Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 14 Jun 2009 16:09:39 -0400
From:      Adam McDougall <mcdouga9@egr.msu.edu>
To:        Ken Smith <kensmith@cse.Buffalo.EDU>
Cc:        freebsd-hubs <freebsd-hubs@freebsd.org>
Subject:   Re: TLD cvsup mirrors
Message-ID:  <4A355903.3040408@egr.msu.edu>
In-Reply-To: <1244913371.55746.12.camel@neo.cse.buffalo.edu>

index | next in thread | previous in thread | raw e-mail

Ken Smith wrote:
> I'm looking for several new cvsup mirrors for the top level domain
> (cvsup*.freebsd.org).  My preference is to "promote" sites that have
> been working for a while as sites within a country code that feel they
> can take on more traffic.  At this point most of the TLD cvsup servers
> are in the US (and not to put too fine a point on it a depressing number
> are in buffalo.edu because I "temporarily" took over for another server
> that went AWOL and I'm just now getting around to looking for a real
> replacement...) so sites in other countries are definitely encouraged to
> make an offer.  I'd like to diversify it a bit if possible.
>
> Just to give you a feel for what you'd be getting into...  Some of these
> sites currently handle as many as 1000 to 2000 connects/day.  From my
> experiences if you want to support that many connects/day in a
> reasonable fashion it takes a machine that can handle 20 to 30
> concurrent connections.  To do that it helps to have learned a few
> tricks like making sure to turn off atime on the filesystem the
> repository is in, and bumping up kern.maxvnodes to encourage more file
> caching.
>
> Thanks.
>
>   
I'd be happy to take on more cvsupd clients if you want, lately I've 
been getting around 550 working clients per day and it feels like 
nothing (cvsup15.freebsd.org).  I have 20G of ram in the system and 
lately a lot of the cvsup data seems to get cached. Plenty of CPU to go 
around too.

What would you recommend for kern.maxvnodes?  Thanks


home | help

Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4A355903.3040408>