How to Create a Bioconductor Mirror Site

The Bioconductor package repositories may be mirrored with rsync. If you would like to become a mirror for package and data package repositories, please use the commands below.

Security

Public mirrors

We have recently increased security requirements for the public Bioconductor mirror sites. A mirror is considered “public” if it’s an option in R’s chooseBioCmirror() function and listed on our mirror page. Public site maintainers must support https on their site and use secure rsync when retrieving packages from the master. If you are interested in hosting a publicly available mirror site, please send a suitable public key to valerie.obenchain@roswellpark.org.

Once your key is added to the bioc-rsync account you can rsync from master.bioconductor.org as this user. You can use a config file to manage your ssh keys or include the full path to your key in the -e statement:

rsync -e "ssh -i path/to/your/key" bioc-rsync@master.bioconductor.org ...

Non-public mirrors

To host a non-secure mirror (maybe for internal purposes), invoke the rsync commands without the bioc-rsync user, e.g.,

rsync -avn master.bioconductor.org::release  ## check size
rsync -zrtlv --delete master.bioconductor.org::release /dest/packages/release
rsync -zrtlv --delete master.bioconductor.org::devel/bioc /dest/packages/devel etc ...

BioC release repos

If you want to mirror the current Bioconductor release version (currently 3.4), use the following commands:

Directory structure

Pick a destination directory where files will be mirrored. Let’s say this will be in /dest. This directory should be served by your web server. Under that you’ll need a directory called packages. This directory must be present as it is part of the structure of a Bioconductor repository. Underneath packages should be a directory corresponding to the versions of Bioconductor that you will host. The current release version is 3.4 and the current devel version is 3.5. We recommend you use symlinks called release and devel that always point to the current release and devel versions; this way you will never have to change your rsync commands. But you should change the symlink targets with every Bioconductor release (see the release schedule for exact dates).

The following commands will create the directory structure you’ll need (remember that /dest is just an example of the destination directory you could use; you can put this directory anywhere on your system where there is enough free space).

mkdir -p /dest/packages
mkdir /dest/packages/3.4 # current release
mkdir /dest/packages/3.5 # current devel
ln -s /dest/packages/3.4 /dest/packages/release # change these links
ln -s /dest/packages/3.4 /dest/packages/devel   # every 6 months (with Bioc release)
rsync -e "ssh" -zrtlv --delete bioc-rsync@master.bioconductor.org:release /dest/packages/release

Bioconductor release Software repo

rsync -e "ssh" -zrtlv --delete bioc-rsync@master.bioconductor.org:release/bioc /dest/packages/release /bioc

Bioconductor release Data repos

rsync -e "ssh" -zrtlv --delete bioc-rsync@master.bioconductor.org:release/data /dest/packages/release/data

Bioconductor release Extra repo

rsync -e "ssh" -zrtlv --delete bioc-rsync@master.bioconductor.org:release/extra /dest/packages/release/extra

BioC devel repos

If you want to mirror the Bioconductor devel repos (currently 3.5), please use the following commands:

rsync -e "ssh" -zrtlv --delete bioc-rsync@master.bioconductor.org:devel /dest/packages/devel

Bioconductor release Software repo

rsync -e "ssh" -zrtlv --delete bioc-rsync@master.bioconductor.org:devel/bioc /dest/packages/devel/bioc

Bioconductor devel Data repos

rsync -e "ssh" -zrtlv --delete bioc-rsync@master.bioconductor.org:devel/data /dest/packages/devel/data

Bioconductor devel Extra repo

rsync -e "ssh" -zrtlv --delete bioc-rsync@master.bioconductor.org:devel/extra /dest/packages/devel/extra

Additional information

Make sure the directory above packages is served by a web server.

Bioconductor is big (> 188GB for BioC 3.5). Please check the size of what will be transferred with e.g. rsync -e "ssh" -avn bioc-rsync@master.bioconductor.org:release and make sure you have enough room on your local disk before you start.

It is recommended that package repositories be synced once per day, scheduled with cron.

Begin using your new local repository by making it accessible on your webserver. See the “contriburl” option to install.packages() (utils) for more information.

The Bioconductor master package repositories reside at https://master.bioconductor.org in the Amazon cloud.