RE: 9.2.0.6 for win32

>
>
> The number of responses I have gotten on this on and off-line
> indicates =
> it happens a lot. Niall, maybe this would be a good subject for a =
> thorough thrashing on paper. How to identify and fix?
>

Best advice I can think of (apart from not using Windows;) would be to stop all Oracle related services on Windows "box", set them to manual startup then re-boot to be 100% sure of shaking off any lingering file locks (uptime is NOT a "sacred cow" in Windows world, especially not on a dedicated Oracle server that have to "go down" anyway in order to be patched).

Other patch-related-behavior I stick to - just after flipping back related database service(s) from manual to automatic startup I'd go into registry HKLM\SOFTWARE\ORACLE\HOME<N> to temporary change value of (any) key(s) like ORA_<SID>_AUTOSTART from TRUE to FALSE. Doing so will ensure that after reboot and intentional or accidental start up of Oracle database service(s) will NOT automatically mount related database(s) but will give "something-little" (the service) to SQL*Plus to connect to making it safe 'n easy for "startup migrate" that follows (call it a superstition - but I am somehow reluctant to allow freshly patched Oracle binaries to mount and/or try to open the database that did not yet went through the catpatch routine).

Branimir

--
http://www.freelists.org/webpage/oracle-l