On Tue, Nov 27, 2012 at 2:03 PM, Vojtech Juranek <vjura...@redhat.com> wrote:
>> Automatic fail-over adds a lot of complexity and a lot of new things
>> to go wrong. An easier approach would be a moderately frequent rsync
>> of the jenkins data to a backup server where you would manually change
>
> you can also use e.g. NFS volume mounted to both slaves so that everything is
> up-to-date without any need of synchronization. However, this all these
> approaches aren't real HA - running jobs and all in-memory stuff will be lost
Right. But as a tradeoff, a 'warm' spare won't automatically be
broken by whatever it was that killed the active server (and maybe
deleted its files...). In most situations a failed build isn't a huge
problem. It is really the SCM that needs the best protection.
--
Les Mikesell
lesmikes...@gmail.com
