Re: [cobalt-developers] screwed up file ownership and permissions

[Rob Montgomery <Rob@xxxxxxxxxxxxxxxxxx>]

Duh, never mind.  You ever have one of those things that you just cant
figure out, but its because of a really stupid reason?  Well this was it...
I was sidetracked completely by his previous chown -R blunder, blaming
it for the problem!

My client emailed me saying he cant get into  <virtualsite>/admin, and
that something was wrong.... Since I never use the admin web-interface
in the first place, and this is the only client I have using cobalt,  I didnt
know that it is <virtualsite>/siteadmin, and not /admin...  I finally figured
it out after investigating the admserv httpd.conf file...

Thanks for your time and help!

Rob




Chris Johnson wrote:

>         > From: Rob Montgomery <Rob@xxxxxxxxxxxxxxxxxx>
>         >
>         > [My client] did a chown -R as root and royally screwed
>         > up the machine.
>
>     It isn't pleasant, but you can do
>         % rpm -V -a
>         .
>         .
>         .
>         S.5....T c /etc/printcap
>
>     and get a list of all the files that the RPM database thinks
>     have been modified.  The files that have had their mode bits
>     changes will get a 'M' flag as the second character in the
>     first field.
>
>     So
>         % rpm -V -a | egrep '^.M'
>     Should give you a list of everything that has been changed.
>
>     You can then try to fix things by hand, or look at another
>     Cobalt box for the right value.  You can also grab most of
>     our rpms from
>         ftp.cobaltnet.com/pub/products/
>
>     Be careful about installing rpms to be sure that you don't
>     downrev things.  Use 'rpm -q -f' to get the rpm version, and
>     look for that RPM, either in the product directories, or the
>     updates.
>
>     It seems to me that you should be able to get the target
>     modes from the RPM databases in /var/lib/rpm, but I don't
>     see an easy way other than hacking up a perl script and
>     decyphering the rpm record format.
>
>     cj*