Load on the master when using pt-table-sync

Hi Percona User group,

I’m trying to access the performance implication of using pt-table-sync tool to replicate a collection of tables from a schema on production server (on load) to a replica schema on the same server. When the pt-table-sync is invoked, does the tool scan through all the data in the source schema and copy to the target schema ?
If yes, it would seem that the tool would create quite a bit of read load on the destination server (if the we have gigantic tables to the order of a TB) to make a copy on the target server ?

Could anyone share their experience in handling the replication scenario’s ?

1 Like

Hello, pt-table-sync is used to resolve data inconsistencies. For what you are trying to do I suggest you use pt-archiver instead. It can be configured to copy the data in small chunks to reduce impact. Check out Want to archive tables? Use Percona Toolkit's pt-archiver for more info

1 Like