pt-archiver should autolearn txn-size from limit if txn-size is not implied

[David Busby]

As the summary suggests, if --limit=100 is specified at runtime and --txn-size is not, then --txn-size should default to the --limit value, instead of defaulting to 1.

[Daniel Nichter (daniel-nichter)]

On second look/thought, I'm not sure that --limit and --txn-size should be linked like this. --limit is "how many rows to select to archive", but --txn-size is "how many rows to commit for each txn", i.e. they do very different things. If one only wants to select the first 10,000 rows to archive, that doesn't imply they want them all in a 10k row txn; maybe they want 100 rows per txn.

[Daniel Nichter (daniel-nichter)]

Since this isn't a bug, I'm going to close it. Perhaps we could discuss this on the Percona forums, see what other people think? But my feeling atm is that --txn-size and --limit are not/should not be related like this.

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PT-1089