[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [leafnode-list] leafnode is not pulling in any news
On Wed, Jul 02, 2003 at 11:00:47AM +0200, Matthias Andree wrote:
> Ross Boylan <RossBoylan@xxxxxxxxxxxxxxxxxx> writes:
>
> > I am running leafnode 1.9.41-1 on Debian GNU/linux testing system,
> > with pan 0.13.0-4 and a dialup account to earthlink.
> >
> > It seems not to be pulling in any new articles or headers, though it
> > was a few weeks ago (erratically). I am looking for help in how to
> > diagnose the problem, since I've run into a number of peculiarities.
> >
> > 1) The FAQ says to check permissions on /var/spool/news. This is a
> > symlink to somewhere else. It and the files under are user/group
> > news. Could the symlink be causing trouble?
>
> As long as all directories from the symlink to the root are executable
> for the "news" user (that'd be /var/spool and /var), no.
All are world read and executable.
>
> > 2) The FAQ also says " 1. Your groupinfo file is corrupt. Restart
> > fetchnews with the -f parameter". When I try to do this, I get the
> > error
> > #su news -c "/usr/sbin/fetchnews -f -v -v"
> > leafnode 1.9.41.rel: verbosity level is 2
> > try_lock(timeout=5), fqdn="wheat.dslnorthwest.net"
> > Cannot obtain lock file, aborting.
>
> Some other process, likely another fetchnews or texpire process, is
> already running. Does your system start fetchnews when you've dialled
> in, from /etc/ppp/ip-up or something?
Yes, but it kills it from ip-down. Well, it's supposed to (see
below).
>
> > (same result when did it directly as root). I don't see anything for
> > fetchnews in /var/lock, though there is an entry for leafnode.
>
> Then that's the one you're looking for. If you "cat" it, you should be
> given the PID and the host name of the process holding the lock.
Now I'm confused. leafnode is running all the time. Are you saying
leafnode and fetchnews can't run at the same time?
Ah: when I look at the file, it's the PID of a fetchnews process. And
yet the ppp-down script says
------------------------------------
#!/bin/sh
/etc/news/leafnode/debian-config
# Kill any fetch processes hanging around
if [ "$NETWORK" = "PPP" ]; then
if [ -f /var/lock/news/fetchnews.lck ]; then
/bin/kill -INT $(cat /var/lock/news/fetchnews.lck | head -1)
fi
fi
----------------------------------------------
So I think this may be a problem with the debian package, since
fetchnews.lck seems not to be where the PID lives (the file
fetchnews.lck is never present).
Also slightly odd: the parent process is PID 1, init.
After killing the process, when I dial up I see
------------------------------------------------
Jul 2 16:48:19 wheat fetchnews[1555]: wrote active file with 29068 lines
Jul 2 16:48:19 wheat fetchnews[1555]: child has process ID 7390
Jul 2 16:50:05 wheat fetchnews[7491]: leafnode 1.9.41.rel: verbosity level is 0
Jul 2 16:50:05 wheat fetchnews[7491]: config: debugmode is 1
Jul 2 16:50:05 wheat fetchnews[7491]: config: expire is 20 days
Jul 2 16:50:05 wheat fetchnews[7491]: config: found username for news.earthlink.net
Jul 2 16:50:05 wheat fetchnews[7491]: config: found password for news.earthlink.net
Jul 2 16:50:05 wheat fetchnews[7491]: config: maxage is 30
Jul 2 16:50:05 wheat fetchnews[7491]: try_lock(timeout=5), fqdn="wheat.dslnorthwest.net"
Jul 2 16:50:05 wheat fetchnews[7491]: Last LIST done 3083732 seconds ago: NEWGROUPS
Jul 2 16:50:10 wheat fetchnews[7491]: <200 EarthLink News Service Ready (Typhoon v1.2.4)
Jul 2 16:50:10 wheat fetchnews[7491]: connected to 207.217.77.23:119, reply: 200
Jul 2 16:50:13 wheat fetchnews[7491]: <381 More Authentication Required
Jul 2 16:50:15 wheat fetchnews[7491]: <281 Authentication Accepted
Jul 2 16:50:15 wheat fetchnews[7491]: >MODE READER
Jul 2 16:50:17 wheat fetchnews[7491]: <200 (Typhoon v1.2.4)
Jul 2 16:50:17 wheat fetchnews[7491]: news.earthlink.net: 0 articles posted
Jul 2 16:50:17 wheat fetchnews[7491]: >DATE
Jul 2 16:50:19 wheat fetchnews[7491]: <111 20030702234749
Jul 2 16:50:19 wheat fetchnews[7491]: check_date: news.earthlink.net: server time 1057189669, our time 1057189819
Jul 2 16:50:19 wheat fetchnews[7491]: >NEWGROUPS 030613 165600 GMT
Jul 2 16:50:21 wheat fetchnews[7491]: <231 New newsgroups follow.
Jul 2 16:50:24 wheat fetchnews[7491]: Reading server info from /var/spool/news/leaf.node/news.earthlink.net
Jul 2 16:50:24 wheat fetchnews[7491]: clamping maxage for comp.text.tex to global expire 20
Jul 2 16:50:24 wheat fetchnews[7491]: >GROUP comp.text.tex
Jul 2 16:50:26 wheat fetchnews[7491]: <211 3221 220422 223838 comp.text.tex
Jul 2 16:50:26 wheat fetchnews[7491]: comp.text.tex: considering articles 221981 - 223838
Jul 2 16:50:26 wheat fetchnews[7491]: >XOVER 221981-223838
Jul 2 16:50:28 wheat fetchnews[7491]: <224 Overview Information Follows
Jul 2 16:52:35 wheat fetchnews[7491]: >ARTICLE 222322
Jul 2 16:52:40 wheat fetchnews[7491]: >ARTICLE 222323
Jul 2 16:52:45 wheat fetchnews[7491]: >ARTICLE 222324
Jul 2 16:52:50 wheat fetchnews[7491]: >ARTICLE 222325
Jul 2 16:52:54 wheat fetchnews[7491]: >ARTICLE 222326
Jul 2 16:52:59 wheat fetchnews[7491]: >ARTICLE 222327
Jul 2 16:53:03 wheat fetchnews[7491]: >ARTICLE 222328
Jul 2 16:53:07 wheat fetchnews[7491]: >ARTICLE 222329
Jul 2 16:53:12 wheat fetchnews[7491]: >ARTICLE 222330
Jul 2 16:53:16 wheat fetchnews[7491]: >ARTICLE 222331
etc
-------------------------------
So I think that was the problem.
My nntp retrieval still seems very slow (I think I noticed that over a
year ago; it's been a constant). I'm getting about 0.5 kB/s, whereas
my dialup connection can manage 5 or 6kb/s (e.g., pulling down email).
>
> > I tried this while dialed up (when fetchnews should be running) and when
> > disconnected (when it shouldn't).
>
> fetchnews may not detect your connection is down for an extended amount
> of time.
If the ppp-down were working right, it wouldn't need to detect the
connection was down, would it?
>
> > The groupinfo file is large, and dated a day before my last messages
> > were retrieved--suspicious.
> >
> > 3) I enabled logging, and all I see is this:
> > Jul 1 23:58:03 wheat fetchnews[3018]: leafnode 1.9.41.rel: verbosity level is 0
> > Jul 1 23:58:03 wheat fetchnews[3018]: config: debugmode is 1
> > Jul 1 23:58:03 wheat fetchnews[3018]: config: expire is 20 days
> > Jul 1 23:58:03 wheat fetchnews[3018]: config: found username for news.earthlink.net
> > Jul 1 23:58:03 wheat fetchnews[3018]: config: found password for news.earthlink.net
> > Jul 1 23:58:03 wheat fetchnews[3018]: config: maxage is 30
> > Jul 1 23:58:03 wheat fetchnews[3018]: try_lock(timeout=5), fqdn="wheat.dslnorthwest.net"
> >
> > Is this normal?
>
> There should be more logging after five seconds.
>
OK, doing better now.
> > 4) I notice my interesting.groups subdirectory has no entries dated
> > more recently than June 25, despite my trying to access the groups
> > through pan.
>
> You need to list the "ctime" rather than the "mtime". Try ls -lc or ls
> -ltc on interesting.groups.
ctime looks OK. Is there a reason you're touching the times in this
way? It seems unusual.
> Could you try running fetchnews -nvvv (with debugmode still set to 1)
> and see if you get more logging? It may be a good idea to redirect or
> copy the output to a file so you can grep it later, like:
>
> fetchnews -nvvv 2>&1 | tee fetchnews.log
>
Haven't done that, but it looks as if the problem was the earlier
process wasn't getting cleaned ou.
--
leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx -- mailing list for leafnode
To unsubscribe, send mail with "unsubscribe" in the subject to the list