Messages per Month
21 messages in com.mysql.lists.clusterRe: Disk-based storage...5.1?| From | Sent On | Attachments |
|---|---|---|
| Jason Williams | 10 Nov 2005 13:49 | |
| Adam Dixon | 10 Nov 2005 14:41 | |
| Simon Garner | 10 Nov 2005 14:42 | |
| Jason Williams | 10 Nov 2005 15:16 | |
| pek...@mysql.com | 10 Nov 2005 15:45 | |
| Fabricio Mota | 10 Nov 2005 16:12 | |
| Stewart Smith | 10 Nov 2005 16:52 | |
| Fabricio Mota | 10 Nov 2005 17:23 | |
| Jason Williams | 10 Nov 2005 17:54 | |
| Simon Garner | 10 Nov 2005 18:01 | |
| pek...@mysql.com | 11 Nov 2005 01:05 | |
| Kenji HIROHAMA | 11 Nov 2005 01:22 | |
| Stewart Smith | 11 Nov 2005 19:46 | |
| Mark Kirkwood | 13 Nov 2005 20:00 | |
| pek...@mysql.com | 14 Nov 2005 01:06 | |
| Mikael Ronström | 14 Nov 2005 15:02 | |
| Jason Williams | 14 Nov 2005 17:55 | |
| Stewart Smith | 14 Nov 2005 19:09 | |
| Jason Williams | 14 Nov 2005 20:11 | |
| Simon Garner | 14 Nov 2005 20:28 | |
| Stewart Smith | 31 Jan 2006 19:16 |
| Subject: | Re: Disk-based storage...5.1?![]() |
|---|---|
| From: | pek...@mysql.com (pek...@mysql.com) |
| Date: | 11/10/2005 03:45:04 PM |
| List: | com.mysql.lists.cluster |
Which is totally unclear as to exactly how it will work. eg is it whole tables only on disk, partitions of tables etc.
Given _attributes_ on disk. No indexed attributes in 5.1. In particular PK must be in memory.
More importantly, will there be a significant performance penalty?
Disk pages are cached (like in any real rdbms). If cache hit ratio is good, performance should be close to memory data.
...Performance was a lot poorer than InnoDB at the time.
InnoDB is within same process (mysqld). NDB is networked.
-- Pekka Nousiainen, Software Engineer MySQL AB, www.mysql.com pek...@mysql.X.com +46 (0) 73 068 4978




