Обсуждение: recovery.conf
It is not obvious to me what is the minimum configuration of the recovery.conf?
The online manuals does not give the exact minumum.
Can anyone help?
Thanks.
--
Med venlig hilsen / Best regards
Poul Kristensen
Poul Kristensen
On 5 September 2016 at 10:50, Poul Kristensen <bcc5226@gmail.com> wrote: > It is not obvious to me what is the minimum configuration of the > recovery.conf? > > The online manuals does not give the exact minumum. > > Can anyone help? The docs are written to describe how to setup various use cases. The minimal config varies depending on what you are trying to achieve. -- Simon Riggs http://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
Poul Kristensen wrote: > It is not obvious to me what is the minimum configuration of the recovery.conf? > > The online manuals does not give the exact minumum. > > Can anyone help? None of the settings in recovery.conf must be specified. One that you will almost always have for recovery is "restore_command", but if all the WAL you need is in pg_xlog, you don't have to specify that either. And you don't need it for streaming replication. Some settings are necessary for point-in-time-recovery, and some for streaming replication, which are two different use cases for the same recovery mechanism. Maybe that's what's confusing you. Yours, Laurenz Albe
Thanks a lot for fast responce.
Do you have any idea of why PostgreSQL don't use cluster like normally used e.g
One common disk for both master and nodes?
Thanks.
Poul
2016-09-05 14:24 GMT+02:00 Albe Laurenz <laurenz.albe@wien.gv.at>:
Poul Kristensen wrote:
> It is not obvious to me what is the minimum configuration of the recovery.conf?
>
> The online manuals does not give the exact minumum.
>
> Can anyone help?
None of the settings in recovery.conf must be specified.
One that you will almost always have for recovery is "restore_command", but if all
the WAL you need is in pg_xlog, you don't have to specify that either.
And you don't need it for streaming replication.
Some settings are necessary for point-in-time-recovery, and some
for streaming replication, which are two different use cases for the same recovery mechanism.
Maybe that's what's confusing you.
Yours,
Laurenz Albe
Med venlig hilsen / Best regards
Poul Kristensen
Linux-OS/Virtualizationexpert and Oracle DBA
Poul Kristensen
Linux-OS/Virtualizationexpert and Oracle DBA
Sorry for asking a lot.
pg_replication_slots.
is this meant to be a slot like a ethernetcard and controller in a physical server.
Thanks a lot.
/Poul
2016-09-06 10:41 GMT+02:00 Poul Kristensen <bcc5226@gmail.com>:
Thanks a lot for fast responce.Do you have any idea of why PostgreSQL don't use cluster like normally used e.gOne common disk for both master and nodes?Thanks.Poul--2016-09-05 14:24 GMT+02:00 Albe Laurenz <laurenz.albe@wien.gv.at>:Poul Kristensen wrote:
> It is not obvious to me what is the minimum configuration of the recovery.conf?
>
> The online manuals does not give the exact minumum.
>
> Can anyone help?
None of the settings in recovery.conf must be specified.
One that you will almost always have for recovery is "restore_command", but if all
the WAL you need is in pg_xlog, you don't have to specify that either.
And you don't need it for streaming replication.
Some settings are necessary for point-in-time-recovery, and some
for streaming replication, which are two different use cases for the same recovery mechanism.
Maybe that's what's confusing you.
Yours,
Laurenz AlbeMed venlig hilsen / Best regards
Poul Kristensen
Linux-OS/Virtualizationexpert and Oracle DBA
Med venlig hilsen / Best regards
Poul Kristensen
Linux-OS/Virtualizationexpert and Oracle DBA
Poul Kristensen
Linux-OS/Virtualizationexpert and Oracle DBA
Poul Kristensen wrote: > Sorry for asking a lot. > > pg_replication_slots. > > is this meant to be a slot like a ethernetcard and controller in a physical server. No, it is a data structure with which the PostgreSQL replication master can keep track how far behind a replication slave is. That way it can avoid discarding WAL that is needed by to replication slave to catch up. Yours, Laurenz Albe
On Tue, Sep 6, 2016 at 10:41:39AM +0200, Poul Kristensen wrote: > Thanks a lot for fast responce. > > Do you have any idea of why PostgreSQL don't use cluster like normally used > e.g > One common disk for both master and nodes? You can read our opinion of that solution here: https://www.postgresql.org/message-id/flat/ncotsa$lhg$1@pye-srv-01.telemetry.co.uk#ncotsa$lhg$1@pye-srv-01.telemetry.co.uk -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + As you are, so once was I. As I am, so you will be. + + Ancient Roman grave inscription +