[leafnode-list] Follow-up posts are lost?

Matthias Andree ma at dt.e-technik.uni-dortmund.de
Mon Oct 25 10:29:26 CEST 2004


"Mike Vanecek" <leaf_list at mm-vanecek.cc> writes:

>> In 1.10.X, the default is to only show the article locally when the
>> upstream has it, in 2.0.0 with default settings means show it
>> immediately. (2.0.0 can be switched to 1.X behavior by 
>> no_direct_spool=1).
>
> The last time I can find that a post like the following worked was
> around the 20th of October. I did a search using google.groups. Don't
> remember making any changes to leafnode or slrn after that time.

Has the upstream news server changed in that time?
 
> Oct 23 09:35:03 www fetchnews[5537]: news-east.newscene.com: postarticles: 
> trying to post file 5503-1098541941-1 Message-ID <lk6q42-
> vb5.ln1 at news.invalid.99computer>

The upstream servers might discard articles that have illegal domains in
the Message-ID - Google have fixed some issues on their site
recently. I'm not sure what changes exactly they made.

> It looks like it was process correctly on my system should have been 
> appropriately posted on newscene.
>
>> Can you telnet into the upstream server and issue
>> HEAD <message.id.of.your.followup> and check that? If you can't 
>> figure it out, just post the header here and we'll work it out together.
>
> 281 Ok
> group alt.agnosticism
> 211 429 137111 137539 alt.agnosticism
> HEAD lk6q42-vb5.ln1 at news.invalid.99computer
> 430 No such article
>
> I also tried it without the group command.

The angle brackets are mandatory for these commands.
Try: HEAD <lk6q42-vb5.ln1 at news.invalid.99computer>


> As an aside, looking at my logs I notice I started having leafnode timeouts 
> (900s) on a pretty regular basis on Sept 11. Probably not related although I 
> would think that 15 minutes would be sufficient time for just about anything.
>
> Sep 11 09:20:06 www leafnode[27815]: < (ERROR: timeout - no response in 900 s)

If you're posting an article and editing the text takes longer than 900 s,
this is "normal".

> What is frustrating is that other posts are working correctly. I am going to 
> try a reply where I just tell the system to post to all groups rather than 
> setting the follow-up to header. Maybe newscene has a problem with those?

We have too little information to figure that out at this
time. Systematic tests across several _test_ groups (de.test for
instance) can show potential problems.

-- 
Matthias Andree



More information about the leafnode-list mailing list