6 messages in com.mysql.lists.clusterRe: Alter table clarification| From | Sent On | Attachments |
|---|---|---|
| Nathan McCall | 05 May 2005 12:12 | |
| Jonas Oreland | 06 May 2005 00:51 | |
| Kris Buytaert | 17 May 2005 08:07 | |
| Jon Stephens | 17 May 2005 08:29 | |
| Stewart Smith | 17 May 2005 19:51 | |
| Jon Stephens | 17 May 2005 23:47 |
| Subject: | Re: Alter table clarification |
|---|---|
| From: | Jon Stephens (jo...@mysql.com) |
| Date: | 05/17/2005 11:47:10 PM |
| List: | com.mysql.lists.cluster |
Hi Stewart,
Stewart Smith wrote:
On Wed, 2005-05-18 at 01:30 +1000, Jon Stephens wrote:
This is correct and is intended behaviour. (Autodiscovery of schema changes tends to slow things down quite a bit, or so I've been told by one of the NDB developers.)
We could possibly extend the Event API in 5.1 to do events on schema changes (although I envisage this great horrible hack spanning mysqld and ndb... but maybe it can be done cleanly). However, this wouldn't be synchronous and you could still get these errors, just less often (maybe, in certain load situations).
The right fix would probably be one where if we encounter this error, we re-get the frm and try again. This is quite possibly more than a few lines of changes though.
Something for the future though...
Maybe this could also be configurable, like say a --mutation-event-retry=[on|off]|[# of tries]|[interval between tries] at startup or in the config.ini? Not sure which executable(s) would need to have such an option, though...
-- Jon Stephens Technical Writer, MySQL Documentation Team Brisbane, Australia Voice: +61 (7) 3388-2228 Email: jo...@mysql.com MySQL AB www.mysql.com
Jumpstart your cluster! http://www.mysql.com/consulting/packaged/cluster.html