[ale] OT - Postgresql Hot Spare time-lag config question

Jim Kinney jim.kinney at gmail.com
Mon Dec 19 13:03:12 EST 2011


time lag on replication is considered bad and thus avoided like the plague.

However, it can be done by a scripted write-ahead log migration so that
only the logs from 30 minutes ago are processed.

you might want to look at techniques for using WAL and roll back. The idea
is the keep things as close to in-sync as possible but use the WALs to back
out of a wrong commit.

NOTE: by using atomic transactions you will not get corrupt data unless the
logic flow around the insertion is bad. You can get incorrect data. As
PostgreSQL performs extensive type-checking on inserts, the corrupt issue
is moot. Unless all fields are text...

On Mon, Dec 19, 2011 at 11:17 AM, Wolf Halton <wolf.halton at gmail.com> wrote:

> A month ago I was talking to Jim Kinney about PostgreSQL replication, but
> I can't remember if we got down to specifics of how to make PostgreSQL hot
> spare take a 30-minute lag to its write.  The stuff I am seeing all looks
> like the kind of feature most people are looking for is 1:1 realtime synced
> hot-spare.  I want an offset so if the data is corrupt, it might be
> possible to recover on main before the hot-spare data is also corrupt.
>
> I have been noodling around on the PostgreSQL site all morning and nobody
> seems to have that piece there.
>
>
>
> --
> This Apt Has Super Cow Powers - http://sourcefreedom.com
> Advancing Libraries Together - http://LYRASIS.org
>
>
> _______________________________________________
> Ale mailing list
> Ale at ale.org
> http://mail.ale.org/mailman/listinfo/ale
> See JOBS, ANNOUNCE and SCHOOLS lists at
> http://mail.ale.org/mailman/listinfo
>
>


-- 
-- 
James P. Kinney III

As long as the general population is passive, apathetic, diverted to
consumerism or hatred of the vulnerable, then the powerful can do as they
please, and those who survive will be left to contemplate the outcome.
- *2011 Noam Chomsky

http://heretothereideas.blogspot.com/
*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.ale.org/pipermail/ale/attachments/20111219/c1ed2c64/attachment.html 


More information about the Ale mailing list