Wendy Smoak wrote:
On Thu, May 29, 2008 at 12:44 PM, Dennis Lundberg <[EMAIL PROTECTED]> wrote:
Hi
Welcome :)
1. Get the stuff from our file: repos into Archiva
Can we just tar/zip up the entire file repo and copy the whole thing over to
the data directory in Archiva?
The location for each repo is configurable, so you don't necessarily
have to move them. Once you decide where, configure each managed repo
in Archiva.
I think you misunderstood me. We have already configured the Archiva
repos to reside in a location outside of the Archiva installation, as
per the docs.
The file: repo we have is on a completely other machine.
Anyway I copied everything from the file: repos to the Archiva repos and
restarted Archiva. No matter how we tried rescanning and database
updating we were not able to browse the jars from the old file: repo. I
finally gave up and nuked the entire repo database. After a restart all
artifacts were browsable.
2. Have the configuration in the "best" place
I read the previous thread "Choosing which repositories you install to" with
interest. Judging by that, we should remove the <repositories> configuration
from our parent POM and move it into each developers settings.xml, right?
Consider creating a custom Maven distribution to use internally--
replace conf/settings.xml with one that contains your internal
repository config. This also allows you to control (or at least
influence) what version the developers are using, since you have the
"official" one.
If that doesn't suit, then ~/.m2/settings.xml is the next best option.
We've had a discussion about this and came to the conclusion that using
~/.m2/settings.xml will work best for us. The deciding factor was what
we do most, upgrade Maven or add new developers. The former will happen
more often so not having to patch the Maven distributions is a good
thing for us.
4. Use Archiva as a proxy
Set up a mirror configuration in every developers settings.xml like this
<mirror>
<id>company.com</id>
<name>Our Archiva Server</name>
<url>http://archiva.company.com/archiva</url>
<mirrorOf>*</mirrorOf>
</mirror>
So that all downloads, both for company internal artifacts and for open
source artifacts available in the central repo, will be made through
Archiva.
I've never seen a setup that has only *one* managed repository-- for a
single mirror to work you'd have to have all your snapshots, releases,
and third-party artifacts in a single repository. At least, you would
until the 'virtual repository' feature is released, unless you're
willing to try out a snapshot. I think I'd still group releases and
snapshots separately, though I haven't experimented with that feature
yet.
Without any mirror configuration one of our builds, that we used to test
this on, downloaded all artifacts from Archiva instead of the central repo.
--
Dennis Lundberg
