Read the chapter 2 of Practical Time Series Analysis MemSQL.pdf to understand the consideration between SQL and NoSQL.

I checked the number of data in ATCamera_bonnShutter, ATCamera_wreb, and ATCamera_wrebPower on ts-efd-srv-01, and the values are 953088, 953190, and 953040 individually.

Check the memory usage by "free -m" as the following (unit is MB):

<ts-efd-srv-01> total: 991, used: 409, available: 318.

<ts-efd-data-01> total: 991, used: 665, available: 144.

The memory of data is set to be 512 MB (DataMemory = 512M) on ts-efd-mgmt-01.

Check the data size in ts-efd-srv-01 (/mnt/tier1), the value is 135 MB.

Check the data size in ts-efd-data-01 (/mnt/tier1), the value is 807 MB.

The disk sizes VBs on "/dev/sda1"  are 43 GB. Therefore, the disk size should be enough.

I got the following message after the execution of ndb_mgm -e "ALL REPORT MEMORYUSAGE" under ts-efd-mgmt-01:

Connected to Management Server at: localhost:1186
Node 16: Data usage is 94%(14254 32K pages of total 15074)
Node 16: Index usage is 61%(1310 32K pages of total 2130)

The Node 16 is "ts-efd-data-01".

The usage is 94% here because the 5% is reserved for the "restart". The default value of MinFreePct is 5, which means 5%.

Check the data in ts-efd-data-01, there is 521 MB under "/mnt/tier1/data/ndb_16_fs/LCP". This looks like the stored data.

Based on https://dev.mysql.com/doc/refman/5.7/en/mysql-cluster-ndbd-definition.html#ndbparam-ndbd-datamemory, the "PARTITION_BALANCE" is something valuable to try. The "MinFreePct" is valuable to use.

Discussed with Andres A. and he suggested the following page to clean the data in RAM for the test purpose:

https://confluence.lsstcorp.org/display/LTS/MySQL+Internal+configuration

We discussed about different technologies, Te-Wei got familiar with current puppet setup to install a MySQL Cluster. He talked to Andres V to get more familiar with the installation procedure in Puppet + Vargrant VMs.