Hi,
Well, the 'complex' stuff is only as there for larger or high-traffic
DBs. Besides at 60GB that is a largish DB in itself and you should begin
to try out a few other backup methods nonetheless. That is moreso, if
you are taking entire DB backups everyday, you would save a considerable
lot on (backup) storage.
Anyway, as for pgdump, we have a DB 20x bigger than you mention (1.3TB)
and it takes only half a day to do a pgdump+gzip (both). One thing that
comes to mind, how are you compressing? I hope you are doing this in one
operation (or at least piping pgdump to gzip before writing to disk)?
--
Robins Tharakan
On 11/13/2011 05:08 PM, Phoenix Kiula wrote:
> Hi.
>
> I currently have a cronjob to do a full pgdump of the database every
> day. And then gzip it for saving to my backup drive.
>
> However, my db is now 60GB in size, so this daily operation is making
> less and less sense. (Some of you may think this is foolish to begin
> with).
>
> Question: what can I do to rsync only the new additions in every table
> starting 00:00:01 until 23:59:59 for each day?
>
> Searching google leads to complex things like "incremental WAL" and
> whatnot, or talks of stuff like pgcluster. I'm hoping there's a more
> straightforward core solution without additional software or PHD
> degrees.
>
> Many thanks for any ideas!
> PK
>
