Observation using trn-test76 and INN 2.4.0

[Patrick Vervoorn]

On Wed, 27 Nov 2002, greg andruk wrote:

> There's a thread here that either shows that this overview performance
> quirk has been there for a couple years, or that Barry frightens people,
> or something.
>
> At least this rules out the new overview API as the culprit.
>
> http://www.isc.org/ml-archives/inn-workers/2000/12/msg00059.html

Ah, yes, that was me again. However, at that time, the problem seemed to
be fixed by switching to a new INN release on the server (it went from,
off the top of my head, 2.3.0 to a 2.4.0 prerelease version).

However, it has reappeared now, and only seems to pop up with Really Large
Groups. I.e. a group with about 40000 unread articles works pretty good
most of the time, it's only when going above 100K articles (and big ones
at that probably) that problems started to appear. There weren't too many
of those groups around at that time. I never got to testing the OV_CHUNK
advice then.

Patrick.