I having a table which contains around 2 million records, all of which are for same node. I cannot partition based on date, as data is continually inserted and deleted each day.
The remaining data is in string format.
Every 15 minutes, the application takes backups from this table (only partial data which was collected for that 15 minutes); while doing this operation, the table is locked and unable to be accessed.
What kind of partitioning can I use to help solve this problem?