Thanks for clarifications,

I just found out that I need a separate pgsql.conf for each cluster that
you want to backup.  That part was really not clear in my mind even if I
read the plugin documentation many times.
I have read about backups on the postgresql.org site and starting to learn
about it.  Thanks for the links about the PITR. Now I understand the
differences between a cluster and a database.

I had compiled the plugin from the git and the bareos version installed was
14.2.  The git is at 14.4 now, thus my plugin would not work with 14.2.
So I had to search a bit to find the correct version of the source of 14.2,
but I got it now.  And since the bareos installed was from a ubuntu/debian
package, I had to found out the correct configuration to compile bareos
since the plugin need files from bareos to be able to compile.  There is a
"debian" directory in the git but not on the download site (
http://download.bareos.org/bareos/release/14.2/xUbuntu_14.04/), there is a
"rules" files that gives all the config defaults used to compile the .deb.
So I managed to compile with same options for Bareos and contrib-

Right now I'm struggling to correct a permission problem on /etc brought by
the compilation of bareos.  The permissions of /etc were changed and I
cannot go in sudo anymore since my user cannot read /etc/sudoers anymore...
  The machine I'm using is on a cloud, so I have to find out how to boot it
from a live cd or an image of a live cd.

But I'm out of the subject.  I think I have now a good configuration to
backup a test cluster.  I only have to get rid of those annoyances before I
succeed!

Thanks again for your help, it is appreciated,

              Bernard


2015-01-23 9:39 GMT-05:00 Francesco Suavuro <[email protected]>:

> Hello,
>
> On Monday, January 19, 2015 at 9:51:43 PM UTC+1, Bernard Tremblay wrote:
> >
> >     Could you tell me exactly what database will be backup'ed in your
> sample config ?
>
> As I understand plugin will backup a whole PostgreSQL instance and all
> database files. Check what Point-In-Time-Recovery is for and how it works:
> http://www.postgresql.org/docs/9.0/static/continuous-archiving.html
>
> >
> > I understant there is a catalog database catbac on 10.80.7.11 and there
> seems to be a client called "testosteronprd".  Does it backup all
> postgresql database on testosteronprd ?  Can you backup only one database  ?
>
> What do you mean by postgresql database? What is a postgresql database in
> your dictionary?
>
> When you install a PostgreSQL cluster (using for example initdb utility)
> and you start your instance on that cluster you can find it consist of a
> number of databases. You can create any number of new databases in your
> cluster with createdb utility.
>
> I'm asking because I met with different meanings, i.e. some admins use a
> word "database" in the mean of the a cluster.
>
> With PITR you have to backup and restore a whole database cluster, not a
> single database. It is a PostgreSQL feature and not limitation of the
> plugin. Sure, if you have a single database in you cluster you will backup
> and restore a single database.
>
> The most important benefits of the PITR are:
> - you can recover your database to any commited transaction
> - your backup and especially restore could be faster then backup with
> pg_dump utility
>
> >
> > To my understanding, the backup catalog is the postgresql equivalent of
> the catalog for the director and regular files.
>
> Yes.
>
> >
> > Once we have the pgsql.conf done, how do wee add a client with one DB to
> backup ??? Do we add a fileset, a job and a schedule ?
>
> You need two jobs for every PostgreSQL cluster: one for online backup of
> database files with dedicated fileset, second for archive logs with another
> dedicated fileset.
> Check: https://github.com/inteos/pgsql-plugin/wiki/FileSet-configuration
>
> For example: You have a three separate servers every with its own one
> PostgreSQL cluster installed. Call it Pixie, Dixie and Sam.
> You need to define (copy/paste if you are a very lazy men) two filesets
> call it: FileSetPostgreSQLDBSet01, FileSetPostgreSQLArchSet01 as they are
> defined in above wiki page.
> Then as for every Bareos backup Job you need a schedule and client
> definitions, easy. Next you need to define a 6 backup jobs - two (db and
> logs) for every database cluster, i.e:
> Pixie: Pixie-PGSQL-DB-JOB, Pixie-PGSQL-ARCH-JOB
> Dixie: Dixie-PGSQL-DB-JOB, Dixie-PGSQL-ARCH-JOB
> Sam: Sam-PGSQL-DB-JOB, Sam-PGSQL-ARCH-JOB
> You need a three pgsql.conf files, one for every database cluster.
>
> >
> > As you can see I'm quite a noob in bareos and postgresql. But, may be it
> could be an opportunity to clarify the documentation ?
> >
>
> First you should understand a what PITR for PostgreSQL is. Check a link
> above for continous archiving.
>
> --
> You received this message because you are subscribed to a topic in the
> Google Groups "bareos-users" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/bareos-users/6xygN5m9GIs/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to
> [email protected].
> To post to this group, send email to [email protected].
> For more options, visit https://groups.google.com/d/optout.
>



-- 
---------------------------------------------
Bernard Tremblay
[email protected]
(418) 658-1411
---------------------------------------------
Le but de Linux est de gérer vos ressources et faire le travail,
le but des OS propriétaire est de vous vendre d'autres licences...
Portae Tenebrosus Abiciantur, Infinitio!

-- 
You received this message because you are subscribed to the Google Groups 
"bareos-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to