SCIENTIFIC-LINUX-DEVEL Archives

July 2007

SCIENTIFIC-LINUX-DEVEL@LISTSERV.FNAL.GOV

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Jon Peatfield <[log in to unmask]>
Reply To:
Jon Peatfield <[log in to unmask]>
Date:
Mon, 23 Jul 2007 18:59:14 +0100
Content-Type:
TEXT/PLAIN
Parts/Attachments:
TEXT/PLAIN (32 lines)
On Fri, 20 Jul 2007, Stephen John Smoogen wrote:

>>  Personally I think that *some* updates should be marked 'reboot needed'
>>  since carrying on running with the old version means that tools will say
>>  'no new updates are needed' but the machine is still running with the
>>  old/insecure/broken version.
>
> The EL5 Gui has support for this and flags updates to glibc,kernel,
> and some others as requiring a reboot. However, doing this via a cron
> job at 4am means an email or some other such thing to tell a person a
> reboot should be done. Which is probably a good idea for the
> yum-cronjob program to have in it.

If you are meaning 'pup' then it seems to simply have a fixed list of 
package names to check for (which is clearly better than nothing):

rebootpkgs = ("kernel", "kernel-smp", "kernel-xen-hypervisor", "kernel-PAE",
               "kernel-xen0", "kernel-xenU", "kernel-xen", "kernel-xen-guest",
               "glibc", "hal", "dbus")

and check if we are updating (or obsoleting) anything which matches any of 
those names.

Clearly it would be more generic if a package contained that info as part 
of it's own metadata -- e.g. if libfoobar is updated and it will require a 
reboot we arn't forced to fix all scripts which want to know if updates 
need a reboot...

-- 
Jon Peatfield,  Computer Officer,  DAMTP,  University of Cambridge
Mail:  [log in to unmask]     Web:  http://www.damtp.cam.ac.uk/

ATOM RSS1 RSS2