[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Doing Kernel Upgrades



>> > The RedHat errata says to use ivh so that you have an old kernel to fall ba
>> ck
>> > on should something go wrong. I set up my lilo.conf so that I always have
>> > a kernel or 2 to fall back on. Sometimes thay just will not boot because
>> > you enabled something that your system is not equipped to handle (devfs 
>> > comes to mind).
>> 
>> Actually this should be bugzillaed against rpm itself. It should refuse to
>> -U any kernel. That simple trick would save newbies from one of the nastiest
>
>It's a bad principle that such rules be built into the tool.
>
>There might be a case for implementing a rule that can be spcified in the spec 
>file (and so he rpm) that says "upgrade means install" so that if someone 
>tries to upgrade such a package then rpm installs the new one instead.

I don't think "upgrade means install" is a good idea as it does
something that was not specified.  However a flag of some kind in
the .spec file that indicates that a package must not be
"Upgraded" or "Freshened", causing RPM to say:

rpm: this packages must be installed with -i, and can not be
     upgraded via -U or -F


>It would still be open to users to 'rpm --erase' the package. 
>
>It might be a good move to mark kernels:
>
>chatt -u /boot/vmlinuz-2.2.16

It's "chattr", and "-u" is meaningless.  ;o)  The chattr manpage
incorrectly states that -u does something.  It might set the
flag, but the kernel does NOT mark files for undeletion.  That
never got implemented.  Actually technically it did, but never
got integrated into the official kernel, and is unmaintained
AFAIR.  Besides "-u" would remove the u bit, not set it.

chattr +i /boot/vmlinuz-2.2.16

Marks the kernel as immutable, and thus is immortal until 
chattr -i /boot/vmlinuz-2.2.16

is done on it.

>hoping that by the time people figure how to undo that, they've also read up 
>the bit about installing new kernels.

Except that as noted above, the ext2 "u" attribute does
nothing.  ;o)

I've cc'd this to rpm-list so that we can see what is
realistically possible.


----------------------------------------------------------------------
   NOTE:  My new email address is:  mharris@opensourceadvocate.org

      Mike A. Harris  -  Linux advocate  -  Open source advocate
              Computer Consultant - Capslock Consulting
                 Copyright 2000 all rights reserved
----------------------------------------------------------------------

[Favorite quotes of Linus Torvalds - Sept 6, 2000]
I'm a bastard. I have absolutely no clue why people can ever think
otherwise. Yet they do. People think I'm a nice guy, and the fact is that
I'm a scheming, conniving bastard who doesn't care for any hurt feelings
or lost hours of work if it just results in what I consider to be a better
system.  And I'm not just saying that. I'm really not a very nice person. 
I can say "I don't care" with a straight face, and really mean it.
        -- Linus Torvalds on linux-kernel mailing list





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index] []