Karl Fogel wrote:
> Max Bowsher writes:
>> Today, svn.collab.net's hard disc died.
>> The last daily backup was missing the last 19 revisions - fortunately,
>> using mine and Malcolm's personal svnsync'ed copies of the repo, we
>> closed the gap to just the last 3 revisions, and I reconstructed those
>> from commit email diffs.
>> The repositories are now being hosted at red-bean.com until such time as
>> we make longer term plans.
> Thank you, Max and everyone else, for arranging a lifeboat in such
> short order!
> I'm not sure what you had in mind by "longer term plans", but I'm very
> much +1 on continuing to host at svn.collab.net, once that box is back
> in good health. Anyone can have outages (Red Bean has them too), and
> in general we've had excellent uptime from CollabNet.
Oh, definitely. The "longer term plans" was an oblique reference to the
fact that we have no information about how long it will take for a
reborn svn.collab.net to be constructed - and if it turns out to require
purchase of a new hard disc and a full reinstall and reconfigure of the
OS, then, given the quasi-unofficial nature of the box, that could
easily take a week or two, I'd guess.
Mindful of the fact that everyone has outages sometimes, what do people
think about assembling a system where we could freely commit to either
of red-bean.com or svn.collab.net, and the changes would be svnsync-ed
to the other node? Then, if either was unavailable, the other would
continue business as usual.
Received on Sun Mar 18 12:00:08 2007