Обсуждение: Pg_basebackup does not do aything

Hello,

I have a configuration in master (A) /slave (B) streaming replication.
I bought a third server (C), and I want to use C as my next master server.
In a first time, i would like to sync C as a slave, and in a second time
i would like to promote C to master, and change my current master (A) to
slave (the third server (B) would be out of the equation)

So I tried to use pg_basebackup on C, either synced on B if possible
(cascading replication), or in A if it is needed, but when i use the
command, nothing happened

My command is  :  pg_basebackup -D /data/postgresql/9.3/main -P -v -h
MASTER -U replicator -X s
(this command worked when I used it on my current slave)

In view pg_stat_replication on master, I have this (1st is my streaming
on slave, 2nd is the pg_basebackup I try to use)

  506 |  5828189 | replicator | walreceiver      | 172.25.1.2
|                 |       48022 | 2014-06-25 21:54:34.561107+02 |
streaming | 3E3/597E0000  | 3E3/597A0000   | 3E3/597A0000   |
3E3/5977F290    |             0 | async
  512 |  5828189 | replicator | pg_basebackup    | 172.25.1.3
|                 |       45573 | 2014-06-25 21:54:39.324114+02 |
backup  | 0/0           | 3DD/BDF88000   | 3DD/BDF88000   |
3DD/BDF67C98    |             0 | async


I try to use it on master and slave, but it does the same thing (nothing)
I am wrong in the way I try to use replication ?

Thanks
Thomas

Yes, it did the trick !
Thanks Matheus

Something I don't understand is how my 2nd slave will be synced to master, because master's archive command only sent WAL files to 1st slave.
Does I have to archive on both slaves ? (is it possible) ?
And what I am supposed to use with parameter "restore_command" ?
In this case, my new slave will not start if i dont start it just after the end of the pg_basebackup ?

On Wed, Jun 25, 2014 at 4:58 PM, thomas simon <tsimon@neteven.com> wrote:

My command is  :  pg_basebackup -D /data/postgresql/9.3/main -P -v -h MASTER -U replicator -X s

I'm betting you are just waiting for a "spread" checkpoint. If that is the case, you can issue a CHECKPOINT command on the master and see if your backup start running, or you you can use "--checkpoint=fast" option for pg_basebackup.

--
Matheus de Oliveira
Analista de Banco de Dados
Dextra Sistemas - MPS.Br nível F!
www.dextra.com.br/postgres

From your last message:
> It is recommended to configure it to grab archive files from somewhere (can be from other node) if you have archiving
enabled.So, if the receiver can't keep up with the sender  
> (e.g. when the WAL files are rotated and the sender does not have the files the receiver wants anymore), it can get
thefiles it need from archiving. 
>
According to the postgresql documentation at 25.2.3 Preparing the Master for Standby Servers, Set up continuous
archivingon the primary to an archive directory accessible from the  
standby.

I took this to mean the archive_command = "scp %p postgres@<IP of standby>:/<directory on standby>"

And yes, you can add a ";" at the end of this and specify a second server. That's what I'm doing. Oh, and you have to
setupssh keyrings so the scp will not prompt for a password.  
Anyway, I got the impression that the primary should push and the standby's should not pull WAL segments from this part
ofthe docs. Although, I can see how you could just specify  
the scp in the restore_command and do the pull. I just don't know the safety of doing that.
--
Jay