qpid-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Fraser Adams <fraser.ad...@blueyonder.co.uk>
Subject Re: Review Request 14361: QPID-5197: Remove obsolete --cluster-durable/persistLastNode options
Date Mon, 07 Oct 2013 17:39:50 GMT
On 07/10/13 14:30, Alan Conway wrote:
> Hi Fraser,
>
> It's always a bit of a juggling act isn't it?
Yeah isn't it just :-)
>
> My thinking was that since the feature is now gone on the broker, it 
> is confusing to keep it in the tools. However you have a point about 
> new tools against old brokers.
Yeah Gordon made a good point about the change to clustering being by 
far the bigger change. This comment was as much a reflection/food for 
thought as anything. As I say I got bitten on another change that would 
probably have caused a lot more issues had I not been so familiar with 
QMF, so it's probably a bit fresh in my mind.

>
> Do you think we should keep the old features in the tools and put a 
> deprecation in the release notes, then take them out next cycle? That 
> would mean that we have to test that new tools do actually work with 
> old brokers, which is something we haven't done so far (but probably 
> should)
I guess that I probably wouldn't push the issue for this change (though 
perhaps a note in the documentation for clustering to the effect that 
0.25+ tools can't control the legacy clustering on older brokers won't 
hurt). But on your general point about a "deprecation strategy" that 
sounds a good call. There's probably a line to be drawn somewhere but 
equally helping users avoid the need for big bang upgrades is the 
friendly thing to do.

Thinking about it, for this particular case I *think* that it might be 
possible to use qpid-config, there's the key/value options
     --argument=<NAME=VALUE>
                         Specify a key-value pair to add to queue arguments

So I think someone who actually knew the correct property name could 
still set it, but that doesn't help anyone who may have scripted a 
control mechanism, and if someone has the experience to know to use this 
approach they're probably well sorted anyway :-)

Frase


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


Mime
View raw message