Re: Technical Interviews

Daniel Roy wrote:

> I have a generic question about the "technical interviews" we as DBA
> often have to go through to get a posting (I'm a consultant) or
> position: When interviewed by another DBA, do you ever feel that your
> answers should be tailored by the knowledge/myths of the DBA
> interviewing? For example, I got asked just last week during one of
> these interviews if "a full table scan is bad?". An "ordinary" DBA
> would answer "yes" right away. But anyone either doing some research
> on the topic on internet or trying different scenarios himself knows
> that the answer is "it depends!" (it seems to me that this is the
> answer to most of the intelligent Oracle-related questions). Some more
> examples of tricky questions I had to answer include "how often should
> we rebuild indexes?" (the correct answer is that some indexes should
> never be rebuilt, but try to explain that to someone who rebuilds ALL
> the indexes religeously every 3 months) or "How much bigger would you
> make the buffer pool if the hit ratio was only 80%?". It seems that I
> could go on and talk about each of these questions for half an hour
> easily. I went successfully through that interview, but it's only
> because I played the role of a DBA who follows the myths. Had I
> answered honestly to the questions, I'm sure he would have thought
> "what a freak!", and I would never have heard of them again!
>
> Daniel

When in situations where there is a mythology and a correct answer I always give both. For example:

Many people believe that ... because that was the case in some earlier versions ... but in 9i with the CBO ....

--
Daniel Morgan
http://www.outreach.washington.edu/extinfo/certprog/oad/oad_crs.asp
damorgan_at_x.washington.edu
(replace 'x' with a 'u' to reply)