Posts Categorized: Pythian
Over-the-Top Tales from the Trenches:: Bringing order to the chaos of everyday DBA life.
Do you want to know a secret? Understanding it will prevent pain and gnashing of teeth, and also leave your face comfortably free of egg. Here it is: Some Oracle functions assume there are 31 days for each month of the year! The moral of the story: be careful when using MONTHS_BETWEEN for fractional dates.
Sheeri Kritzer has published the 25th edition of Log Buffer, the weekly review of database blogs, on The MySQL She-BA.
A short story depicting the life of an on call DBA
Welcome to the 24th edition of Log Buffer, the weekly review of the database blogosphere.
Carl Sagan is a personal hero of mine. Ten years ago today this gifted communicator died of cancer. And so today the community of bloggers that loved and admired Carl Sagan is having a spontaneous blogathon in his memory. You might ask, what does Carl Sagan have to do with being a good DBA? Believe it or not, a lot. Yes, really.
Welcome, reader, to the 23rd edition of Log Buffer, the weekly, human-edited review of the database blogosphere.
Oracle thinks it can make queries run better by not having to parse them repeatedly, so it grows the shared pool to keep as many queries as it can. Problem: the shared pool is now 4G+. That takes a while to trawl through, which of course adds to the spins on the library cache lock and pushes up CPU usage. I’m going to get into more detail on this when I have the time, but for this db, I think I’m going to switch to manual SGA and set a hard limit of 500M-1G on the shared pool
Brian Kelley has published Log Buffer #22, the weekly review of database blogs, on SQLServerCentral.com blogs.
In MySQL–land, failovers for redundancy, disaster recovery, or load balancing are performed by master databases and slave databases, the most popular method using binlog replication. There are a couple more methods of replication which aren’t covered here. Also see the MySQL Replication FAQ. The methods are the same, but the formatting of the procedure(s) are less than ideal.
Anyone interested in the architecture of highly scalable and stable infrastructure will enjoy this Presentation by Randy Shoup and Dan Pritchett of Ebay on how they, in their words, “strike a balance between site stability, feature velocity, performance, and cost.”