Re: Restoring DB using Continuous Archive Backup - Mailing list pgsql-novice

From salman
Subject Re: Restoring DB using Continuous Archive Backup
Date
Msg-id F810C78E-60AD-4764-8F54-B7A5AFFDB453@quietcaresystems.com
Whole thread Raw
In response to Re: Restoring DB using Continuous Archive Backup  (nick <nicktjh@solutionx.com.my>)
List pgsql-novice
On Mar 28, 2008, at 12:15 AM, nick wrote:

> Thanks for the reply, ashish.
>
> Sorry, should’ve been more specific.
>
> By running WAL, I mean reprocessing the WALs up to a certain point
> of time.
>
> For example:
>
> A backup is made at 0000 hours. Something happens. Restore has to be
> made at 2000.
>
> Since the backup is at 0000, 20 hours’ worth of transactions will be
> stored in WAL. To get the data up to the 2000 the WALs have to be
> reprocessed. I realize the only way for this to work is by data file
> duplication, so this mean the service has to be stopped anyway to
> allow the files to be restored.
>
> But there may be a requirement to not take the system offline so I’m
> looking for a method to do the restore and WAL reprocessing without
> stopping then starting the service like described in the manual.
>
> It would be of great help if it’s possible.
>
> Thanks in advance,
>
> Nick
>


You need to put postgres in recovery mode to restore from WAL logs; at
this time, there isn't any way to have any write or read-only access
to the service while a restore is in progress. For 20 hours' worth of
data, depending on your hardware, you're probably looking at 1-3 hours
of restore time before you're caught up.

I don't believe there's a way to have db offline and others available
for usage since that will change the timeline.

HTH,

-salman


pgsql-novice by date:

Previous
From: nick
Date:
Subject: Re: Restoring DB using Continuous Archive Backup
Next
From: ashish
Date:
Subject: Re: Restoring DB using Continuous Archive Backup