On Tue, 2002-01-29 at 19:33, Bob Tanner wrote:
> I must be very different then most who subscribe to this list. Turn around time
> of an email is not that big of a deal to me. 45min, 1 hour, 1 week, doesn't
> matter to me.

we just need to make people aware that there may be some lag.

> If I want immediate results, I jump on to irc, there is usually someone always
> there to help. If not #tclug, then #<your distro here>.

Not all people know about irc. Too bad, there's tons of us on the
channel who have no life and enjoy helping people.

> What is the list's expectation in turn around time?

within an hour would be nice in my opinion.

> Instead of throwing more hardware at it, how about a software solution. Carl
> posted why mailman's archiving sucks. How about some of you Python guru's "fix"
> the performance issues?

setting mailman to archive weekly should trim down the size of the mbox
files. I host a higher traffic list than the lug list and don't see
these problems. The key difference is that my list doesn't have _nearly_
as many subscribers. Bob's little busy ass box has several lists with
tons of subscribers and this _may_ be were the latency is comming from
as well.

I think carls suggestion about mysql is a great idea albeit a bit too
much work. IMHO spreading the load of our little list to another box is
a better idea i think and then going further to put the archive store
and webdata for mailman on it's own SCSI disc (even if it's slow it'll
block on I/O alot less long because of the nature of SCSI the userland
processes don't have to fight against the kernel for CPU time because
much of it in done on the processor of the SCSI controller for the I/O)
 
> As an aside, we are not the only group feeling the pinch. Lots of talk about
> performance problems, scalability of mailman on the mailman developer list.

precisely another reason why the geeks (that's us) should pitch in and
help real-time and as a result that Tanner guy. Bob, if you can clear us
a small corner of your space (maybe next to a desk) we'll help out with
the rest.

> I'd be willing to write the middle-ware and back-end stuff, if someone would
> work with me on designing the database to hold it all. Something like an open
> source www.geocrawler.com. 
>
> So, time to put up or shut-up. Open source is about technical people solving
> their own problems (aka scratching their own itch). Seems like we got a big-ass
> mosquito bit. Who wants to help scratch?

agreed. Bring all hardware you'd like to donate to the install fest.
I'll bring a U160 SCSI card and 9Gb disc and an ethernet card. I'll also
take care of setting up an OS, and MTA, and mailman on said box _at_ the
installfest (provided I manage to get transportation to the damn thing)
and I'll volunteer to help manage it. Hell if I had the bandwidth I'd
host the bastard myself. In fact, I'm planning on talking to my bosses
about hosting it on the sistina list server.

How's that for putting up or shutting up.

-- 
Ben Lutgens				http://people.sistina.com/~blutgens/
System Administrator			Sistina Software Inc.	

"If you love someone, set them free. If they come home, set them on
fire."
	- George Carlin
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 232 bytes
Desc: This is a digitally signed message part
Url : http://shadowknight.real-time.com/pipermail/tclug-list/attachments/20020129/da982dad/attachment.pgp