Hi,
On 02/10/2015 07:40 PM, Christoph Berg wrote:
> Re: Georges Racinet 2015-02-08 <54D7542E.2090503@anybox.fr>
>> Hello list,
>>
>> first message here, so first things first: thank you all for
>> apt.postgresql.org (apt.p.o for short). We've been using and
>> recommending it for a while at Anybox, and that has been tremendously
>> useful.
> Welcome :)
thanks !
>> Now to my question:
>>
>> I'd like to evaluate pg_rewind [1] with the idea to maybe use it on our
>> production replica sets.
>> Are there any plans to package and include it in apt.p.o ? If not, and
>> we turn out willing to use it, I'd probably try and package it anyway,
>> and would look for the proper place to contribute that.
> There's no plans for that, but if you provide a proper package, we can
> include it.
Last week, I couldn't find time to give it a try, but it's still on the
todo-list.
>
>> As of now, pg_rewind is not in contrib/ ; would that be against apt.p.o
>> policy ?
> We have plenty of non-core modules as packages.
>
>> Lastly, the FAQ seems to suggest that apt.p.o is actually downstream of
>> Debian unstable ("rebuilt"), but I'm not sure to get it right,
>> especially for stuff that's not part of PostgreSQL releases.
> That means that the packages on apt.postgresql.org should preferably
> be in unstable as well, but that's not a hard requirement. You'll have
> to argue with me if you don't want that, though ;)
I was just wondering about the workflow (which one should come first)
and the consequences of Debian's freeze, don't worry :-)
>
>> [1] https://github.com/vmware/pg_rewind (has separate branches for 9.3
>> and 9.4)
> Hmm that might be a bit of an issue - the source packages usually
> support several PG versions in parallel, and then there's individual
> binary packages built from that. But it's not like that's an
> unsolvable problem.
I'll make a first try with the 9.4 branch, with explicit PG version pinning.
--
Georges Racinet
Anybox SAS, http://anybox.fr
Bureau: 09 72 39 50 97 / 09 72 39 13 06
Portable: 06 51 32 07 27
GPG: 0x33AB0A35, sur serveurs publics