Ed Swierk wrote:
On 11/20/07, Warren Togami <wtogami@xxxxxxxxxx> wrote:
BUGS TO FIX TO MAKE IT ACTUALLY WORK
- wget --server-response http://PATH/TO/LARGE/FILE
It constantly redirects over and over again rather than waiting.
FILE.tmp never gets larger than 8192 or 16384 bytes.
Hmm, my setup doesn't do that. What upstream server are you using, and
which release of Fedora? I am running it on an FC6 machine, using
http://linux.nssl.noaa.gov .
Fedora 8 with http://download.fedora.redhat.com
- We MUST find a way so the client can begin downloading the file while
the mirror downloads the file. Any other reverse proxy server would
allow this. This might be difficult to implement (perhaps impossible
with mod_python?) but this might be the only sane way to fix the
previous problem.
I don't think this has anything to do with the problem above. This
would be a nice enhancement, though. It does complicate the
implementation, especially to handle byte range requests properly.
It does complicate the implementation, but unfortunately this will be a
completely necessary feature.
- Even if apache:apache owns DocumentRoot, SELinux denies by default.
Need sane solution.
I haven't yet tested on a machine with SELinux enabled.
I've attached an updated version of ApacheMirror.py with a redirection
bug fixed; previously it omitted the http://server part of the URL,
confusing some clients like Anaconda. I don't expect this to fix the
problem you're seeing, though.
--Ed
Did you checkout the initial repo?
Warren
--
fedora-devel-list mailing list
fedora-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-devel-list
