Re: optimal size for rollback

Two options -

1. Remember that Oracle doesn't discard extents just because they are past their sell-by (undo_retention), smon kicks in from time to time (every 12 hours) and does the equivalent of 'optimal'. So if an extent is freeable and waiting for smon to free it, then it is "safe" for another segment to steal it
2. An extent isn't past its sell-by, but another segment needs the space. In this case, it isn't "safe" to nab it, but the segment steals it anyway.

If you look at v$undostat, there are various columns there which report the number of times the various extent-related activities have occurred - if you see excessive stealing, you just have to increase the tablespace size (or fix the application).

--
Regards

Jonathan Lewis
http://www.jlcomp.demon.co.uk

Next Seminar dates:
(see http://www.jlcomp.demon.co.uk/seminar.html )

____England______September 24/26, November 12/14

____USA__________November 7/9 (MI), 19/21 (TX)

The Co-operative Oracle Users' FAQ
http://www.jlcomp.demon.co.uk/faq/ind_faq.html








Niall Litchfield wrote in message
<3d8f6ea4$0$1293$cc9e4d1f_at_news.dial.pipex.com>...


>>                                             Adding/removing segs on the

fly, but


>> more significantly, extent stealing, ie one segment to another 'ah,

>> you've finished with that extent, I'll nab it'.

>

>This is one of the things that I'm fully prepared to say 'nope don't see

how


>it does that' or more accurately 'Can I be absolutely sure that oracle will

>never reallocate an extent elsewhere when it shouldn't'. If I can then

maybe


>I'll go with it but its all a bit hocus-pocus to me at the mo.