Log Buffer #186: A Carnival of the Vanities for DBAs
Apr 23, 2010 / By Vanessa Simmons
Welcome to the 186th Edition of Log Buffer. Lots to report this week, so read on…
In Oracle news:
We begin with Gary Myers at the Sydney Oracle Lab who mixes GUI and CLI and shows how to manage your database from EMACS. You have to read a post that starts with: “There is a place of shadow, a place between the dark lands of the command-line interface, and the shining brightness of the GUI. In the days of yore, many dwelled in the shadow lands, but almost all have been attracted to the lights of SQL Developer…”
Tanel Poder gives a step by step tour of his latest troubleshooting adventure, solving mysterious wait events with a cause that AWR failed to clarify.
Tom Kyte discovered that even if you wait with a timeout, you may still get deadlocks. A bug, or merely unexpected feature? Find out…
Michael Janke at the Last In – First Out blog warns us about a nasty ZFS bug which can cause data loss in crash scenarios. Read his war story and upgrade your Solaris.
Alex Gorbachev posts his slides on Oracle Clusterware internals and how to design and run a 1TB MySQL database in a high availability setup from what he deems to be the longest Oracle conference ever, MiracleOpenWorld 2010. Alex is happy to be on his way home, after being stranded in Denmark for a week following the event.
On the MySQL front:
The new MySQL 5.5 beta was announced at the conference last week. Vadim tests performance in tpcc-like workload on Percona’a MySQL Performance Blog. On TheOpenforce.com, Zack Urlocker gives his take calling it a “tremendous improvement.” Baron Schwartz notes on Xaprb that great things are afoot in mysql 5.5, and offers the good, the bad, and the ugly.
Kristian Köhntopp goes down the dirty road with his latest hack for somebody who is using INSERT ON DUPLICATE KEY UPDATE.
Adam Machanic provides an excellent tip, and A Warning to Those Using sys.dm_exec_query_stats.
Antirez updated his blog with new features that will be included in the Redis 2.0 release. Among them is publish/subscribe interface that may turn Redis to something more than just another key-value data store. The release will also include non-blocking rehashing which will significantly improve performance when adding or removing instances with large amounts of memory.
The Apache Foundation released Cassandra 0.6. The new release includes an Hadoop interface, allowing Hadoop map/reduce processes to process data stored on Cassandra directly. Also included is an integrated row-cache, reducing the need for a seperate caching layer.
If you have any misconceptions around Cassandra, Spyced will be happy to correct you with his myth busting blog post. Think Cassandra is not reliable? Perhaps you should read this post and update your views.
Stuart Smith posts Ramblings about Storage Engine API: write_row, CREATE SELECT, and DDL. And Eric Day recaps Drizzle Developer Day.
Lastly, for MySQL, Postgres, Drizzle and others OpenSQL Camp was updated this week to show both of the opensqlcamps scheduled for 2010 — August in Germany, and October in Boston.
Leave a Reply