Percona Toolkit moved to https://jira.percona.com/projects/PT

  1. Bug #925781
  2. Comment #7

Comment 7 for bug 925781

Revision history for this message
Kolbe (kolbe) wrote :

I think the suggestion to use utf8mb4 is misguided. The data going into lower_boundary and upper_boundary never needs to be manipulated as character data. If someone is using a BINARY PK, there will always be sequences that are illegal in both utf8 and utf8mb4.

Why not use BLOB instead of TEXT for these columns? That should resolve the issue completely.