Using the History Database (MON-ExplorerHDB) - Latest

The MON-ExplorerHDB is an optionaldatabase tier that can be added to collect storage management events for analysis. The database tier is required only when frequent real time access to complex queries is needed. For example, if a large corporation needs to find the 100 largest datasets that reside across 5000 volumes, a database can provide a nearly instant answer. With a two-tiered solution, Monitor would otherwise have to analyze all 5000 volumes to find the largest datasets.

The database can also be used for historical analysis of the storage subsystem. For example, you can generate a list of datasets that occupied a specific volume at 3 am on the previous night.

The Explorer History Database (MON-ExplorerHDB) service collects and analyzes dataset and volume event records produced by the MON-RTM host component. This feature can be used to research past events or for trending and capacity planning via the Explorer GUI PC application.

The Explorer History Database runs on a Linux server. ExplorerHDB is written in Java and stores records produced by MON-RTM in a MySQL database server. The Explorer GUI contains predefined scripts used to query the Explorer history database and produce reports for analysis.

Running MON-ExplorerHDB

Once installed and configured, MON-ExplorerHDB runs continuously in the background as a Unix daemon.

When MON-ExplorerHDB is running you can view the expdb.log file for status information and you may stop and restart MON-ExplorerHDB when required for configuration or database changes. Everything else happens automatically.

Querying the MON-ExplorerHDB

The Explorer GUI PC interface is the vehicle for obtaining information from the Explorer history database. Once the connection to the database has been established (refer to Connect the Explorer GUI and to the Default History Database), the Database tab on the Explorer GUI navigation panel can be used to run existing queries or create new ones.

The Event Log folder contains scripts to view the dataset event records generated by the MON-RTM host component. The scripts may contain prompts that allow filtering so a subset of records can be viewed.

Once the grid is displayed, further filtering and sorting can be performed as well as charting, reporting and exporting. The Event Log folder also contains a sub-folder called Analysis that contains additional scripts for further analyzing the event records.

For example:

• The Dataset Growth script is a complex SQL query that determines how much datasets have grown in size for a specified date range. The output is sorted by datasets that have experienced the largest growth. When datasets grow steadily over time and span multiple volumes, it can be difficult to track which datasets are causing space problems. The Dataset Growth script can quickly and easily track this usage.

• The GDG Datasets script tallies all generations of a GDG base. It identifies how many current generations exist and the total space usage for all current generations. Once the output is displayed in the grid, right-clicking the GDG base name produces a sub-menu in which the listcat command can be invoked to list all current generations.

• The Multi-volume Dataset script tallies and totals multi-volume datasets so true dataset space usage across volumes can be identified.

For detailed information on how to modify or add scripts and how to manipulate data returned from script execution, refer to Using the Explorer GUI.

The History folder contains scripts for obtaining volume snapshot records for space utilization, used for trending and growth analysis.

Database maintenance

Trim database

Modify the file /usr/local/expdb/expdb_trim.sh as required. This file defines the interval periods used to remove old event and volume records. Currently the defaults are:

• event records: 2 months
• volume records: 2 years

Move database location on native Linux system

After installation the Exp MySQL database files are located in /var/lib/mysql/exp. The following steps move the MySQL data files to another filesystem.

Note: Replace /xxxx/data with the target location that will contain the moved MySQL data files.

1. Stop the MySQL service.

   /etc/init.d/mysqld stop

2. Create the directory that will hold the moved MySQL data.

   mkdir -p /xxxx/data

3. Copy all databases from the current data directory to the new directory.

   cp -R /var/lib/mysql/ /xxxx/data/

4. Back up the MySQL configuration file.

   cp /etc/my.cnf /root/my.cnf.backup

5. Edit the MySQL configuration file to update the data directory and socket values (if necessary).

   nano /etc/my.cnf

   Note: After editing, Save and exit your editor.

6. Set the datadir and, if required, the socket values to the new locations.

   datadir=/xxxx/data/mysql
           socket=/xxxx/data/mysql/mysql.sock

7. Update ownership of the moved data directory so the mysql user and group own it.

   chown -R mysql:mysql /xxxx/data/mysql

   Note: Ensure the mysql user and group exist on the system before changing ownership.

8. Rename the original data directory as a backup.

   mv /var/lib/mysql /var/lib/mysql-old

9. Create a symbolic link from the old data directory location to the new location so programs that expect the old path continue to work.

   ln -s /xxxx/data/mysql /var/lib/mysql

10. Back up the MySQL init script.

    cp /etc/init.d/mysqld /root/mysql.backup

11. Edit the init script if it contains hardcoded paths that need updating.

    nano /etc/init.d/mysqld

    Note: Save any changes after editing.

12. Locate the line(s) that reference get_mysql_option in the init script (this shows the configured datadir).

    grep -n "get_mysql_option" /etc/init.d/mysqld || grep -n "datadir" /etc/init.d/mysqld

13. Adjust the get_mysql_option (or datadir) value so it reflects the new directory.

    get_mysql_option mysqld datadir "/xxxx/mysql"

14. Start the MySQL service.

    /etc/init.d/mysqld start

    Note: If MySQL does not start, check the server error log and system logs, confirm the datadir and socket values are correct, and verify ownership and permissions on the data directory.