On 26/02/19 4:53 PM, Robert Haas wrote:
> On Mon, Feb 25, 2019 at 4:38 PM David Steele <david@pgmasters.net> wrote:
>>> FWIW, if you weren't selling backrest quite so hard everywhere backups
>>> are mentioned, I'd find this thread a lot more convicing.
>> pgBackRest has not used exclusive backups since the new API was
>> introduced in 9.6 so this is not an issue for our users.
>>
>> Over time we have contributed back to Postgres in areas we thought could
>> be improved based on our work on the pgBackRest project: 6ad8ac60,
>> 9fe3c644, 017e4f25, 78874531, 449338cc, 98267ee8, 8694cc96, 920a5e50,
>> c37b3d08, 5fc1670b, b981df4c. This does not include the various backup
>> related patches that we have reviewed.
>>
>> If promoting pgBackRest were our primary concern then it would be in our
>> interest to allow Postgres exclusive backups to stay broken and
>> pg_basebackup to be as primitive as possible.
> Hmm, so what you're saying is that you'd like to disable an API that
> some non-backrest users are relying upon but which no backrest users
> are relying upon. And you don't understand why some non-backrest
> users are opposed to that plan. Is that a correct summary of your
> position?
>
+1 to Robert's Management Summary.
ISTM that the onus should be on the patch submitter to provide additions
to pg_basebackup that make it as painless as possible for those people
*not* using pgBackRest to continue making backups. Breaking this is just
not right. Submitting patches that mean that people *must* use
pgBackRest is also not right IMHO.
Finally, Open Source is about is working together to make the a common
project (in this case Poistgres) better - not forcing us to use
something else you have written (even if it is good).
regards
Mark