Re: pg_xlogdump and .partial files
От | Michael Paquier |
---|---|
Тема | Re: pg_xlogdump and .partial files |
Дата | |
Msg-id | CAB7nPqRSxc4oAFMaN2uYPOoV67Xvn2Kt95OkKrhYDGb-GRmtLw@mail.gmail.com обсуждение исходный текст |
Ответ на | pg_xlogdump and .partial files (Peter Eisentraut <peter_e@gmx.net>) |
Список | pgsql-hackers |
On Mon, Sep 28, 2015 at 9:47 AM, Peter Eisentraut <peter_e@gmx.net> wrote: > The pg_xlogdump man page states that it cannot read .partial WAL files > and that they need to be renamed. It seems to me with about the same > number of lines we could just make it accept those files, no? FWIW, the discussion has happened here: http://www.postgresql.org/message-id/CAB7nPqS_X5-M+S9yo_xd7OxWq1ETVwV-1TsjUqNOVQYYAztReA@mail.gmail.com In short, I was under the impression that this was not worth it, and I still am under this impression knowing that this is not a common use case. Now, and I imagine that this is actually what is disturbing for the user: we could allow the definition of a .partial file as an end segment. We would need to be careful though in XLogDumpXLogRead when doing the segment switch when the last segment is being read. As pg_xlogdump does not handle timeline jumps this would be useful to allow the user to have a look at the segments of a previous timeline up to the point the promoted standby has switched to, particularly if the .partial file, archived by a promoted standby, and its complete version, archived by old master are present at the same position. Still this seems like a very narrow use-case to me, so opinions are welcome. Regards, -- Michael
В списке pgsql-hackers по дате отправления: