Hi, Colin,

Thanks for the KIP. Sorry for the late reply. A couple of comments below.

100. It would be useful to think through how to support SASL scram in the
new world. If the only broker listener is SASL scram, we need to bootstrap
the scram credential in each broker before we can start the broker.
Currently, this can be done by adding the scram credential in ZK first and
then start the brokers. It's not clear how this works when the metadata
service is part of the brokers.

101. It would be useful to think through the broker failure detection
logic. From the diagram, it seems that in the new approach, the controller
can fence off a broker but a broker can also fence itself off if it can't
contact the controller. This is a bit different from the current approach
that ZK server solely determines which broker's session expires. There are
tradeoffs between the two approaches. One downslide of the new approach is
that it's not clear if the decision from the controller is always
consistent with that from the broker and whose decision is the source of
truth.

102. For the upgrade process, could we make it clear whether downgrading is
supported or not?

Jun

On Fri, Sep 13, 2019 at 6:12 PM Jason Gustafson <ja...@confluent.io> wrote:

> Late here, but I am a big +1. Great to see this finally happening.
>
> -Jason
>
> On Fri, Sep 13, 2019 at 11:43 AM Colin McCabe <cmcc...@apache.org> wrote:
>
> > Hi all,
> >
> > With 3 binding +1 votes from Gwen Shapira, Ismael Juma, and Bill Bejeck
> > and 5 non-binding +1 votes from Tom Bentley, Ron Dagostino, David Jacot,
> > Dhruvil Shah, Mickael Maison, the vote passes.
> >
> > Thanks, everyone!
> >
> > best,
> > Colin
> >
> >
> > On Fri, Sep 13, 2019, at 01:14, Mickael Maison wrote:
> > > +1 non-binding
> > > Thanks Colin
> > >
> > > On Fri, Sep 13, 2019 at 5:21 AM Dhruvil Shah <dhru...@confluent.io>
> > wrote:
> > > >
> > > > This is exciting! +1 (non-binding)
> > > >
> > > > - Dhruvil
> > > >
> > > > On Thu, Sep 12, 2019 at 12:29 PM Bill Bejeck <bbej...@gmail.com>
> > wrote:
> > > >
> > > > > Thanks for the KIP!
> > > > >
> > > > > +1 (binding)
> > > > >
> > > > > -Bill
> > > > >
> > > > > On Thu, Sep 12, 2019 at 3:26 PM Ismael Juma <ism...@juma.me.uk>
> > wrote:
> > > > >
> > > > > > Thanks for the KIP, +1 (binding).
> > > > > >
> > > > > > Ismael
> > > > > >
> > > > > > On Mon, Sep 9, 2019, 8:28 AM Colin McCabe <cmcc...@apache.org>
> > wrote:
> > > > > >
> > > > > > > Hi all,
> > > > > > >
> > > > > > > I'd like to start the vote for KIP-500: Replace ZooKeeper with
> a
> > > > > > > Self-Managed Metadata Quorum.
> > > > > > >
> > > > > > > The DISCUSS thread from the mailing list is here:
> > > > > > >
> > > > > > >
> > > > > >
> > > > >
> >
> https://lists.apache.org/thread.html/cce5313ebe72bde34bf0da3af5a1723db3ee871667b1fd8edf2ee7ab@%3Cdev.kafka.apache.org%3E
> > > > > > >
> > > > > > > The KIP is here:
> > > > > > >
> > > > > > >
> > > > > >
> > > > >
> >
> https://cwiki.apache.org/confluence/display/KAFKA/KIP-500%3A+Replace+ZooKeeper+with+a+Self-Managed+Metadata+Quorum
> > > > > > >
> > > > > > > regards,
> > > > > > > Colin
> > > > > > >
> > > > > >
> > > > >
> > >
> >
>

Reply via email to