Remote polling executes on the central server. It does not require a copy
of the source code on the central server, but it does require read access
to the source code, since the "git ls-remote" command is used to perform
the remote polling.
If you're unwilling to have the source code on your cent
Mark, correct I want no source code on the jenkins server. Currently I use
"Restrict where this project can be run" to force the build to be run from
one specific computer. Your logic would also work, the only issue being
that if the build is started one day on slave A and the next day on slave
B,
I may have misunderstood your comment about not wanting to clone the source
code onto the Jenkins server. If so, my apologies. I assumed you meant
that you don't want the project source code to ever arrive on the central
Jenkins server.
One way to do that might be to constrain the jobs to never
Sure and I see how using a lot of small git repos, as opposed to one large
repo, could be beneficial. But for established code bases this is often not
the case.
On Mon, Dec 16, 2013 at 10:01 PM, David Gayman wrote:
> Nice, I wasn't aware of shallow or reference clones. This could be helpful.
>
Nice, I wasn't aware of shallow or reference clones. This could be helpful.
One more note: The jenkins server is set up as a web server to the outside
world. Our build machines are not exposed as servers. For security reasons
no source code is allowed on the jenkins server; but it would be prefera
Sounds like you want a "shallow clone" rather than a full one. you have thus
option on the git plugin for about a year.
That way you won't have all of the history for the clone and hence less disk
space and network traffic is required.
--
You received this message because you are subscribed to
>
> I still wonder why there are no options out-of-the-box to:
Probably noone needed it badly enough to do it or to pay for it :)
It seems to me that in the Git world, several smaller repos are preferred
instead of a big monolithic one.
Not sure about you, but in our case the big repo was inheri
Greg,
Thank you for responding, I am somewhat new to git and had no idea about
the --reference option. Your solution looks like basically the best way to
do this.
I still wonder why there are no options out-of-the-box to:
1) Set up only one cloned repository per project, and
2) Refuse to check o
Not sure how have the exact same "checkout" multiple times is useful (most
of my jobs modify the workspace so I need to isolate them from each other
-> separate workspace == separate "checkouts").
OTOH, I was also interested to save on disk space when it comes to large
git repos cloned multiple ti
Has anyone considered a modification to the git plugin (or other SCM
plugins), to reduce the total number of checkouts physically required on a
slave machine?
We run 6 mult-configuration projects (debug and release configurations) on
a handful of computers. Each configuration requires one code
10 matches
Mail list logo
