2 messages in com.mysql.lists.clusterRe: what's the safest way to alter ta...| From | Sent On | Attachments |
|---|---|---|
| mar...@idefix.net | 25 Nov 2004 06:16 | |
| Tomas Ulin | 26 Nov 2004 05:25 |
| Subject: | Re: what's the safest way to alter tables ? |
|---|---|
| From: | Tomas Ulin (tom...@mysql.com) |
| Date: | 11/26/2004 05:25:05 AM |
| List: | com.mysql.lists.cluster |
bash-2.05b$ perror --ndb 800 Error code 800: Too many ordered indexes (increase MaxNoOfOrderedIndexes): Permanent error: Insufficient space
after you get an "internal" ndb message yuo can also do:
mysql> show warnings; or mysql> show errors;
mar...@idefix.net wrote:
Hi all,
I've got a cluster running (1 mgmd, 2 ndbd's and 2 mysqld) on 2 machines.
I'm using MySQL 4.1.7 on Debian btw.
So far, so good.
Now, one of the machines is already running a database (with 48 tables)
which I want to convert to ndb.
As my SQL skills are less developed then my Perl skills I wrote a little
program to fetch the tablenames and do an ALTER TABLE after that.
The first 15 tables converted without a problem, but then the script got
"DBD::mysql::st execute failed: Can't create table './nagios/#sql-6e3a_115.frm'
(errno: 800) at ./hop.pl line 32." errors.
I suppose the sql-6e3a_115.frm file is a temporary table.
When this happened I was running out of time so I had drop all tables (twice, that was fun to find out) stop the mysqld's, remove all clusterfiles, remove the tables on both machines and put back the backup I earlier made.
Now I've got 2 questions.
1) is there a way to prevent the errno 800 of popping up ? 2) is there an easier way to drop all data in the cluster ?
Thanks for your time, Martin