2019-06-11 09:23 UTC

Fun MySQL fact of the day: binlog row image

Considering yesterday's downsides to the ROW binary log format, you may have spent the night sad and upset, maybe even unable to sleep. Likely, the narrative was along the lines of, "why, oh why, MySQL, must you make me choose between binary log size and safety?!" Well, rest-assured, tonight, you'll ... read more

2019-06-10 12:56 UTC

Fun MySQL fact of the day: ROW format

Over the last couple weeks, all of the examples we've considered have used the STATEMENT binary log format. We've also discussed, no less than two times, that the STATEMENT format is prone to incorrect results for non-deterministic DML (Data Manipulation Language) statements, and last Friday, we considered the ROW ... read more

2019-06-06 12:43 UTC

Fun MySQL fact of the day: PITR without GTIDs

When we isolated yesterday's bad transaction using GTIDs, life was pretty simple. But in some cases, GTIDs won't be an option for us. So, today, we'll take a slightly closer look at the binary log and figure out an alternative. ... read more

2019-06-05 11:05 UTC

Fun MySQL fact of the day: PITR with GTIDs

Yesterday, we looked at MySQL's MASTER_DELAY functionality, which we may use to maintain a delayed replica in our important database clusters. A couple weeks ago, we also looked at a very basic point-in-time recovery technique using mysqlbinlog. Today, we're going to combine the two with a universal ... read more

2019-06-04 12:37 UTC

Fun MySQL fact of the day: replication lag can be good

We've spent number of days discussing replication delay, and we've mostly considered replication delay as bad. And, true, in most cases, it usually is pretty bad, but today we're going to consider a case when replication delay is as desirable as it is important. ... read more

2019-06-03 23:18 UTC

Fun MySQL fact of the day: MVCC-caused replication delay

Looking way back to March 11, we had discussed InnoDB's MVCC implementation. We also had discussed how InnoDB (by default) needs to keep around all versions of a record in the InnoDB undo log until all queries referencing the record either commit or roll-back. And while it may not ... read more

2019-05-31 08:00 UTC

Fun MySQL fact of the day: absolute delay

Yesterday, we established an understanding of Seconds_Behind_Master and identified some of its shortcomings. While it's not a completely useless metric, it's also not a very helpful one. What we probably really want is an absolute measure of replication latency, and we'll look at one way of accomplishing that today. ... read more

2019-05-30 08:00 UTC

Fun MySQL fact of the day: seconds behind master

Yesterday, we started considering some causes of replication delay. Given L. Peter Deutsch's assertions about the fallacies of distributed computing, I think we can safely agree that replication delay is going to happen. And, not only will it happen, it will happen frequently and unpredictably. So while we ... read more

2019-05-29 10:33 UTC

Fun MySQL fact of the day: replication delay

Now that we have a basic understanding of how MySQL's replication mechanism works, let's poke it with a stick and see what happens. That is, of course, fun. ... read more

2019-05-28 12:39 UTC

Fun MySQL fact of the day: the SQL thread

Yesterday, we looked at how a MySQL replica's IO thread downloads a master's binary logs and stores them as its own "relay logs". We also said this happens until the IO thread is stopped or the replica database runs out of disk space, and while that's true, there's another ... read more