Re: [SQL] Re: [HACKERS] Re: INSERT/UPDATE waiting (another example)

Hi,

> 6.5 beta speeds up vacuuming with existing indexes, thanks to Vadim.
> Also, accessing during vacuuming may be better too.

That is good news :) When I first heard about MVCC I remember someone
suggested it would be possible to still do SELECT on tables being
vacuumed, is this right or not in the current 6.5?

When we were developing the system we spent a lot of time working out
ways of getting around vacuum, and I've learned a lot from it. I am going
to try it out on a full dump of our current database and test some
examples to see what kind of improvement there is.

> 6.5 beta also has a faster and smarter optimizer.
>
> It may be wise for you to test 6.5beta to see how many problems we fix.

This week I intend to test out the patches I've received, and hopefully
they will fix up my big problems (the one with the backend locking up)
then I will grab the latest 6.5 and try that out with some test data to
see what happens.

Unfortunately, I can't test 6.5 like I would the real thing because many
of my problems only occur when everyone is busy firing off queries and the
box is running an unusually high load and things start waiting on locks.
I'll see what I can do here although the only true way is to go live with
it - but I'm not ready for that yet :)

I should be able to check the optimiser improvements though, I've got a
lot of code which does the SET COST_HEAP/COST_INDEX hack to make things
work :)

thanks,
Wayne

------------------------------------------------------------------------------
Wayne Piekarski Tel: (08) 8221 5221
Research & Development Manager Fax: (08) 8221 5220
SE Network Access Pty Ltd Mob: 0407 395 889
222 Grote Street Email: wayne(at)senet(dot)com(dot)au
Adelaide SA 5000 WWW: http://www.senet.com.au