Re: PG Upgrade with hardlinks, when to start/stop master and replicas

Поиск
Список
Период
Сортировка
От Martín Fernández
Тема Re: PG Upgrade with hardlinks, when to start/stop master and replicas
Дата
Msg-id CAH0BA3=o4djA60vy1=eykVzgjcMc7e-jcHbJEiG32qn7eKhzkQ@mail.gmail.com
обсуждение исходный текст
Ответ на Re: PG Upgrade with hardlinks, when to start/stop master and replicas  (Hellmuth Vargas <hivs77@gmail.com>)
Список pgsql-general
Laurenz and Hellmuth, replying to both of you!

Thanks for the quick replies BTW!
Martín


On Mon, Feb 18, 2019 at 5:32 PM Hellmuth Vargas <hivs77@gmail.com> wrote:

Hola Martin

Pues si uno sigue la secuencia de la pagina de ayuda de PostgreSQL 


Usage
(...)
7. Stop both servers
(...)
10. Upgrade Streaming Replication and Log-Shipping standby servers
(...)
12. Start the new server

The new server can now be safely started, and then any rsync'ed standby servers.
 

Pensaría que toca esperar a que terminen todo los rsync de las replicas antes de iniciar la master... Pero tengo incluso una pregunta adicional: si tengo 2 o mas replicas.. se podria  sincronizar primero la master con una replica (rsync), iniciar la master y luego emplear la replica sincronizada para sincronizar el resto de replicas mientras que la máster ya se encuentra en funcionamiento?.

I asked myself that question as well. I'm trying to basically minimize the amount of standbys that are going to be affected by the upgrade. The documentation states that you can do something similar to what you describe, the only thing that you need to make sure is to no start the standby if you are going to use it as a source of rsync to another standby. Would that make sense ?

 

El lun., 18 de feb. de 2019 a la(s) 15:21, Laurenz Albe (laurenz.albe@cybertec.at) escribió:
Martín Fernández wrote:
> After reading the pg_upgrade documentation multiple times, it seems that after running pg_upgrade on the primary instance, we can't start it until we run rsync from the primary to the standby. I'm understanding this from the following section in the pg_upgrade manual page.
>
> You will not be running pg_upgrade on the standby servers, but rather rsync on the
>            primary. Do not start any servers yet.

Immediately following, you can read:

 If you did not use link mode, do not have or do not want to use rsync, or want an easier
 solution, skip the instructions in this section and simply recreate the standby servers
 once pg_upgrade completes and the new primary is running.

So this is not compulsory, it's just an efficient method to quickly get the standby
server updated.

There is nothing wrong with rebuilding the standby later.


I think that by `recreate` here we are talking about pg_basebackup right ? That won't be acceptable because our database would take days to complete. We need to use rsync and leverage the hardlinks.
 
Yours,
Laurenz Albe
--
Cybertec | https://www.cybertec-postgresql.com




--
Cordialmente,

Ing. Hellmuth I. Vargas S.
Esp. Telemática y Negocios por Internet 
Oracle Database 10g Administrator Certified Associate
EnterpriseDB Certified PostgreSQL 9.3 Associate

В списке pgsql-general по дате отправления:

Предыдущее
От: Hellmuth Vargas
Дата:
Сообщение: Re: PG Upgrade with hardlinks, when to start/stop master and replicas
Следующее
От: Adrian Klaver
Дата:
Сообщение: Re: Statement-level trigger results in recursion