pt-table-checksum suggestion needed

What is the suggested
To run pt-table-checksum on master or slave, what will be the difference in running on master and slave.


pt-table-checksum needs to be run on the master. It automatically detects the slaves by methods defined in --recursion-method, where the default being processlist, hosts. It runs the checksum queries on the master, and saves the results to a table. It uses statement-based replication for the session and so the same checksum query will be executed on the slaves as well. The default table that stores the checksum results is percona.checksums, which could be set by --replicate option. At the end, the tool queries the checksum table on the slaves and output any inconsistency.

Please refer to the document for more details and options:–recursion-method

Give it a try and take a look at percona.checksums table on master and slave to see what the tool recorded.

Hello psong,
Appreciate your time and response, am also expecting the same thing. It shows you are more familiarity with checksum tool, have posted another query for checksum about it’s behavior when run on master and read from the table percona.checksum, can you please have a look at that query and suggest me how to over come that error.


[I]I tried the tool on slave and the slave was able to find the data drift between master and slave and send me an alert.

So i got the doubt like may be the tool was designed to run on slave and check the data drift between master and slave.[/I][/URL]

Appreciate your time and energy.


Just saw your reply and I will check the other post now.