Re: Algorithm for calculating extent size in LMT

Howard J. Rogers doodled thusly:
>table*, subject to the proviso I made about the extent map. (For the record,
>I still don't like seeing more than 500 or so extents for anything on an 8K
>block system).

Yup, I can live with that one.

>
>How many extents were we talking about for the rollback segment, though?

In the high hundreds. To give you an idea: the uniform allocation size was 16K, the extent size for the RLB was 160K and the overall available space was 120Mb (small db). This was what I found from looking at the dict tables and v$ views (remember, this is 8.1.7). Unfortunately the DBA in this database didn't specify a proper uniform size for the RLB. Just re-used the stuff from other tablespaces.

Fortunately we caught it in time. Before prod was too far advanced. The offending recursive code is now fixed and the allocations for that tablespace have also been fixed. But it could have been a problem if we hadn't trapped it indirectly through the recursive problem.

>I've never encountered anything as serious as you describe, but I can well
>imagine that internal rollback handling is not bazillion-extent aware.
>

Yes, that is my feeling too. I've got a feeling we need more rteal-life experiences with LMTs before going for them in a big way. Of course, with 9i it makes sense. But if I was running 8iR3, I'd think twice before committing my disk space to LMT.

Cheers
Nuno Souto
nsouto_at_optushome.com.au.nospam Received on Tue Mar 05 2002 - 03:28:45 CST