On Tue, Mar 30, 2010 at 11:26:50AM -0400, Ward Vandewege via RT wrote: > > [yavor - Tue Mar 30 10:39:02 2010]: > > > > Hossam Hossny wrote: > > > This looks like something to drive savannah-hackers attention to > > > > Actually no, the sysadmins are the ones who can fix this. > > > > Some time ago, they added a size restriction in order to reduce > > excessive bandwith resulting from large messages occasionally sent to > > the lists. This affected commits with huge diffs, and we agreed > > subsequently to ignore messages exceeding the default size limit if > > they originate from Savannah: > > > > http://lists.gnu.org/archive/html/savannah-hackers/2008-12/msg00000.html > > > > Apparently this stopped working since at least few months, and I > > intended to reraise the issue again. (I'm adding the RT ticket number > > to the Subject, which would hopefully reopen it again.) > > Yes. We do that based on IP. All the IPs we have allocated to Savannah > in that whitelist, except for the dom0 (colonialone). I didn't realize > that you actually use the dom0 for production services like routing > mail; that's something that's usually not a good idea... > > Anyway; I've added the colonialone IP to the whitelist, so this should > now be fixed. > > I would recommend you review the setup though; ideally the dom0 has *no* > services that touch the outside world, leaving all that - and its > inherent risks - to the various domUs.
Hi, It's just a matter of routing internal IPs - just like for cvs->web webservice that we fixed last month or so. -- Sylvain
