UKOUG 2012 – Day One
Posted by John Hallas on December 4, 2012
Yes, I know I had the wrong year in the title, at least I cannot be accused of looking backwards all the time.
The UKOUG conference comes around again and the first two presentations re-confirmed why I attend. Jonathan Lewis was talking about generating data for test cases and I realised that I could generate the data for an interesting issue at work involving the non use of an index and a histogram with a bucket size of 60. It wasn’t so much that the data was hard to create it was that it had just not occurred to me to do so. I have seen Jonathan speak a number of times and I think I enjoyed this the most in that it was all real-world and practicable whereas sometimes I drift off as dives deeper than I either need or care to go.
My second eureka moment was when Graham Woods was talking about dynamic connection pooling room the middleware tier and the effect it has on the database and it really did make me wonder about how some of our tiered apps manage that, something that is relatively invisible to me at the moment but doesn’t have to be. My key takeaways were the lines ‘the apps tier was invented to queue’ and also ‘one of the most expensive oracle operations you can perform on a database is to connect’.
To contrast the two positives I will throw in two negatives, the perennial lack of seating at the ICC when trying to eat lunch and the obsession this year by the UKOUG with mobile apps. I normally use their web agenda app to book my sessions and then print it off. That has all been changed this year in favour of mobile apps. Not very good with a Blackberry and not particularly well thought through or implemented in my opinion. However I was even more surprised to find out that the session feedback forms have been binned in favour of inputting your data from a mobile app. I can see the reasons in terms of collating the data and saving cost but not everyone has a Smartphone and the experience on Blackberry is distinctly lacking. In fact I met another two friends at the evening reception who both had Blackberries and had experienced the same difficulties and had just given up on it. Just call me a Luddite but I did not see many people inputting data at the end of the sessions I attended.
Carl Dudley did a good talk on auditing which chimes with some work I am doing at the moment around the ‘db, extended’ parameter for the audit-trail parameter and writing to the syslog file. From what he was saying I think there might be some differences between *nix and windows environments – My experience is solely on *unix whereas Carl was talking about Wintel. If you do nothing else from this blog then have a look at the new (11GR2) v$xml_audit_trail view.
My second sql test case session of the day was Carlos Sierra on the various permutations of SQLTEXPLAIN. The options were somewhat confusing but there was no doubt that it is a powerful tool set. I will be interested to see what the SQL Health Check talk is about tomorrow because it sounds to be something similar.
My final session was Edgars Rudans talking about Grid Infrastructure upgrade from 188.8.131.52 to 184.108.40.206. I wonder how many people in the room had missed out the ‘on AIX’ in the presentation title that was not printed in the schedule, as I did. However it did shed light on similar issues we had experienced on HPUX, especially around the requirement for several scripts to be run directly as root and not as sudo root which we normally do.
A very good day and a pageful of things to look at when I get back to the office on Thursday.