From: | Michael Paquier <michael(dot)paquier(at)gmail(dot)com> |
---|---|
To: | Andreas Kretschmer <andreas(at)a-kretschmer(dot)de> |
Cc: | Leonardo M(dot) Ramé <l(dot)rame(at)griensu(dot)com>, PostgreSQL mailing lists <pgsql-general(at)postgresql(dot)org> |
Subject: | Re: Streaming Replication archive_command is really needed? |
Date: | 2017-07-21 06:24:50 |
Message-ID: | CAB7nPqRG78R_ZffJdsrw+PRTOwo74x5oWtZ-3FF6WwGattdKAQ@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Fri, Jul 21, 2017 at 8:15 AM, Andreas Kretschmer
<andreas(at)a-kretschmer(dot)de> wrote:
> Am 21.07.2017 um 08:01 schrieb Michael Paquier:
>> "No" is not completely exact and lacks in details. There are two cases
>> where having an archive is helpful:
>> 1) The standby has disconnected from its primary for a time long
>> enough that WAL segments have been rotated by two completed
>> checkpoints. If that happens, when the standby reconnects it would
>> fail, and you would need to take a new base backup.
>
> you can prevent that using replication slots, but i'm pretty sure you
> (Michael) knows that ;-)
> http://paquier.xyz/postgresql-2/postgres-9-4-feature-highlight-replication-phydical-slots/
There is a typo on my URL here. Well that's too late to fix it even if
that's bad style.
>> 2) Backup strategies. Keeping a larger history set of WAL segments is
>> helpful for incremental backups, which is partially the point actually
>> raised upthread about PITR.
>
> Ack, that's right. Using both (streaming and wal-shipping/archiving) will
> make it more robust, and you have (with archiving) the posibility for PITR.
> BUT, you can build a streaming replication without archiving, even you can
> build a continuous backup using only streaming (Barman, streaming only
> mode).
Backup solutions developed by experts on the topic are paths to reliability.
--
Michael
From | Date | Subject | |
---|---|---|---|
Next Message | Michael Paquier | 2017-07-21 06:27:15 | Re: Logging at schema level |
Previous Message | Andreas Kretschmer | 2017-07-21 06:24:04 | Re: Logging at schema level |