Re: What Would Cause....

From: Howard J. Rogers <hjr_at_dizwell.com>
Date: Sun, 05 Dec 2004 07:53:28 +1100
Message-ID: <41b223bf$0$20859$afc38c87@news.optusnet.com.au>

Niall Litchfield wrote:

> "Harvey" <harveyb_at_NoSpambeliveau.ccHere> wrote in message 
> news:41B06DE3.9C679990_at_NoSpambeliveau.ccHere...
> 

>>Hi All:
>>
>>I've encountered this situation twice, and I'm not sure of what the
>>cause is. If anyone has a hint of where to look I'd appreciate it.
>>
>>The problem - when looking at a database via DBA Studio I go to Storage
>>and then Tablespaces to display the list/utilization. Generally
>>speaking this takes a few seconds to load/display, however, what I'm
>>encountering are significant delays, in the order of up to ten minutes.
>>I've had two databases that were configured identically, with one
>>displaying this behavior and the other working fine. There's nothing
>>remarkable in the alert logs. Any ideas/suggestions?
>>
>>The environment - Oracle 8.1.7.4 on Solaris 8. (The environment for the
>>first encounter of this type was Oracle 8.1.7.4 on Windows 2000).

> 
> 
> I wonder if the *slow* tablespaces are in fact dictionary managed and have 
> rather a lot of extents in the objects in them, whereas the *fast* ones are 
> locally managed.
> You can however find out rather than guessing. You can trace the dba studio 
> session and then see (using tkprof, traceanalyser or some other tool) what 
> is consuming the time. ten minutes does sound excessive.
> 
> Cheers
> 
> 

I realise we don't have to guess, because the OP could test as you describe. But I'd like to predict that the slowly-displaying tablespaces will actually be the locally-managed ones. And the fast-displaying ones will be dictionary-managed. Not out of perversity, but because extent calculations are the one thing LMTs do worse than DMTs.

It will be interesting to see whether the style of extent management has anything to do with this at all!

Regards
HJR Received on Sat Dec 04 2004 - 14:53:28 CST