Log Buffer #234, A Carnival of the Vanities for DBAs

Well like all good things, summer is also ending and its closing days are rapidly approaching. For most of the database techies, there is nothing to get remorseful, because they are thrilled to be back in full flow managing and loving their databases. They are gearing up, refreshing up, re-evaluating and reassessing themselves and also blogging about it. This Log Buffer enjoys the last days of summer in this Log Buffer #234.

Oracle:

Many Oracle DBA’s are probably familiar with what Optimizer trace files are and likely know how to create them. Greg Rahn has more.

Ratios have become pariah in the Oracle performance tuning world, and Nial Litchfield hates them too.

Kerry Osborne has modified has great sql profile script to allow any arbitrary text sting including quote characters.

Kevin Closson blogs about that the Facebook adopts Oracle Exalogic Elastic Cloud, and replaces Tens Of Thousands Of Servers, well read along….

Doug Burns carries on with his blogging about the Real-Time SQL Monitoring and comes up with a new gem.

SQL Server:

Patrick O’Keeffe shares a graphical representation of SQL Server Version breakdown across all Project Lucy users.

CUME_DIST() calculates relative position of a value relative to a group of values. The value returned by CUME_DIST() is > 0 and <= 1, which represents percentage of number of rows with value less than (for ascending order) or equal to current row. Vishal has more.

Colin Stasiuk shared the session abstract: Unraveling Tangled Code – A Spellbinding Tale of Victory Over Chaos.

Mike Hillwig is not a fan of maintenance plans on SQL Server despite of the fact that they’re easy to set up and they work well in small environments.

When working with SQL in a cluster, the account rights on both nodes of the cluster need to be the same, Denny Cherry blogs.

MySQL:

Mark Callaghan says that If you run a MySQL master in durable mode (sync_binlog=1, innodb_flush_log_at_trx_commit=1) then your commit rate will not exceed the rate at which your server can fsync to disk, flash or a HW RAID card and in many cases that is a rate less than 10,000/second.

Set server id in my.cnf is one of the necessary steps to implement or start replication. The general rule is, server id must be unique across the environment where we are implementing MySQL replication. Nilnandan Joshi has more.

Multi-threaded (aka parallel) slave, and allowing out-of-order commits on such a parallel slave are two concepts which have been advocated by Henrik Ingo for the past some years and the Galera supports both of them.

In this blog post, Tony Darnell connects to a MySQL database, retrieves multiple rows of data, and prints a table in HTML with the data.

Gerardo Narvaja writes a spectacular post about MySQL HA Agent Mini HowTo.

Enjoy Summer !!!!