You're not alone. While attempting to install 10.1.0.4 patch for EM
Grid (different from the 10.1.0.4 patch for the DB) on Linux, the README
says "Don't install this -- grab 10.1.0.4.1 instead". But 10.1.0.4.1
for Linux doesn't exist. Support says "Yeah, it's really the 10.1.0.4
patch. It was never correctly renamed." Upon installing it, it blew up
which apparently trashed my repository, damagement server, and
additional agent. Or at least I wasn't able to reinstall, uninstall, or
use the existing components any more and decided I've wasted enough time
getting this "Enterprise" Manager working.
The Grid is no more.
Sent: Tuesday, June 28, 2005 2:35 PM
To: Oracle-L (oracle-l@(protected))
Subject: 10.1.0.4 patchset 2 catcpu.sql slop
w2k3 server std ed, 32 bit.
Oracle 10g R1, 10.1.0 patchset 3 applied (10.1.0.4)
While running the 10.1.0.4 patchset 2 post-installation script
(catcpu.sql) against a newly created database there were a rather
large amount of ORA-00955
(457) and ORA-02303
These can sometimes be ignored, but basically, this means having to
compare the previous definitions of the objects against the new
definition on a line by line basis in order to determine if in fact
the object definition has changed. This is a huge waste of time, IMHO.
I was thinking of rather than taking that approach, of manually
dropping all of these objects prior to re-running catcpu.sql and
examining the results (HUGE hammer approach).
this is in testing in a db with no users.
yeah, I could just ignore the patchset and stick with 10.1.0.4 patchset
we've hit an issue covered by 10.1.0.4 patchset 2 so I'd rather get
the bugfix out there than wait for the next iteration.
just trolling for some thoughts before I vent into an iTAR.