On Tue, Jan 7, 2014 at 11:00 AM, <k.thiel...@comcast.net> wrote:
> It works for what we need it for, which actually isn't matrix jobs. Our
> setup was more in an interest of saving space and not having to keep
> expanding VM hard disk space as jobs are added. With a much larger network
> location storage space we can have multiple smaller slaves to handle the job
> load. We also use SVN with the emulate fresh checkout option so a lot of
> the workspace is already there for all the slaves.
>
> Again it works for what we need it for, it may not work for everyone else's
> setup necessarily.
I wouldn't expect it to work well for typical matrix jobs unless the
jobs don't actually use the same files (in which case there isn't much
point in trying to share them) or something within the job itself
arbitrates the concurrent access. We do use some network mapping for
read-only access to a common set of tools and libraries, but write
everything locally - on VMs the workspace is on a separate volume that
can be rebuild as needed instead of having to expand a disk in place.
Even if we used network mapped workspace, I'd keep it unique per
slave to avoid any contention.
--
Les Mikesell
lesmikes...@gmail.com
--
You received this message because you are subscribed to the Google Groups
"Jenkins Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to jenkinsci-users+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
