On 9/12/16 3:24 PM, Jeff Janes wrote:
> The man page you quote seems clear to me that setting it to 1, rather
> than leaving it at 0, makes the opportunity for corruption wider, not
> narrower.
Yeah, I actually read it backwards. :/ I don't see how --modify-window
is helpful at all here; you need to use --ignore-times.
> I thought that David's "-1" suggestions was tongue in cheek. But it
> turns out that that actually does work. Of course, it works by forcing
> every file to be copied, which removes the point of using this over
> pg_basebackup, but nonetheless it would preserve the integrity of the data.
AFAIK pg_basebackup blindly copies all data files, while rsync will
transfer only the parts of the files that have actually changed (see
--block-size). If the source and destination are on different servers,
that can mean less data transferred over the network.
--
Jim Nasby, Data Architect, Blue Treble Consulting, Austin TX
Experts in Analytics, Data Architecture and PostgreSQL
Data in Trouble? Get it in Treble! http://BlueTreble.com
855-TREBLE2 (855-873-2532) mobile: 512-569-9461