Re: [DISCUSS] CASSANDRA-16840 - Close native transport port before hint transfer during decommission

[Matt Fleming <ma...@codeblueprint.co.uk>]

Thanks Jeff. Since no one has objected I've submitted a patch to this
jira ticket for review.

On Tue, 10 Aug, at 01:27:55PM, Jeff Jirsa wrote:
> The hint behavior aside, stopping native protocol once you begin a decom
> seems like something most people would benefit from, even if they dont
> realize that's what they want to happen.
> 
> 
> 
> On Tue, Aug 10, 2021 at 12:53 PM Matt Fleming <ma...@codeblueprint.co.uk>
> wrote:
> 
> > Hi,
> >
> > With the way that hint transfer currently works during decommission,
> > it's possible to leave hints on the disk of the decommissioning node if
> > those hints are generated after the decommission begins.
> >
> > I'd like to discuss automatically closing the native transport port
> > before hints are transferred to peers during the decommission process to
> > prevent this from happening.
> >
> > While it's possible to manually prevent this problem today if an
> > operator runs "nodetool disablebinary" before starting the decom, I
> > think the default behaviour is surprising enough that doing it
> > automatically makes more sense.
> >
> > Thanks,
> > Matt
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe@cassandra.apache.org
> > For additional commands, e-mail: dev-help@cassandra.apache.org
> >
> >

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