High-availability inn?

Russ Allbery rra at stanford.edu
Wed Sep 14 22:48:02 UTC 2005


Lars Magne Ingebrigtsen <larsi at gnus.org> writes:
> Russ Allbery <rra at stanford.edu> writes:

>> Note that the newgroup has to be received and processed before any
>> messages in that group are processed or things won't work the way that
>> you want.

> Hm...  if I rejuggle the feed file created by just taking all the tokens
> from the history file to have all the control messages first, then I
> should probably get the right result?  (Assuming that I haven't expired
> the newgroup messages, that is.)

Yeah, or you can just sync up the active file first and then it doesn't
really matter.  (It only matters for going forward.)

> Hm, I see that I don't have the control.{newgroup,rmgroup} groups, but
> if I create those two groups, and issue newgroup messages for all the
> groups that I have now, and then have the tokens for these two groups
> first in the feed, then that might probably be the easiest method.  Sort
> of.

I usually just copy over the active file from the current server, run a
script that resets all the numbers to 0000000001 0000000000, and then do
the feed.  (Although these days, just modifying the active file doesn't
create all the groups in overview properly; tdx-util -F may be needed as
well.  I haven't poked at that stuff in a long time.)

Alternately, you could just generate ctlinnd newgroup commands from the
active file and do it that way; it should work, just be a bit slower.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>

    Please send questions to the list rather than mailing me directly.
     <http://www.eyrie.org/~eagle/faqs/questions.html> explains why.


More information about the inn-workers mailing list