Draft specification for future X-Trace header
sven at tin.org
Fri Jul 7 22:37:48 UTC 2000
In article <Pine.LNX.4.21.0007071514570.17326-100000 at pcmy.sei.cmu.edu> you wrote:
|> The main reason I want to see this is because I want to be able to
|> squash any supplied Date: header of submitted articles and have nnrpd
|> generate the Date: header itself.
Don't do that. Think of the - nowadays common - case, where your customer
runs an INN installation on his home system, using suck and rpost for
pulling and reposting articles. This customer get his newsfeeds setup wrong
and reposts every article he fetched back to your server. Since he has setup
a cron job for doing this, he doesn't even know, that he reoffers every
article. Usually this is no problem, since the articles' Message-IDs are
already present in your history file, so innd will refuse the articles
coming from the customer (via nnrpd).
No this customers decides to switch his PC off while going into a one month
summer holiday. On the last night before he leaves some hundred articles get
sucked to his system and references are written to the outbound spool file,
waiting for rpost to handle them. Since the PC is switched of, this happens
immediatly after the customer comes back from his holidays a month later.
What has happened in the meantime on your server? The articles got expired
after, lets say 20 days. Your /remember/ is set to 10 days, so this is
irrelevant for these articles. So there is no reference to the articles in
the history anymore.
Usually your innd would reject the one month old articles coming from the
customer as being to old (that's why artcutoff <= /remember/), but if you
rewrite Date headers, all the articles get accepted.
What happens now? The articles spread all over the net, having old
Message-IDs and a new Date header. Some sites keep articles for a long time,
these will reject the articles. Other will accept them because the
Nessage-IDs are not in their history. Thousands of customers of these sites
will have to download old articles again and will be confused about old
threads appearing again.
Even cancelling these duplicate articles won't work, since you would erase
the real articles (with the original Date header) on sites which still have
Can you follow me? :)
More information about the inn-workers