On Wed, 2003-07-30 at 16:24, Aleksander Demko wrote: > On Tue, 2003-07-29 at 21:45, Tom Diehl wrote: > > If you can do ssh you can usually do rsync. Just add a "-e ssh" to the > > rsync command line. rsync is a truely wonderful program for mirroring > > stuff. > > Almost. The DMZed machine itself cannot open ANY sockets to the outside > world - only respond to httpd and ssh requests. So what I did is I > ssh'ed INTO it and used -R port:some.proxy:port so the server tunnels > http requests back down my ssh connection and to a proxy, and then to > the rest of the world. > > After giving up with wget, I decided to just hack together my own > yum/http mirroring tool. It's quite a hack, but seems to work and is > available at: Hey, I just looked through what you were doing. neat idea. Summary of it: grab the header.info file for a repository. traverse the list and build out the .hdrs files and the .rpms corresponding to each and download. nifty. a couple of pointers that will help. look at clientStuff.stripENVRA() that does all the LHS stripping of header.info Then you have: epoch, name, ver, rel, arch and the relative path to the rpm. then you can rebuild from epoch, name, ver, rel, arch the .hdr filename and get it. I would suggest using that function to grab the info and using urlgrabber.py from yum to do the grabbing. you should be able to do the whole script in a lot less lines if you use those few functions. -sv
