How to backup all Nexus 3 artifacts?

Question

I was wondering how to download all (not one or two specified ones) artifacts from a Nexus 3 repo to local disk. In Nexus 2 it was easy since everything was stored on disk and I would just rsync all the artifacts to my local disk.

But in Nexus 3 all artifacts are stored in the OrientDB and I will have to take an other route. I was thinking about downloading them per http after getting a complete list somehow.

Does anybody has an idea on how to perform such an export?

user_2777965 · Answer

One could download all artifacts from a Nexus3 repository by using n3dr.

user_3170715 · Answer

I modified following groovy scirpt, which can be uploaded using the API: https://gist.github.com/kellyrob99/2d1483828c5de0e41732327ded3ab224

To upload and execute the script I took a look at the examples: https://github.com/sonatype/nexus-book-examples/tree/nexus-3.x

user_6195590 · Answer

You can retrieve all DownloadURLs using the REST-API, then download them all. I did this with a simple Python Script.

user_4444822 · Answer

If you see the version upgrade on the Dockerfile they have modified the location for the base app. You need to point your storage to

/opt/sonatype/sonatype-work

Source: https://github.com/sonatype/docker-nexus3/blob/master/Dockerfile

user_5406499 · Answer

I had to do move all artifacts from one Nexus 3 repository to another one. Once this was accidental, I developed the procedure (Windows):

from Nexus' web interface / Browse / Assets took the JSON with the list of assets
extracted the assets' URLs by collecting the values of their "name" attribute
delete all assets that are .md5, .sha1, maven-metadata.xml
delete all assets that are pom.xml for another atrifact

split the artifact URLS into

path in the repository - it is in the format /<groupId path>/<artifactId>/<versionId>/<file name>
groupId- parse the path, replacing / with .
artifactId - parse the path,
version - parse the path, 
file name - parse the path
packaging - the file name extension

for each such parsed URL call the script (named publish.bat):

@echo off
rem %1 = from repository URL
rem %2 = to repository URL
rem %3 = path
rem %4 = groupId
rem %5 = artifactId
rem %6 = version
rem %7 = file name
rem %8 = packaging
echo.
echo %1%3/%5/%6/%7
echo.

curl --remote-name --create-dirs --output %7 %1%3/%5/%6/%7

call mvn deploy:deploy-file -DgroupId=%4 -DartifactId=%5 -Dversion=%6 -DgeneratePom=true -Dpackaging=%8 -DrepositoryId=admin -Durl=%2 -Dfile=%7

del %7

NOTE: -DrepositoryId=admin is a reference to a server definition in the Maven's settings.xml defining the user and password to publish in the target repository.

Example:

set FROM=source repository URL
set TO=target repository URL

call publish.bat %FROM% %TO% net/xyz/webtools net.xyz.webtools buildServer 03.06.00.01 buildServer-03.06.00.01.war war

How to backup all Nexus 3 artifacts?

5 Answers