how to create a seperate databases in W4Net ?

Hi Sachin,

Can you please clarify the question? Are you requesting information on the database split procedure?

Has your database grown beyond the max recommended size of 1 million metrics?

A single Watch4net server can manage thousands of objects and process millions of indicators. It is recommended that a single backend service be limited to process 1 million metrics. Databases beyond this size may suffer data corruption, poor performance and crashed tables.

To check on the metrics in the database to make sure it is necessary to split the database:

1.) Log into the MySQL database /opt/APG/bin/mysql-client.sh

2.) Run the following query from the mysql prompt mysql> select count(*) from data_variable;

3.) Check the output of the query is greater than 1 million:

+----------+

| count(*) |

+----------+

| 1501819  |

+----------+

1 row in set (0.11 sec)

In the above example, the output displays approximately 1.5 million for a returned metric count. This exceeds the max 1 million recommendation, and therefore, may result in performance degradation.

The database split procedure is a complex one and requires a Professional Services engagement.

If this is required, please contact you EMC Account Manager to initiate this engagement.

Please let me know if you have any additional questions.

Kind Regards,

Paul O'Rourke

Thank you Paul for the detailed reply.

Just wanted to know if this limit (1 million ) is only for events_live table or for both events_live and events_archive ?

Also is this limit there for datastore also or is only applicable for mysql database ?

Hi Paul,

Many thanks for the info.

What is data_variable and what is its significance ?

Hi Sachin,

As mentioned above, the data_variable is a table in the mysql database which keeps a record of the metrics collected during polling.

To see the contents of this table :

1.) Log into the MySQL database /opt/APG/bin/mysql-client.sh

2.) Run the following query from the mysql prompt mysql> select * from data_variable;

Thanks,

Kind Regards,

Paul O'Rourke