Can we articulate the value of packages over tarballs? In my view, packages are useful for managing dependencies and in-place updates.
If users want to install only parts of BigTop then dependency management between BigTop components is useful. If users always install the entire stack, then dependencies between packages is a moot point. Dependencies on system packages makes installation easier and more foolproof but it's hard to get it right and requires testing on many platforms. For example, I was recently bitten by a (non BigTop) RPM that required Java 7 but Fedora 21 only has Java 8. If this isn't done rigorously, it may be better to avoid specifying dependencies. Packages also offer versioning which helps with updates. But it really depends on whether users do in-place updates or provision fresh installs for new versions. Ultimately, packages DO have value but it requires a lot of effort to reap the benefits. Tarballs may be easier and free up resources though. I see containers / microservices as a provisioning issue. The same packages / tarballs can be provisioned in multiple ways and I think it would be a mistake to lock users into a single approach. Offering provisioning examples (via puppet, ansible, etc.) is useful though IMO. Related question: what are BigTop's goals? Just integration testing? Full blown distro targeted at end users? Packaging for others to build distros on top of? > On Feb 10, 2015, at 7:19 PM, Konstantin Boudnik <[email protected]> wrote: > > I think betting it all on an over-night shift to microservices might be a bit > too risky. Also, I don't see how packages are opposite to the containers? > >> On Tue, Feb 10, 2015 at 11:44AM, jay vyas wrote: >> Hi bigtop. Just a thought... (Thanks @rnowling for providing a seed for >> this discussion this morning) >> >> In BigTop 1x (the experimental branch) we have a chance to do things >> radically different. >> >> - How long are we planning on actually running bigtop on multitenant >> systems Months or years? >> - I we move towards individual processes/microservices/containers, who will >> be consuming our RPM/DEB packages ? >> >> ...If nobody... then bigtop 1x could purely focus on nothing other than >> - integration testing, >> - cutting edge deployment, and >> - puppet recipes >> >> without worrying at all about packaging details. There are tools for DNS >> for containers and so on. >> >> >> -- >> jay vyas >
