We’ve been using pt-table-checksum for several years to check replication consistency. Just recently we started getting errors for slaves that are 2 hops away from the master. I was under the impression that pt-table-checksum only checks a master and its IMMEDIATE (i.e. 1 hop) slaves. I looked through the documentation and googled for a while but couldn’t find anything saying whether pt-table-checksum finds all slaves in a chained replication topology. Can someone clarify this issue? btw, we are using the “hosts” method of slave discovery in case that makes any difference.
By default pt-table-checksum tries to discover all possible slaves. This can be controled with the --recurse option to limit how far down the chain it goes to find additional slaves.
Scott, thank you so much for that info. I guess I completely skipped over that during my reading, and I thought I was thorough. That helps a lot!
No problem. Glad I could help! =)