MySQL Forums
Forum List  »  NDB clusters

NDBCLUSTER engine - Table is full
Posted by: taner selim
Date: May 26, 2009 03:39AM

Hi all i am having a very big problem trying to test the performance of mysql cluster MySQL 5.1.32-ndb-7.0.5-cluster-gpl (64bit). I cannot insert more than around 1,200,000 rows in a very simple table.

I need a resolution ASAP. I am really looking forward to migrating to mysql cluster but at this point i am in the dark with issues like 'table is full' which in this case doesnt make sense to me!!!

I am sure that my RAM and CPU are not the problem, but here goes the config info:

Machine configuration
2 storage Nodes (64bit os 2GB Ram Each), 1 management server, 2 replicas

=================================================
Config.ini from management server:
[NDBD DEFAULT]
NoOfReplicas=2
DataMemory=1500M
IndexMemory=256M

[MYSQLD DEFAULT]

[NDB_MGMD DEFAULT]

[TCP DEFAULT]

# Section for the cluster management node
[NDB_MGMD]
# IP address of the management node (this system)
HostName=192.168.1.21

# Section for the storage nodes
[NDBD]
# IP address of the first storage node
HostName=192.168.1.22
DataDir= /var/lib/mysql-cluster

[NDBD]
# IP address of the second storage node
HostName=192.168.1.23
DataDir=/var/lib/mysql-cluster

# one [MYSQLD] per storage node
[MYSQLD]
[MYSQLD]


=================================================

table definition:
CREATE TABLE `section` (
`s_id` int(10) unsigned NOT NULL AUTO_INCREMENT,
`s_name` varchar(50) NOT NULL,
PRIMARY KEY (`s_id`)
) ENGINE=ndbcluster DEFAULT CHARSET=utf8

=================================================
current limit;

mysql> select count(*) from section;
+----------+
| count(*) |
+----------+
| 1201566 |
+----------+
1 row in set (0.00 sec)

=================================================

Node memory usage :

dbClusterDB1:~# free -t -m
total used free shared buffers cached
Mem: 1968 875 1092 0 141 476
-/+ buffers/cache: 257 1710
Swap: 400 0 400
Total: 2368 875 1492


dbClusterDB2:~# free -t -m
total used free shared buffers cached
Mem: 1968 871 1096 0 140 474
-/+ buffers/cache: 256 1711
Swap: 400 0 400
Total: 2368 871 1496

=================================================

I must also mention that the storage nodes are in fact virtual servers but that should not be a problem since they have been allocated enough ram for this testing purpose. I need to be able to accommodate around 27,000,000 rows in a single table with around 15 columns (no blobs). Of course i am aware that for this i will need 'a bigger boat' but if i cant get more than 1.2 million rows in a two column table with 2GB ram, then its kinda worrying !

Please help. THanx
Taner



Edited 1 time(s). Last edit at 05/26/2009 03:43AM by taner selim.

Options: ReplyQuote


Subject
Views
Written By
Posted
NDBCLUSTER engine - Table is full
18291
May 26, 2009 03:39AM


Sorry, you can't reply to this topic. It has been closed.

Content reproduced on this site is the property of the respective copyright holders. It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.