[Bucardo-general] how to backup and restore postgres9.6 running on bucardo cluster

[arshi shaikh]

Hi David,

Thanks a lot for your reply. Below is the scenario that we are trying.

As explained earlier we have 3 node Multimaster cluster and bucardo is
running on separate node.
We would like to take the backup from one of the nodes so that in event
when any of the node goes down we can do
recover from the backup.

Assume that backup is taken at 10:00AM from one of the cluster node and
bucardo node. Backup is taken using pg_basebackup utility of Postgres.
Consider node 3 goes down at 10:30AM and we want to recover it so that it
can join the existing cluster and be in sync with
other Master nodes.

We tried following steps to achieve this

1. Performed restore on node3 from the backup taken at 10:00AM.
2.  After restore was completed node3 was able to join the cluster. However
Transaction that happend between 10:00AM to 10:30AM
where missing from this node. Also bucardo did not recognizes that node3 is
not in sync.

We would like to know that if there are any additional steps need to
perform to bring node3 in sync. Also let us know if we are doing anything
wrong.

Any help or suggestion are appreciated.

Thanks in Advance,
Arshi Shaikh


On Thu, Nov 29, 2018 at 11:24 PM David Christensen <david at endpoint.com>
wrote:

> > Hi ,
> >
> >   Can some one please point me to the document or guide me steps by
> which I can take the backup of postgres9.6 running bucado cluster.
> >
> > I have 3 node. Node1 is running bucardo server and other 2 are data node.
> > I want to know exact steps by which I can take backup of bucardo cluster
> and prerequisites that I have to check before triggering the backup.
> > Also steps to restore the backup on failed node and joining back in
> cluster or adding new node in cluster.
> > Any Help is appreciated ..
>
> Hi Arshi,
>
> There shouldn’t really be any specific trick to this; you’d just backup
> the database as normal, making sure to also backup your “bucardo” database
> if that runs on a separate database cluster.  “pg_dump”,  “pg_backrest”,
> etc should just work as normal to save the state of the Bucardo database.
>
> HTH,
>
> David
> --
> David Christensen
> End Point Corporation
> david at endpoint.com
> 785-727-1171
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.endcrypt.com/pipermail/bucardo-general/attachments/20181211/f4480d71/attachment.html>