Comments on: HCC – 2

By: Anon.

Anon. — Tue, 11 Oct 2011 16:01:35 +0000

I forgot to add the whole point of my comment in relation to this post … Ive been having “fun” with oracle and indexes on HCC. We wanted to use it due to the shere size of the databases and the fact that its hard to get a fast backup solution in place. Ie cannot reliably/speedily backup 20Tb of data, so reduce the size down instead. Unfortunately, right now we are having to stick with regular compression.

By: Anon.

Anon. — Tue, 11 Oct 2011 15:16:38 +0000

Im in the, some would say fortunate, position of having 3 exadata boxes that I work on. Now, Im no expert .. but lets just say that when people talk about the “bleeding” edge, I know what they mean.
Sure I see the things flying, when they are up that is. But, I have to say, Im impressed … most of all with this flash cache and intelligent storage … BIIGGG queries running somewhat faster than any previous RAC system Ive worked on.

By: Jonathan Lewis

Jonathan Lewis — Wed, 05 Oct 2011 15:07:21 +0000

Possibly but I’ll have to wait a bit before I can restart my laptop to check exactly what’s there.

It’s also possible, of course that the odd 1000001 rows is because this ‘rownum = 1′ is followed by ‘insert 1,000,000′ – and I was only
looking for create, alter and drop statement.

Update: I’ve looked at the original trace, and it had “<= 1000001". I have no idea how I got the wrong predicate while extracting and editing the text.

By: Flado

Flado — Wed, 05 Oct 2011 14:30:05 +0000

Is there a “less than” and maybe six zeroes missing in the trace? I see four tables of one row (where rownum = 1) each created and then dropped.