Hi,
I have preoccupations with a slowness in my select with a cluster MySQL.
Hardware (two server):
*
Dual-Xeon@3.00GHz
* 2Go of Ram with dual channel
* Harddisk 120Go SATA2
* Gigabits Network
I tested:
1) 1 Data node, 2 SQL nodes
2) 2 Data node, 2 SQL nodes
The table with a problem made to me : 700K entries (+/- 60Mo data, +/- 60Mo index).
When i select the 30th last entries, that lasts 80s. on the node data+sql, and 140s. on the node data, config 1)
When i select the 30th last entries, that lasts 80s. on the two node, config 2). Other informations : during the select, CPU < 4.0%, Network < 4Mbps, MEM < 20%.
Without MySQL Cluster, with load at 10 (same hardware) : 48s.
config.ini:
---------------
[NDBD DEFAULT]
NoOfReplicas= 1
DataMemory= 900M
IndexMemory= 730M
RedoBuffer= 50M
TimeBetweenLocalCheckpoints= 19
MaxNoOfUniqueHashIndexes= 128
MaxNoOfTables= 256
MaxNoOfOrderedIndexes= 384
MaxNoOfConcurrentOperations= 100000
[MYSQLD DEFAULT]
BatchSize= 128
BatchByteSize= 64K
[NDB_MGMD DEFAULT]
[TCP DEFAULT]
[NDB_MGMD]
Id= 1
HostName= **
[NDBD]
Id= 2
HostName= **
DataDir= /var/lib/mysql-cluster
[NDBD]
Id= 5
HostName= **
DataDir= /var/lib/mysql-cluster
[MYSQLD]
Id= 3
HostName= **
[MYSQLD]
Id= 4
HostName= **
---------------
When I import the dump (280 Mo with ENGINE NDB), CPU < 4% and network < 4Mbps.
Any idea ?