Diary of a Paranoid MySQL Upgrade - 250GB database to 5.x

Interesting enough. But honestly all I took away from it was that Second Life (all of it?!) fits on a laptop drive.

Storage estimation is a mind bender in the modern world.

> We also used the maatkit tools to verify replication integrity.  I hooked up a 5.0.84 slave to a 4.1 master, replayed 24 hours worth of logs, and ran mk-table-checksum/mk-table-sync on the whole thing as a consistency check.  Unfortunately, the consistency check showed that the slave and master had slightly inconsistent data after only a few hours.  Doubly-unfortunately, these inconsistencies were not unique to the test scenario; we also see them in replication between 4.1->4.1, 4.1->5.0, and 5.0->5.0.  This is probably a topic for another blog post; for now, suffice it to say, I verified that the upgrade was not going to introduce any new bugs, and left it alone.a1

Emph. mine.

I wonder if dropping in a few SSDs would have fixed all their perf problems without having to spend that much time investigating it.

this time, we needed to know for damn sure that our shiny new db was as fast or faster than the older one. Not faster in some abstract sense, but verifiably faster * with our query set *

This is presented as a great insight, but I have here a book, Scaling Oracle 8i by James Morle from a decade ago that covers it, and people were doing this sort of thing 20 years before that!

So Linden Labs problem #1 is inexperienced DBAs, no wonder they made such a mess of it.