Re: [leafnode-list] Small posting problem with

[Leopold Toetsch <lt@xxxxxxxxxx>]

<x-flowed>
Hi Matthias
Leopold Toetsch <lt@xxxxxxxxxx> writes:

Sometimes leafnode doesn't post articles on the 1. fetchnews run, but
only on the 2. run. I could not detect a logic behind this, but it may
be related to if I did already read this posted article or not.


BTW I again checked it, it's not related.


Both leafnode versions, 1.9.21 and 2.0b8_ma*, in their current versions
only delete the article from out.going if it is found to be already
available upstream. 


Ah, that's the point. So it's just a timing problem. Article get's 
posted to the server and server may or may not report the article 
already there, depending on housekeeping.

I work around this "problem" by the following snippet in my ip-up:

  ( /usr/sbin/fetchnews -vv >> /var/log/news/fetch.log 2>&1
    /usr/bin/newsq | /usr/bin/grep -q From && \
    /usr/sbin/fetchnews -vv -P >> /var/log/news/fetch.log 2>&1 )&
So this is totally useless.

So it always takes two runs for the posts to clear
from out.going if you have only one server. 


No, I have just one upstream. But, articles are first posted and might 
or might not be ready depending on server load and on articles fetched 
until this group is queried.


And 2. I would prefer the 1.9.x method, that the article first goes to
the ISP and then is fetched from there, so I'm really sure, the article
was posted.


Maybe as an option, I'll put this on my TODO list.


This would be a nice feature (I'm currently modifying articles before 
posting, locally I have only the unmodified article)

Thanks for the detailled answer,
leo
--
leafnode-list@xxxxxxxxxxxxxxxxxxxxxxxxxxxx -- mailing list for leafnode
To unsubscribe, send mail with "unsubscribe" in the subject to the list
</x-flowed>