Re: SECUREFILES disaster

From: Jonathan Lewis <>
Date: Sat, 28 Apr 2012 10:01:37 +0100
Message-ID: <>

"Mladen Gogala" <> wrote in message
| On Fri, 27 Apr 2012 20:57:51 +0100, Jonathan Lewis wrote:
| > More like extent-level locking - and then only with the concurrency
| > impact of buffer busy waits, not table locks.
| > I've seen some undesirable effects with LOBs that can't be worked
| > around, but nothing catastrophic (yet).
| How about abnormal growth and space consumption?
| --

I have seen that, but not for quite a long time, and it was with ASSM tablespaces.

But I've also seen that with bitmap indexes, btree indexes, and simple heap tables in various versions and under various circumstances in the complete absence of LOBs. Apart from things that you could definitely call bugs, these phenomena also appear when some Oracle developer has missed a possible boundary condition in how an application may use their feature so my typical approach is to figure out where the collision is between the types of things that Oracle is probably doing and the nature of the application activity.


Jonathan Lewis
Oracle Core (Apress 2011)
Received on Sat Apr 28 2012 - 04:01:37 CDT

Original text of this message