Announcement Module
No announcement yet.

pt-table-checksum false positives?

Page Title Module
Move Remove Collapse
Conversation Detail Module
  • Filter
  • Time
  • Show
Clear All
new posts

  • pt-table-checksum false positives?

    We are running the latest version of Percona Toolkit 2.2.7 on MySQL 5.1 with row based replication. Here is an example command that I am using:

    pt-table-checksum -u user-p pass -h master_ip_addr --set-vars innodb_lock_wait_timeout=50 --nocheck-plan --no-check-binlog-format --chunk-size-limit 100.0 --chunk-time 10.0 --nocheck-replication-filters --replicate percona.checksums > /path/to/checksum_output.log 2> /path/to/checksum_errors.log || pt-table-checksum -u user -p pass -h master_ip_addr --set-vars innodb_lock_wait_timeout=50 --replicate-check-only > /path/to/checksum_results_report.log

    The issue I am running into is that in the first command (using --replicate to compute the actual checksums), I see 1's appear in the DIFFS column for various tables. However, when I run --replicate-check-only command after the checksum command, it reports no DIFFS. This seems contradictory. I then checked percona.checksums table in the database for some of the tables in question and the CRC is identical on master and slave. In addition, there are no row mismatches between master and slave. Therefore, I'm left to wonder, why did the checksum output report a 1 in the DIFF column?

    Has anyone else run into this or know of why this would occur?