| From: | Martijn van Oosterhout <kleptog(at)svana(dot)org> |
|---|---|
| To: | Simon Riggs <simon(at)2ndquadrant(dot)com> |
| Cc: | Gokulakannan Somasundaram <gokul007(at)gmail(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: Index Page Split logging |
| Date: | 2008-01-02 14:01:50 |
| Message-ID: | 20080102140150.GE8566@svana.org |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Wed, Jan 02, 2008 at 01:17:03PM +0000, Simon Riggs wrote:
> That's right; shame I forgot this before I started the thread...
Actually, I think your idea has merit, it's just not as easy as
originally thought. All splits, including multiple-level splits, can be
described as a sequence of "split page X with items {S} going to the
left" followed by an "insert item into page X". The trick is that for
multi-level splits you have to split from the top down. Then each split
becomes a simple loggable operation.
But, I think in the live system we split from the bottom up (the split
and insert is a single operation), so I don't think you can actually
combine the current split algorithm with the logging operation I
suggest above.
Have a nice day,
--
Martijn van Oosterhout <kleptog(at)svana(dot)org> http://svana.org/kleptog/
> Those who make peaceful revolution impossible will make violent revolution inevitable.
> -- John F Kennedy
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Usama Dar | 2008-01-02 15:00:10 | Autovacuum & Table List Ordering |
| Previous Message | Gokulakannan Somasundaram | 2008-01-02 13:48:39 | Re: Index Page Split logging |