--- Paul Drake [EMAIL PROTECTED] wrote:
--- Jacques Kilchoer [EMAIL PROTECTED]
wrote:
In what cases does the SERIAL# need to be used?
orakill.
sqlnet.expire_timeout did not work on NT for 8.1.7.
garbage middle-tier apps that don't close
connections
require the use of orakill.
Thank you for your answer. My question was more along the lines why doesn't kill
session use audsid? Mr. Kangaraj provided the answer that audsid is 0 for some
sessions, or, in older databases, when AUDIT_TRAIL is not TRUE.
My question would then be: why doesn't auditing use SID/SERIAL# instead
Jacques,
In what cases does the SERIAL# need to be used? Can someone give an example
where a session-level command would be
applied to an incorrect session object if SERIAL# were not available?
For backward compatibility reasons :) Looks like AUDSID wasn't generated 7.2
and prior unless
Thank you John. I guess I can revise my question, then, to say Why not use
SID/SERIAL# all the time? :)
I suppose there are reasons but it seems unnecessary to have two sets of values that
uniquely identify a session. Though of course, from the documentation, AUDSID is
unique over the lifetime
--- Jacques Kilchoer [EMAIL PROTECTED]
wrote:
In what cases does the SERIAL# need to be used?
orakill.
sqlnet.expire_timeout did not work on NT for 8.1.7.
garbage middle-tier apps that don't close connections
require the use of orakill.
Pd
__
Do you Yahoo!?
