Page 1 of 1

Incomplete files

Posted: Sat Jun 03, 2006 11:36 pm
by caroux
Hello,

Sine june 2 many files that i downloaded are incomplete with this error
"430 no such article (server)", my configuration haven't change. I also tried another
grabber and the files downloaded fine with no error and they were all
complete.

do you have solution for my problem?

Posted: Sun Jun 04, 2006 3:36 am
by alex
it is most probably server.

try to create a duplicate newsgroup with "article numbers" type and set "download bodies" method in properties->articles to "article number".

the if it works the server has its message-id index corrupted (so contact their support). it also means no newsreader in principle is able to use the server to download through nzb file or any search service since the server only succeeds if you download article bodies by its local article numbers so you must to have them somewhere the newsreader database.

the catch (apart forget about nzbs and search service) is also article numbers type is the most greedy RAM type, you can only load about 14M headers per 1GB RAM, with compact binary you can load hundreds of millions headers.

most resource economy can be achieved with properly working servers and for newer downloading header methods (not direct downloading headers) it is a requirement.

in the past (until october, 2005 when they fixed it) newsfeeds had such a problem after one contacted their support they rebuilt message-id index and it was starting working again.

Posted: Sun Jun 04, 2006 3:27 pm
by caroux
I tried your method and i have the same problem as before. The error
message is " 430 no such article (server) ". My server is giganews and
it's working fine with 2 other grabber ( downloading the same files) so
i don't think the server is the problem.
What can the problem be?

thanks

Posted: Sun Jun 04, 2006 3:41 pm
by alex
with giganews i didn't hear about the corrupted message-id index problem in the past so it is unlikely to be it.

but giganews did have the problem with incompletes since other programs who tried to use them as exclusive indexing service header source always had user complaints about incompletes. are you talking about direct header download or downloading through nzb file or search service?

some newsreaders ignore such errors and save attachments (did you recover the files or they were complete right after the outset?). usenet explorer will save such attachments as well (it is set in properties->articles->ignore "no such articles" errors) but it will also indicate error so you'll have a chance to recover the missing parts later if par2 files won't be enough.

it cannot be anything else, the newsreader just gives newsserver article message id and the server either returns article body or returns "no such article" error, you can verify it even using telnet with very few commands.

telnet news.giganews.com 119
authinfo user ...
authinfo pass ...
group ...
article <message-id>

(message id paste from the failed part properties).

nothing can go wrong here.
if you pm me the whereabouts of any such post in question i could verify it with someone who has giganews account.