On 3/21/07, David Summers <david@summersoft.fay.ar.us> wrote:
> My understanding was that this *should* allow us to continue to access our
> (local) mirror repository when the master was down, but this turns out to
> not be the case.
It *should* - but something might have crept in.
Does the slave have an identical copy of the authorization and
authentication configs? Ideally, the slave should handle the
authn/authz locally before passing it on to the master.
> Our master server was down for a couple of days and when we tried to
> update on the slave server, we got a variety of error messages, usually
> (but not always) it would take several minutes (like a wait timeout)
> before we got the error message.
>
> I verified that the local (mirror) httpd was running and our local
> svnserver process was up and they were doing fine. Our Daily build could
> check out via the "svn" metthod but when we tried to update via http from
> the local mirror it would not allow us.
>
> Sometimes (but not always) the error would be authorization.
>
> Do I have something mis-configured on my local mirror proxy or is this a
> bug in Subversion?
>
> My first guess is that it is re-directing the authorization requests to
> the master instead of handling them locally, but that is just a gues.
>
> Any thoughts, info pointers, etc., would be greatly appreciated.
I'd like to see a TCP dump of what the slave is trying to pass along
to the master. Feel free to send it to me off-list if it has
passwords or such not suitable for the list... We may need to tweak
some things to ensure that all read-only ops work with the master
being unreachable. But, if I can see a network trace, it'll likely be
very obvious what's going on and how to fix it. -- justin
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Mar 22 19:00:45 2007