0

at this moment I have configured a cluster that has 5 nodes:

2 data nodes 2 SQL nodes 1 node manager

My file.config.ini is the following:

[ndbd default]

NoOfReplicas = 2 # Number of replicas
DataMemory = 256M IndexMemory = 100M
ServerPort = 2202.
MaxNoOfConcurrentTransactions = 429496
MaxNoOfConcurrentOperations = 472496
MaxNoOfLocalOperations = 519745

[ndb_mgmd]

NodeId = 1
HostName = 192.168.10.145 # Hostname or IP address of MGM node
DataDir = / var / lib / mysql-cluster # Directory for MGM node log files

[ndbd]


HostName = 192.168.10.181 # Hostname or IP address
NodeId = 2 # Node ID for this data node
DataDir = / mnt / dataPartition / mysql / data # Directory for this data node's data files

[ndbd]

HostName = 192.168.10.183 # Hostname or IP address
NodeId = 3 # Node ID for this data node
DataDir = / mnt / dataPartition / mysql / data # Directory for this data node's data files

[mysqld]

HostName = 192.168.10.140 # Hostname or IP address
                                # (additional mysqld connections can be
                                # specified for this node for various
                                # purposes such as running ndb_restore)

[mysqld]

HostName = 192.168.10.184 # Hostname or IP address
                                # (additional mysqld connections can be
                                # specified for this node for various
                                # purposes such as running ndb_restore)

[mysqld]

IN THE NDB_MGM REPORT, IT SAYS THE FOLLOWING:

ndb_mgm> 2 REPORT MEMORYUSAGE
Node 2: Data usage is 18% (1476 32K pages of total 8192)
Node 2: Index usage is 11% (1461 8K pages of total 12832)
ndb_mgm> 3 REPORT MEMORYUSAGE
Node 3: Data usage is 18% (1480 32K pages of total 8192)
Node 3: Index usage is 11% (1462 8K pages of total 12832)

Each data node has 2gb of ram, it is assumed that with this configuration should not consume more than 500 mb of ram the ndbd process, but I am consuming 1.89 gb of ram per node, I am using disk storage for the fields that they are not indexed

What is wrong or what is wrong?

  • How did you compute the "500"? – Rick James Nov 18 '17 at 0:39
0

The 429496 transaction records consume around 400 MBytes (around 900 bytes per record).

The operation records consume also around 400 MBytes, around 300 bytes in transaction coordinator and 500 bytes in ldm thread.

In addition it uses memory for a lot of other buffers such as REDO buffers, SEND buffers, job buffers. These normally consume a few hundred MBytes.

In addition you can decrease memory usage by setting BatchSizePerLocalScan to a smaller value (e.g. 16). This should save a few hundred MBytes of memory space.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.