Re: linux-kernel's extremely slow turnaround time

Florian Lohoff (flo@rfc822.org)
Mon, 11 Jan 1999 16:19:24 +0100


On Sun, Jan 10, 1999 at 06:16:39AM -0800, David S. Miller wrote:
> From: Jordan Mendelson <jordy@wserv.com>
>
> Also, what kind of box is vger right now and what kind of
> CPU/memory load is the box at on average?
>
> SS10, 40Mhz SuperSparc, 128MB ram. Running 2.0.33vger + bunch of
> local networking hacks to kill various bogus sites sending ICMP
> redirects to it (some I caught were from Australia, haha very
> funny... ftp.kernel.org gets these from time to time too)
>
> Keep in mind that vger also does loads of CVS activity (read as: very
> disk and memory intensive) and this can cut into the cpu time zmailer
> gets and occaisionally causes thrashing when many people hit vger with
> anoncvs updates etc. (so for the weekend I have anoncvs turned off so
> vger can catch up with all the email backlog).
>
> A little known fact that caused the recent problems is that the
> subscription list to linux-kernel tripled in size around the time when
> Linus started doing 2.2.0 pre-releases and the masses started to
> notice.

What about splitting machines like routing "*.de" to a couple
of alternative *.de relays (Donations ?)

With "sendmail" someone could to this with the mailertable
like .de EMSPT:[alternative.de.relay.de]

Then the machine should send ONE mail to the *.de relay with
loads of rcpt lines possibly even with smtp pipelining
and the load for all recipients in the *.de domain
is off the machine ...

This could be done redundant - sendmail does this

by [alt1.de.relay.de]:[alt2.de.relay.de]:[alt3.de.relay.de]

Now someone should evaluate what specs these machines should
have - To the end of the week ill be able to have a machine up
and running where this could be done (preferred MTA is postfix)

Flo

-- 
Florian Lohoff		flo@rfc822.org		      	+49-5241-470566
Good, Fast, Cheap: Pick any two (you can't have all three).  (RFC 1925)

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/