Re: multiple keys - Kishore Jalleda - com.mysql.lists.cluster

12 messages in com.mysql.lists.clusterRe: multiple keys

From	Sent On	Attachments
Matthew Story	15 May 2006 14:09
pek...@mysql.com	15 May 2006 14:26
Matthew Story	15 May 2006 14:32
pek...@mysql.com	15 May 2006 14:42
Matthew Story	15 May 2006 15:37
Kishore Jalleda	15 May 2006 15:47
Kishore Jalleda	16 May 2006 10:29
Kishore Jalleda	16 May 2006 11:17
Matthew Story	16 May 2006 11:18
Matthew Story	16 May 2006 11:55
Matthew Story	16 May 2006 12:42
Stewart Smith	17 May 2006 04:44

Subject:	Re: multiple keys
From:	Kishore Jalleda (kjal...@gmail.com)
Date:	05/16/2006 11:17:30 AM
List:	com.mysql.lists.cluster

Please reply to the entire cluster mailing list so that everybody can read it and give some valuable suggestions

Thanks Kishore

On 5/16/06, Matthew Story <matt...@gmail.com> wrote:

When I restart everything as per that guide the error goes away, but the other error (too many indexes 4242) is back now even though I've increased the number of unique hash indexes allowed to 128.

On 5/16/06, Kishore Jalleda <kjal...@gmail.com> wrote:

Look at this previous post on the same issue http://forums.mysql.com/read.php?25,75550,75550 and also the bug http://bugs.mysql.com/bug.php?id=18676

These might give you more hints .....

I'm still getting this "table already exists" error despite the fact that the table is MyISAM and not NDBCLUSTER at this point. I've also tried altering the table by changing the table engine to MyISAM and INNODB and neither of these changes goes through either, failing with the same error. Any ideas?

On 5/15/06, Matthew Story <matt...@gmail.com> wrote:

will have to first create the database ('CREATE DATABASE') on all the SQL nodes in the cluster (table data is then seen by all the sql nodes as there is auto-discovery of tables), only then all you SQL nodes

I should have mentioned that I've already done that in the previous e-mail. Sorry.

On 5/15/06, Kishore Jalleda <kjal...@gmail.com> wrote:

On 5/15/06, Matthew Story <matt...@gmail.com> wrote:

Well we've made some progress. I now get a different error that tells me the table already exists when I try to alter the engine. When I check from another MySQL daemon connected to the cluster I find that these tables aren't actually in the cluster, even if I perform a flush tables command on the other database. This has happened before after i've had troubles, but this usually just required me to restart the mysql server, now I've tried that and I still have the error.

Since the Cluster doesn't support auto-discovery of databases, you will have to first create the database ('CREATE DATABASE') on all the SQL nodes in the cluster (table data is then seen by all the sql nodes as there is auto-discovery of tables), only then all you SQL nodes would see the DB, so I suggest you do this first and then try ALTER TABLE .....

On 5/15/06, pek...@mysql.com <pek...@mysql.com> wrote:

since my first error I doubled the value of max number of tables and the error is still coming up.

How about MaxNoOfUniqueHashIndexes ?