[K12OSN] K12LTSP 3.2 for RHEL, LDAP config wishlist ..

[Andy Rabagliati]

On Mon, 30 Aug 2004, Les Mikesell wrote:

> On Sun, 2004-08-29 at 14:59, Andy Rabagliati wrote:
> 
> > I choose UUCP, because the caller can request delivery of the mail,
> > from a dynamic dialup, even.
> 
> ETRN would work, given tunnels to known addresses.

> 
> > It was developed by people (at the time) who cared about price and
> > service - in the same environment. I do not feel like re-inventing
> > the wheel - it worked fine before.
> 
> Yes, it does do a good job of mail delivery.  It just is hard
> to scale and eventually you'll probably want better access to
> networks behind the connecting machine.  And it is hard to
> beat being able to do 'rsync -essh -z ...' directly to a
> target for efficiency.

Yes, if one had VPN tunnels this would work. But .. I just have UUCP
call across TCP/IP, to pick up traffic. I can wrap it in ssh as well. I
try to avoid protocols that require realtime connections.

I am hoping scaleability can be managed by having the UUCP maps in LDAP.
Maybe I can skip UUCP logins if I could pass authentication to the ssh
wrapper. Then I would need ssh keys in LDAP ..

At this point I am leaning towards a perl script to pull the information
out of the directory and write the UUCP control files.

> If the SMEserver project were in a little better shape I'd
> recommend that for your gateways and local email hubs.
> That's about as easy as it gets for a local admin to
> manage users, groups and local web sites. I have several
> where I've dropped in CIPE as a vpn to a common hub
> but they have been a little bit of trouble to keep
> up to date.

I have my own gateway I put down, available from my website. It handles
DHCP, mail, and uses wwwoffle as an offline web cache, and uses UUCP for
data transfer. However, SMEserver seems fairly mature - I need the
web cache though, and squid does not sit well when offline.

It can even use UUCP to carry web content - so a non-connected school
passes web scoop requests upstream via UUCP, and after scraping the
websites a tarball is passed back via UUCP to be unpacked in the
wwwoffle cache directory.

I also install a snapshot of wikipedia (.org) - 4Gig - so that the
school can have a web-like experience offline.

Cheers,    Andy!

http://wizzy.org.za/