Re: alert #68 patches

Yes, exactly. Thanks for stating it much better than I.

FYI, I granted select on V_$INSTANCE to the local role and switched the scripts to get the instance_name from that instead of restoring access to V_$PARAMETER for all users, which has a lot more in it and maybe that is why it was revoked.

Catherine

At 04:10 PM 9/30/2004, Walter Montalvo wrote:
>I think Catherine's point is that V_$PARAMETER was removed from the SELECT
>privileges for a local role.
>
>So, for instances that have custom roles with grants on V_$PARAMETER, the
>patch might remove them, breaking the applications/scripts that depend on
>the role having the proper grants.
>
>This will not show itself as something wrong with the view
>V_$PARAMETER. It would show itself as objects becoming INVALID after the
>patch, due to the removal of the grant. However, for scripts, you cannot
>tell until you run them.
>
>This should be the case on most installations, I would imagine.
>
>Walter
>
>
>At 11:47 AM 9/30/2004, Jerome Roa wrote:
>
> >I upgraded from 9204 to 9205 and applied the #68 patch and it seems to
> be OK.
> >Sun Solaris. V_$PARAMETER is ok. so is V$PARAMETER
> >
> >
> >At 02:39 PM 9/30/2004 -0400, Catherine LeBlanc wrote:
> > >After applying patch 3811887 for alert #68, all of a sudden we had scripts
> > >failing. It appears that select on V_$PARAMETER was removed from a local
> > >role by the patch. This did not show up in our initial post-patch testing
> > >since my group all has SELECT_CATALOG_ROLE which still includes select on
> > >V_$PARAMETER.
> > >Has anyone else experienced anything like this? We are on 9.2.0.4 on Tru64
> > >UNIX.
> > >
> > >Thanks,
> > >Catherine LeBlanc
> > >
> > >--
> > >http://www.freelists.org/webpage/oracle-l
> >
> >--
> >http://www.freelists.org/webpage/oracle-l
>
>--
>http://www.freelists.org/webpage/oracle-l

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