire --> doexpire will delete articles as they get old (every night at 00:30), and the three last commands does various supervisory and error correcting tasks. <P>One should also ensure that things are cleaned up when starting the machine. As user root, add the following line to /etc/rc.d/rc.local: <BLOCKQUOTE><CODE> <PRE> su news -c /usr/lib/newsbin/maint/newsboot </PRE> </CODE></BLOCKQUOTE> News may be collected via the program <B>NewsX</B>, <!-- NewsX --> picking news from an NNTP-server The program can be found at: <P><CODE> <A HREF="ftp://sunsite.unc.edu/pub/Linux/system/news/transport/newsx-0.9.tar.gz">ftp://sunsite.unc.edu/pub/Linux/system/news/transport/newsx-0.9.tar.gz</A></CODE> Or: <P><CODE> <A HREF="ftp://ftp.sol.no/user/egilk/newsx-0.9.tar.gz">ftp://ftp.sol.no/user/egilk/newsx-0.9.tar.gz</A></CODE> <P>Setting up <CODE>NewsX</CODE> is quite simple. Installation is a classic case of: <BLOCKQUOTE><CODE> <PRE> make su make install exit </PRE> </CODE></BLOCKQUOTE> <P>With the setup outlined here, all you have to do is to create the groups you want to read using the <!-- addgroup --> addgroup command. <P>To fetch articles, user <CODE>news</CODE> issues the following commands (assuming communication via PPP or similar is up): <BLOCKQUOTE><CODE> <PRE> newsrun newsx acme news.acme.xz newsrun </PRE> </CODE></BLOCKQUOTE> <!-- newsrun --> <P>The option <CODE>-d</CODE> gives continuous printout to the screen. Refer to the NewsX documentation for further information. <P>NewsX will also take care of posting of outgoing news. <P>To control disposal of articles as they get old, a file <CODE>explist</CODE> is required. The comments in this example should explain what we want to do: <BLOCKQUOTE><CODE> <PRE> # hold onto history lines 14 days, nobody gets >120 days /expired/ x 14 - /bounds/ x 0-1-120 - # retain these for 2 months comp.sources,comp.os.linux.all x 60 - # noise gets thrown away fast junk,control x 2 - # default: 14 days, no archive all x 14 - </PRE> </CODE></BLOCKQUOTE> <P><B>ALT:</B> In a small news-spool, one will usually not need the newsgroup <CODE>control</CODE>. The traffic is <B>huge</B> compared to the possible usefulness. The main point is that articles will be canceled, and that groups may be created automatically. To ensure that control messages containing <CODE>newgroup</CODE> not shall mess up things for us, a file called <!-- newgroupperm --> newgroupperm specifies what we will allow: <BLOCKQUOTE><CODE> <PRE> comp.os.linux tale@uunet.com yv all any nq </PRE> </CODE></BLOCKQUOTE> In this example, all proper groups under comp.os.linux will be created (y), and the user <CODE>news</CODE> will be notified (v). Everything else will be silently (q) ignored (n). The last line is sufficient if you want to create all groups manually. <P><B>ALT:</B> An alternative to NewsX is suck. <P> <H2><A NAME="ss5.4">5.4 How do I set up Leafnode?</A> </H2> <P>A different solution altogether is to install the integrated package <!-- leafnode --> leafnode. This will handle all tasks required for a personal news spool, and is easy to configure. It is available via: <P><CODE> <A HREF="http://www.troll.no/freebies/leafnode.html">http://www.troll.no/freebies/leafnode.html</A></CODE> <P>As for C News, all news maintenance really should be performed as user <CODE>news</CODE>. <P>The home directory for leafnode is in /usr/lib/leafnode. To install, write: <BLOCKQUOTE><CODE> <PRE> cd /usr/lib/leafnode tar -xzvf leafnode-0.8.tgz cd leafnode-0.8 make su make install </PRE> </CODE></BLOCKQUOTE> <P>Note in the following that the prefix /usr/local/sbin should be replaced with /usr/sbin if you installed leafnode from a package. <P>While still being logged in as <CODE>root</CODE>, change the line that controls NNTP in <!-- /etc/inetd.conf --> /etc/inetd.conf: <BLOCKQUOTE><CODE> <PRE> nntp stream tcp nowait news /usr/sbin/tcpd /usr/local/sbin/leafnode </PRE> </CODE></BLOCKQUOTE> Activate it by: <BLOCKQUOTE><CODE> <PRE> killall -HUP inetd </PRE> </CODE></BLOCKQUOTE> <P>Return to user <CODE>news</CODE> by writing exit. In <!-- /usr/lib/leafnode/config --> /usr/lib/leafnode/config change the line that defines the NNTP server. In our case: <BLOCKQUOTE><CODE> <PRE> server = news.acme.xz </PRE> </CODE></BLOCKQUOTE> <P>Leafnode will look after itself by adding the following command via <CODE>crontab -e</CODE> as user <CODE>news</CODE>: <BLOCKQUOTE><CODE> <PRE> # expire Leafnode, once a day 0 4 * * * /usr/local/sbin/texpire </PRE> </CODE></BLOCKQUOTE> <!-- texpire --> <P>News exchange is also done as user <CODE>news</CODE> by the following command (assuming PPP is up and running): <BLOCKQUOTE><CODE> <PRE> /usr/local/sbin/fetch </PRE> </CODE></BLOCKQUOTE> <!-- fetch --> <P>Users who wants to read news should then use the recipe in <I>How do I set up an online news-reader?</I>, except that they configure for the local machine, i.e: <BLOCKQUOTE><CODE> <PRE> export NNTPSERVER=localhost </PRE> </CODE></BLOCKQUOTE> <P>That should be all there is to it. The first <CODE>fetch</CODE> will transfer a list of available newsgroups. Leafnode will then monitor what groups the users are requesting, and adapt to this the <B>next</B> time it is activated. <P>Note that leafnode does not seems to work in cases where NNTP authorization is required. +.LP <P><B>ALT:</B> An alternative to leafnode is nntpcache, available from: <P><CODE> <A HREF="ftp://ftp.suburbia.net/pub/nntpcache/nntpcache.tgz">ftp://ftp.suburbia.net/pub/nntpcache/nntpcache.tgz</A></CODE> <B>ALT:</B> Another alternative is to use the newsreader <!-- slrn --> slrn together with the <!-- slrn- pull --> slrn-pull package. The newsreader must be compiled with the <CODE>spool</CODE> feature set. <P> <HR> <A HREF="ISP-Hookup-HOWTO-6.html">Next</A> <A HREF="ISP-Hookup-HOWTO-4.html">Previous</A> <A HREF="ISP-Hookup-HOWTO.html#toc5">Contents</A> </BODY> </HTML>