On Wed, Oct 8, 2014 at 11:38 PM, Heikki Linnakangas <hlinnakan...@vmware.com > wrote:
> On 10/08/2014 04:59 PM, Fujii Masao wrote: > >> On Wed, Oct 8, 2014 at 6:54 PM, Heikki Linnakangas >> <hlinnakan...@vmware.com> wrote: >> >>> Instead of creating any .done files during recovery, we could scan >>> pg_xlog >>> at promotion, and create a .done file for every WAL segment that's >>> present >>> at that point. That would be more robust. And then apply your patch, to >>> recycle old segments during archive recovery, ignoring .done files. >>> >> >> What happens if a user shutdowns the standby, removes recovery.conf and >> starts the server as the master? >> > > Um, that's not a safe thing to do anyway, is it? > That's not safe as it bypasses all the consistency checks of promotion. Now, it is also something that repmgr for example does as far as I recall to do a node "promotion". What if we simply document the problem properly then? The apparition of those phantom WAL files is more scary than a user or a utility that does a promotion with a server restart. Not to mention as well that users as free to add themselves files to pg_xlog. -- Michael
