Re: Speed up Switchover
| От | Michael Paquier |
|---|---|
| Тема | Re: Speed up Switchover |
| Дата | |
| Msg-id | CAB7nPqTRwGtNaBnZDVcmYDQsV7CKE7w9P5PJbt9t0kcDUud03Q@mail.gmail.com обсуждение исходный текст |
| Ответ на | Re: Speed up Switchover (Samrat Revagade <revagade.samrat@gmail.com>) |
| Список | pgsql-general |
On Fri, Jul 26, 2013 at 3:00 PM, Samrat Revagade <revagade.samrat@gmail.com> wrote:
-- > secondary without having to rsync the data as it can take up to 10 hours.But I think it has a problem regarding the hint bits which Robert Hass
pg_rewind (https://github.com/vmware/pg_rewind) is what you need.
pointed out.
You can still solve hint bit problem by enabling new checksum feature, But
you have to face consequences such as performance overhead.
You can find the discussion about that on following link:
http://www.postgresql.org/message-id/flat/CA+TgmoY4j+p7JY69ry8GpOSMMdZNYqU6dtiONPrcxaVG+SPByg@mail.gmail.com#CA+TgmoY4j+p7JY69ry8GpOSMMdZNYqU6dtiONPrcxaVG+SPByg@mail.gmail.com
Exactly, using checksums is mandatory or you could finish with corrupted data after rewinding a data directory. Any help in testing or proposing patches for pg_rewind is welcome. It is still in early development.
Thanks,
Michael
В списке pgsql-general по дате отправления: