[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [leafnode-list] groups listed twice in groupinfo, and other



Jan Knutar <jknutar@xxxxxx> writes:

> First run of fetchnews -f cleared up all the duplicates, with lots of 
> warnings of the type:
> in-core active file misorder at pos. 1058: "alt.2600.AOL" vs. 
> "alt.2600.aol"

1. you're pulling from some server that has a group aol.2600.AOL.

2. you subscribe to that group

3. groupinfo is lost/damaged/whatever.

4. fetchnews reads "alt.2600.aol" from an upstream server.

5. at some time, the alt.2600.AOL entry is reconstructed from
   interesting.groups.

> Whatever group that is, it seems to be on the upstream servers...
> Now, nothing seems to have been logged through syslog at all, that a 
> good sign?

Can't say.

> After running fetchnews -f, I ran 'fetchnews -v', which fetched list of 
> all groups from all servers again... I thought that's ok, now with the 
> groupinfo fixed maybe it needs one more run to retrieve list and if any 
> server doesn't work I will see it... But I was once again flooded by 
> the error message mentioned above...
>
> Makes me think the source of the problem is somewhere else and still in 
> there... Could it be upper case lower case thing in the error message 
> above?

To verify my above thesis and obtain information I might need to decide
on the fix, could you do the following:

a. show the output of ls -l /var/spool/news/interesting.groups
b. start ngrep -d ppp0 'alt.2600' port 119 | tee ngrep.log
   (replace ppp0 with the name of your external interface)
c. start fetchnews -nf (add -v to your taste)
d. show ngrep.log

If you don't have ngrep, check your distribution, it might have it
somewhere on the install media, other than that, it's available from
http://ngrep.sourceforge.net/, the current version seems to be 1.40.1.

> One upstream server seems to run newscache as well, I seem to recall 
> reading that newscache isn't quite fully functional?

I don't think this particular problem is NewsCache related. The
NewsCache issue was that STAT <message-ID> was always successful up to
including 0.99.17 regardless if the article was actually there or no.

-- 
Matthias Andree

http://sourceforge.net/projects/leafnode/   - leafnode-1 tarballs
http://mandree.home.pages.de/leafnode/beta/ - leafnode-2 tarballs

-- 
leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx -- mailing list for leafnode
To unsubscribe, send mail with "unsubscribe" in the subject to the list