From: | Marty Scholes <marty(at)outputservices(dot)com> |
---|---|
To: | pgsql-hackers(at)postgresql(dot)org |
Cc: | Marty Scholes <marty(at)outputservices(dot)com> |
Subject: | Indices pinned in pageable RAM? |
Date: | 2004-06-02 19:38:47 |
Message-ID: | 40BE2CC7.9050105@outputservices.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Has anyone seriously looked at how it would impact things to give the
DBA the option of storing certain indices in RAM instead of on disk?
Queries (both select and insert/update) against heavily indexed tables
do most of the reads and writes to the index trees and relatively little
reading and writing to the actual table data.
The index is completely redundant (read: rebuildable during restart) and
its rebuild "should" be very parallelizable. Without the WAL and disk
synchronization munching, even a single-threaded rebuild during restart
should be fairly quick.
This would allow the index to stay in RAM. giving it a higher priority
than all other disk objects in the OS cache. Putting it in pageable RAM
would allow the OS to reclaim core if it absolutely had to.
Am I way off here?
From | Date | Subject | |
---|---|---|---|
Next Message | Jonathan Gardner | 2004-06-02 20:28:46 | Re: constraint upon view |
Previous Message | Andrew Dunstan | 2004-06-02 18:54:48 | Re: Default Locale in initdb |