On 08/29/2014 12:48 PM, Jesper Krogh wrote:
> On 20/08/2014, at 23.22, Dimitri Maziuk <dmaziuk AT bmrb.wisc DOT edu> wrote:
>
>> FWIW, I usually pg_dump the schema to a text file and run a script that
>> does '\copy ... to csv' for each table. Then commit them to git or rcs
>> repository right there. Then rsync the repository to a couple of other
>> servers. No bacula needed.
>
> That is not going to give you a backup gauranteed to be consistent.
It will give *me* a consistent backup, but in general you're right: it
won't. (Wrapping it in a single transaction can be problematic.)
The other problem is, like pg_dump it only works up to a point. Once
your .csv's grow to a couple of gigabytes you'll have the same problems
as with one huge pg_dump file. Dep. on the data and update frequency the
deltas can grow into gigabytes even faster and kill your i/o on both vcs
commits and repo sync. Again, *my* csv's aren't expected to get that big
anytime soon, YMMV.
So buyer beware, when it breaks you get to keep the pieces, and all that.
--
Dimitri Maziuk
Programmer/sysadmin
BioMagResBank, UW-Madison -- http://www.bmrb.wisc.edu
signature.asc
Description: OpenPGP digital signature
------------------------------------------------------------------------------
Slashdot TV.
Video for Nerds. Stuff that matters.
http://tv.slashdot.org/ _______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users
|