I’ve been happily hacking at some packages for hashdist, it’s pretty nice, there other build systems out there for dealing with building applications and libraries with different combinations of compilers and numerical libraries. Out of the lot I think hashdist has been the most satisfying to use so far. It’s still missing some bits and pieces to allow users to use different compilers for key components (or everything).
Without explaining too much, it’s basically taking inputs which define a package and then generating an output hash to store the output of the build. This means that as you modify the the environment and rebuild, it will install the updated packages into a different location. Traditionally I would have just overwritten the old versions or install it into a named directory, then setup a module file for it etc… with hashdist there is less messing around. It doesn’t fully solve the problem of having multiple versions of the software available at any one point in time, but it does let you have many versions installed and you can create different inputs profiles to get access to the ‘builds’. This means you can version control your environment, it also means you can roll back to a previous environment if something is broken or wrong.
Share!
See Also
- building a private cloud with saltstack as the cloud controller
- accelerating development and deployments with ansible
- zfs on linux is usable and safe
- building kelvin supercomputer time lapse
- supercomputing 2012 and the arms race to exascale