Performing a table dump
The RTM operational backup does not include all RTM tables. This is because of the large size of the backup and the impact on performance. You must manually dump some tables and restore them on the new RTM server during migration.
- Run the following dump command as root or as the RTM SQL account:
mysqldump --insert-ignore --no-create-info --lock-tables=FALSE cacti grid_jobs_finished grid_jobs_jobhosts_finished grid_jobs_reqhosts_finished grid_jobs_pendreasons_finished grid_job_daily_stats grid_jobs_rusage grid_job_interval_stats > /tmp/grid_tables.sql mysqldump --insert-ignore --no-create-info --lock-tables=FALSE cacti lic_daily_stats lic_daily_stats_traffic lic_interval_stats > /tmp/lic_tables.sql
- Run the following commands to restore these tables on the new server:
mysql cacti < /tmp/grid_tables.sql mysql cacti < /tmp/lic_tables.sql
Note: If the dump process takes a long time, run gridacct to see the time range between when the dump was started and finished. See the gridacct documentation for the procedure to get legacy LSF account records from clusters.
Important: Schedule the restore process of these tables to avoid periods of database maintenance. Perform the dump and restore as close to the end time as possible for the previous RTM maintenance cycle, but never before it has had a chance to complete.