UKOUG 2010 – 2 Days Later

Blogging from the Speaker Ready room which is now suited on the balcony overlooking the exhibition. I must say that I like the new location. Not only it is closer to the action, but allows you to oversee who is hanging out over the exhibitors.

There’s been a number of interesting sessions, but it’s little things here and there that I learn that provide some real value.

Here’s a few examples

• Within the first 10 minutes of Tanel’s Exadata migrations someone finally shared what exactly is the difference between each of the different types of columnar compression – the compression algorithm. L2ZIP, gzip, gzip high and bzip2. What’s even more interesting is how he determined it.
• Lary Carpenter’s Real Time Dataguard presentation he demoed life, automatic block corruption repair. The production database requests a block from the standby and repairs it, thus not application level errors.
• Julian Dyke’s presentation on “Inside Replication” – truncating a table with an materialized view log performs a “delete” on the log. It’s better to drop the materialized view log before truncating the table.
• Alex Gorbachev’s “Under The Hood of Oracle Clusterware 2.0” – showed a cool trick on running RAC nodes in smaller VM’s by unlocking the clusterware’s memory using gdb. He borrowed that trick from Jeremy Schneider’s blog.

Tomorrow at 8:45 AM GMT time (which is 3:45 AM on my internal EST time clock) I will be doing the updated version of my Memory presentation – “The Answer to Free Memory, Swap, Oracle and everything”. I should perhaps rename it to “Memory Tales from the Middle of the Night” for next year…