Getting Started

Documentation

Glish

Learn More

Programming

Contact Us

Version 1.9 Build 1556

News	FAQ
Search	Home

Mirroring the Repository

The AIPS++ data repository is generated at a central site and is then mirrored by all consortium and user sites. This allows the central site to build and maintain the data, and it allows user sites to easily stay in sync with the master repository. In fact, it is easy for these secondary sites to act as mirrors from which tertiary sites can retrieve the data.

The first step to retrieving the data repository is to install CVSup. While CVSup is optimized for mirroring CVS repositories, it can mirror arbitrary collections of files. CVSup is the mirroring tool used by FreeBSD. CVSup was chosen because it excels at minimizing the information which must be transferred by incrementally updating files using whatever information is known about individual files.

It is easiest to just install CVSup binaries because CVSup is written in Modula-3. CVSup binaries for many architectures can be found at the FreeBSD FTP site. You can choose between versions which include or exclude the GUI. However, the GUI version is recommended since the GUI is nice and can be disabled with a command line flag.

Once you have installed CVSup, you are ready to create a supfile which describes how to retrieve and install the data repository. First log in as aips2mgr and create a directory where the repository should be mirrored, e.g. /aips++/data. It is fine for many AIPS++ installations to share the same data repository. So if you have more than one AIPS++ installation, you will likely want to mirror the repository in a central location and then just create symbolic links from the AIPS++ installations to the repository. If you have only one AIPS++ installation, then calling the directory data and putting it in the root of the AIPS++ installation (at the same level as code) is the best choice.

Next edit a file called supfile in the directory just created. To retrieve the whole repository, which currently isn't so large, you would put the following in the supfile:

   *default host=aips2.nrao.edu compress
   *default base=YOUR_DIRECTORY_HERE
   *default release=all
   *default delete use-rel-suffix
   data

You should replace YOUR_DIRECTORY_HERE in the base assignment with the fully qualified path to the directory you just created to contain the repository.

After you have created the supfile, you are ready to install the repository. Just run:

   yourhost% cvsup supfile

If you don't run this from the newly created directory, you will have to fully qualify the path to the supfile for cvsup. This should run and create a mirror of the data provided on the central AIPS++ server.

If you want to continue to keep your data repository in sync with the repository on the master AIPS++ server, you need to create a cron job entry to run cvsup regularly. This entry might look like:

   00 6 * * 0 /usr/local/bin/cvsup -g /aips++/data/supfile 2>&1 | mail aips2mgr

This would update the data repository once per week on Sunday at 6AM. It is best to synchronize the updates of the data repository with the update of local AIPS++ installation.

If you have multiple AIPS++ installations which will be sharing a single data repository, create a symbolic link called data in the root of each installation (at the same level as code) to the data repository.