At 10:43 AM 1/8/2004 -0800, you wrote:
>X-No-Archive: Yes
>
>Support Auckland <[EMAIL PROTECTED]> wrote:
>
>>>There was one uk.* creation sent message sent out yesterday. For some reason,
>>my copy of DNEWS doesn't even appear to be seeing the uk.* creation message,
>>never mind decrypting it.
>>>
>>>It creates (and deletes) plenty of alt.* groups on a daily basis.
>>>
>
>>That sounds odd, First search the control* log files to see if any
>>record of the message exists.
>>
>>Second I would remove the ,pgp from the uk group
>>rule and see if that fixes it, then the cause will be clear. If that doesn't
>>fix it I'm a bit mystified, next I would try creating a test newgroup message
>>and sending it in locally to see if is stored correctly.
>
>Well, I;ve just started taking the groups via a SUCK feed, but only at 95% to pick up
>missing posts - so they all come via IHAVE really.
>
>But I got the creation message via SUCK today - and it was processed (and rejected
>because I had already created it manually).
>
>So why is it not received via IHAVE?
>
>The creation message is sent out crossposted to the uk .announce group together with
>the new group itself (which obviously doesn't exist on my server at the time). Is
>this somehow causing the problem (it did exist by the time I SUCKed it as I had
>created it manually)? What happens to messages sent to a non-existant group via an
>IHAVE feed?
Control messages are treated differently to normal messages so they
can arrive in non existent groups which is required as newgroup messages
are often sent to the group that is being created even though it doesn't
exist. That's why they are stored in 'control.newgroup'.
My guess is the problem is upstream from you and the upstream
site is failing to send these messages to you. But this is just
a wild guess I admit. Do you have a 'control.newgroup' newsgroup, if not
create one, but I gather you already have that.
I can't think of any easy way to prove the problem is upstream. Since
it's very hard to log a message that doesn't arrive :-).
Possibly the best solution is to continue using the sucking mechanism
to create any missing groups as it will do it in a much simpler manner.
ChrisP.