Re: PITR Base Backup on an idle 8.1 server

Simon Riggs wrote:
> Marco Colombo wrote:
>> my method
>
> ...is dangerous

Ok, but why? Once again, I'm asking: what _exactly_ can go wrong?

> so we don't get loads of new DBAs picking up this idea
but missing the exact point of danger.

I'm one of them. I'm _am_ missing the exact point of danger.

> Making the assumption that its OK to archive WAL files in the pg_xlog
^^^^^^^^^^
> directory exposes you to the risk of having them deleted by the
> archiver, which will invalidate your backup.
^^^^^^^^

I'm sorry I'm really having a hard time following you here... what is
"to archive" and "the archiver"? The archive_command? The tar in the
backup procedure? What do you mean by "deleted"? AFAIK, files in pg_xlog
are first renamed (and only if and after the archive_command returned
true) and later overwritten to. Never deleted. Anyway, how could that
invalidate the backup? It's all about making a self-contained backup.
What happens after that, it's irrelevant.

Hey, I haven't come here proposing a new revolutionary way to perform
backups! I've made pretty clear it was for a not-so-common case. And
anyway, I've just asked what may be wrong with my procedure, since it
seems to fit _my_ needs and it makes _my_ life simpler, and _I_ don't
see any flaw in it. It may be useful to others, _if_ it's correct. If
not, I'd like to know why.

Can you provide a simple failure scenario, please? That would help me
understand what I'm missing...

.TM.