Oracle DBA – A lifelong learning experience

Another really enjoyable UKOUG Database SIG

Posted by John Hallas on May 8, 2014

Today’s venue was the Metropole Hotel in the centre of Leeds and there was a good attendance, encouraged by a strong agenda.

After introductions, health and safety and UKOUG information it was straight into techie talk with Phil Davies from Oracle doing his normal support update (although he does share duties with Owen Ireland). Invariably I make more notes from this session than most others I hear because of the wealth of information it contains. Snippets I jotted down that are worthy of sharing are :-

12.1.0.2 RDBMS is due out Jul-Dec 2014 and is likely to contain new functionality that did not get into the first 12c release – I think in-memory will be one such feature. A straw poll of members saw very little uptake of 12c in production and some but not much take-up in non-production.

Oracle Clud Control (OEM) 12.1.0.4 is due very soon – Q2 2014

I was intrigued enough about MoS note 1557478.1 about a bug with transparent huge pages to make a note of it for further investigation and the same goes for 17761775.8 – Bug 17761775  ORA-600 [kclchkblkdma_3] ORA-600 [3020] or ORA-600 [kcbchg1_16] Join of temp and permanent table in RAC might lead to corruption

Next up was Phil Brown of E-DBA who ran an open forum where the audience picked from a selection of topics.

The three topics covered in about 70 minutes were

1)      Consolidation v Virtualisation

2)      The worst SQL statement in the world

3)      DBA interviews

 

All three allowed plenty of audience participation and competing views and I think almost everyone in the room provided input. My input regarding DBA interviews was a tip given to me several years ago which I think is very valid. Even if you are not thinking of moving on keep updating your CV so it is an accurate record of what you have been doing over the last couple of years.

A quick coffee break and then Neil Johnson talked about Jumbo frames in a RAC environment where you alter the default MTU size from 1500 to 9000 (typically) bytes. He provided a really good test case and lots of in-depth investigatory notes. The obvious question, which I am glad somebody asked because I was thinking it, was, if 9K is so good then why is the default 1.5K. The general acceptance was because things have moved on but the default has not and the analogy was thrown in that disk segment size used to be 512bytes but is now often 4K

A very good lunch, and the Phil Brown again talking about some of the functionality available in 12c Cloud Control around compliance and incident management. He also shared a Jython script to manage database password using the emcli interface which stirred a fair bit of interest.

Joel Goodman of Oracle delved into the innards of Automatic Parallel Execution taking about the pitfalls through the versions and how things had moved in – as always he crammed a lot of detail into a relatively short window but finished on time. There was a bit of conflict from some members of the audience about how much business want quicker finishing jobs if the timings might be unpredictable and the consensus from some was that consistency was of much more importance than dynamic adaptability that might vary the length of a job.

Another coffee break and then Martin Nash provided a live demo of some bugs and issues around Fast Connection Failover on a 2 node RAC cluster. I always admire people who do this sort of thing live and Martin handled it very well and gave a very good, in-depth presentation. However this is not the first presentation I have seen where the audience, or at least myself, get confused with all the jumping around between screens and sometimes let their minds wander. Well actually I don’t know if that happens to anyone else but it does to me.

So that was it, done and dusted and some went off to have a drink. As far as I know we had a first today at a SIG as whilst a speaker was talking an audience member was dealing with a major incident on a production system and diagnosing an issue and both co-ordinating and performing a database restart.

Thank you to Anna Crellin from UKOUG for arranging everything and to those who attended and presented. Another example of the value of UKOUG membership.

 

About these ads

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

 
Follow

Get every new post delivered to your Inbox.

Join 202 other followers

%d bloggers like this: