Prepare for 0.8 release

[Weiwei Yang <ab...@gmail.com>]

Hi folks

I'd like to call out for starting the YuniKorn 0.8 release now (the first
release under Apache incubator).
In the past few months, we got tremendous help from community members to
help the growth of the project, now I think it is a good time to cut for
0.8 release. This release will target for K8s users who have needs to run
Big Data workloads on K8s. Features include:

- Hirecharcy resource queues and cross-queue fairness
- Job ordering support at queue level, FAIR & FIFO
- Configurable node-sorting policies: FAIR/BIN-PACKING
- Automated reservations to avoid starving jobs
- Dynamical queue creation and management via user-defined placement rules
- Support K8s predicates functions and volume bindings
- Configuration hot-update
- Stateful recovery

We've been working really hard to roll out all these features, in order to
make YuniKorn standout and ready for production use.
I propose to do code freeze on *March 27*, and the target release date is *Apr
3*.
Let me know your thoughts, thanks!

Weiwei

Re: Prepare for 0.8 release

[Wilfred Spiegelenburg <wi...@apache.org>]

Sounds good for a first release.
I assume the target release date, 3 April, is for the project to have the
artefact(s) ready and request a vote from the IPMC.
We need a vote on the dev list to finish before we can ask the IPMC to
approve the release.

The main point that we need to figure out is how we release with the 4
dependent repositories.
I think we should release 1 source artefact, which contains the code for
all repositories.
Inside the package we need to have the disclaimers etc and a readme on how
to build after unpacking.
This will give us flexibility going forward.
In the future we might need to generate separate binary artefacts from
different repos from one overarching source release.

Wilfred

On Thu, 19 Mar 2020 at 07:55, Weiwei Yang <ab...@gmail.com> wrote:

> Hi folks
>
> I'd like to call out for starting the YuniKorn 0.8 release now (the first
> release under Apache incubator).
> In the past few months, we got tremendous help from community members to
> help the growth of the project, now I think it is a good time to cut for
> 0.8 release. This release will target for K8s users who have needs to run
> Big Data workloads on K8s. Features include:
>
> - Hirecharcy resource queues and cross-queue fairness
> - Job ordering support at queue level, FAIR & FIFO
> - Configurable node-sorting policies: FAIR/BIN-PACKING
> - Automated reservations to avoid starving jobs
> - Dynamical queue creation and management via user-defined placement rules
> - Support K8s predicates functions and volume bindings
> - Configuration hot-update
> - Stateful recovery
>
> We've been working really hard to roll out all these features, in order to
> make YuniKorn standout and ready for production use.
> I propose to do code freeze on *March 27*, and the target release date is *Apr
> 3*.
> Let me know your thoughts, thanks!
>
> Weiwei
>