[DISCUSS] test emulation for s390x by IBM

2020-06-05 Thread Sarah-Julia Kriesch


Hi everyone,

I am a Bachelor Thesis Student at IBM and my job is to add s390x emulations 
(hardware emulation for Z systems) to different open source projects. One 
project we want to support is your project Apache Cassandra. So you can support 
s390x as an additional architecture.

I have seen that you are using CircleCI for Continuous Integration. I would be 
open to integrate a minimal emulation there. But I need additional requirements 
from your side:

1) Do you prefer a "real" minimal Linux in your test system for s390x or do you 
prefer any special Linux distribution? We are open to creating a test 
environment with qemu with your favourite distribution.

2) Our idea was to use and integrate the rootfs from a Cassandra Docker image 
directly into the emulation. So we can minimalize the system without any 
additional docker installation and you can see in your test environment whether 
all is ok with the Docker image (via Docker Hub) and whether your development 
is compatible with our hardware.

3) Do you have any other requirements for your hardware tests? You can add them 
here.

I am looking forward to your feedback. :)

Thanks,

Sarah Julia Kriesch


-
To unsubscribe, e-mail: dev-unsubscr...@cassandra.apache.org
For additional commands, e-mail: dev-h...@cassandra.apache.org



Re: [DISCUSS] governance on the Apache Cassandra project

2020-06-05 Thread Joshua McKenzie
Integrated feedback from this thread thus far and responded to comments on
the doc. Should be open to everyone for comment now.


On Thu, Jun 4, 2020 at 7:41 PM Jon Haddad  wrote:

> > With regards to CEPs, I personally don't see any value in voting to start
> one.
>
> Agree with this, and I'd go even further - requiring a vote in order to
> propose an idea runs so counter to the idea of a CEP that it would default
> the purpose of even having them.  The CEP is the _proposal_ for a change
> that gets fleshed out enough so people can understand the idea and _then_
> vote on it, not the other way around.
>
> On Thu, Jun 4, 2020 at 2:51 PM Benedict Elliott Smith  >
> wrote:
>
> > I think the 24 hours point that was raised was pointed to being too short
> > was just for the roll-call; I personally that think for closing down a
> > discussion, 24 hours is acceptable in order to assist progress, since it
> > should only be called when it's clear the discussion has halted or
> > consensus has likely been reached.  If in retrospect it appears that was
> > wrong, we can always cancel the vote.
> >
> > With regards to CEPs, I personally don't see any value in voting to start
> > one.  There's nothing to stop proposers seeking advice, discussion and
> > collaborators beforehand, but voting on it seems premature until there's
> at
> > least some concrete proposal that's had some thought put into it, and an
> > initial round of wider discussion.  There's already a community cost to
> the
> > process, too, and we don't want it to be overly burdensome.
> >
> >
> > On 04/06/2020, 22:39, "Joshua McKenzie"  wrote:
> >
> > On the topic of CEP's, I'd advocate for us trying a couple/few out
> > first
> > and seeing what uncertainties arise as being troublesome and see if
> we
> > can't codify a best practice around them. To date we've had only a
> > couple
> > CEP's actively move and a few in draft pre-move pending more progress
> > on
> > 4.0 so I don't think we have enough signal on how they evolve to know
> > what
> > we might want to address through this doc. Does that make sense?
> >
> > 24 hours to close down lazy consensus does feel pretty quick by
> > default; I
> > think a default 72 hour with flexibility based on the topic (i.e.
> like
> > adding testing to the CEP guideline; super non-controversial) we can
> > just
> > run with things and revert if they're off.
> >
> >
> > Speaking of revert - that's one thing that was a real eye opener for
> me
> > personally philosophically in the past few weeks; git revert exists
> > for a
> > reason and if we all changed our posture to periodic reverts being a
> > healthy thing rather than shameful or contentious, we can all move a
> > lot
> > faster together in trust and revert when mistakes invariably happen.
> > Not
> > that we should start ninja'ing in 40k patches of course, but
> hopefully
> > the
> > point makes sense and resonates in terms of it being a continuum
> we're
> > perhaps quite extreme on culturally as a project.
> >
> > And we all have a sense for when something's more controversial, so
> we
> > have
> > CEP's to lean on. I dunno, makes sense in my head. :)
> >
> > On Thu, Jun 4, 2020 at 4:13 PM Mick Semb Wever 
> wrote:
> >
> > > > A link to the current draft of the governance doc is here:
> > > >
> > >
> > >
> >
> https://docs.google.com/document/d/1wOrJBkgudY2BxEVtubq9IbiFFC3d3efJSj9OIrGcqQ8/edit#
> > > >
> > > > The doc is only 2 pages long; if you're interested in engaging
> in a
> > > > discussion about how we evolve and collaborate as a project,
> > please take
> > > > some time to read through the doc, think through things, and
> > engage on
> > > this
> > > > thread here.
> > >
> > >
> > >
> > > Thanks Benedict and Josh. This is an awesome initiative to put out
> > in the
> > > open and include everyone in.
> > >
> > > My question is around the CEP lifecycle, how one is established and
> > how it
> > > exits (or moves into a real implementation stage). I guess that is
> an
> > > evolving discussion, and also depends on the nature of the
> > individual CEP.
> > > But it raises the questions of when do we apply the vote. For
> > example I can
> > > imagine two votes on a CEP: once to accept an CEP to start in
> > earnest, and
> > > a second time on the finalised CEP that the working group has
> > > finalised. As CEPs
> > > can evolve to quite a different place from their original idea.
> > Maybe we
> > > don't need that entry vote, as the document implies, but I'm not
> > entirely
> > > sure about that: i think some initial exposure and discussion can
> be
> > > valuable to prevent wasted adventures.
> > >
> > > regards,
> > > Mick
> > >
> >
> >
> >
> > -
> > To unsubscribe, e-mail