[leafnode-list] Re: [beta] select_group dead slow

[Matthias Andree]

Christian Ebert <blacktrash at gmx.net> writes:

> * Matthias Andree on Monday, September 04, 2006 at 01:40:34 +0200:
>> Christian Ebert <blacktrash at gmx.net> writes:
>>> The only difference /in the log/ is that in the meantime 2 more
>>> message have arrived. Still, with leafnode-2, it took slrn
>>> several seconds to enter a group with 35 unread messages, whereas
>>> with leafnode-1 I can enter instantly.
>>> 
>>> Even more confusing is the fact that both servers behave the same
>>> way, if I select *all* messages in a group.
>> 
>> Are you using the same spool directory for either version?
>
> Now also tried and set up a spool from scratch, in case I screwed
> up migration; but the issue persists (with leafnode-2 + slrn, not
> with tin).
>
> I can live with leafnode-1.11.5 and slrn, but the faster fetch of
> leafnode-2 wetted my appetite, as I have to use flaky modem
> connections quite often when I am abroad.

Well, it is unclear to me from your logs what makes leafnode-2 slow in
the slrn case - can you try enabling excessively high debug modes and
try to get a snippet from one GROUP with some of the XOVER commands?
Feel free to bzip2 and mail me off-list (in fact you'll have to if it's
too large, the list will strip it).

-- 
Matthias Andree