Search this site





Apache httpd config to cache the main maven repo

At work, I'm experimenting with Maven as a way to manage java package builds and dependencies.

The maven mirror documentation says pretty explicitly "Do not rsync the entire repo" - it recommends the use of caching proxies. We can do this trivially with apache httpd.

# httpd.conf:
ProxyPass /maven/proxy

# What http path to cache for
CacheEnable disk /maven/proxy
# Where on disk to store the cached data
CacheRoot   /srv/repo/maven/proxy

CacheDirLength 2
CacheDirLevels 3 

# Override default cache expiration and control
CacheDefaultExpire 2419200
CacheMaxExpire 2419200

# Ignore requests to not serve from cache. Maven data never changes.
CacheIgnoreCacheControl On

# Default max file size is 64K. Set to 1GB.
CacheMaxFileSize 1073741824
The above config will take requests to http://yourserver/maven/main/... and proxy them through to the main maven repo and also cache the fetch local to your webserver so future fetches will be local.

You tell maven to use your local repo in ~/m2/settings.xml:

        <name>Local maven mirror</name>
        <-- Replace 'repo.local' with whatever your webserver's name is -->
Now all my maven dependency fetches are going through a local repo and files get cached to disk for future requests.

Since I already had a repo server for local rpm and rubygems, pushing this 5 line httpd config change with puppet was practically a no-op in terms of implementation.