9:33 a.m.
OK, I have updated yum and then updated the kernel packages on my fc5dev
system. I notice that the old kernel-smp packages are removed so I only have
a couple of kernels install but the corresponding kernel-smp-devel packages
are NOT removed.
Is this a bug (or a feature)?
--
Gene
9:47 a.m.
On 11/11/05, Gene C. <czar(a)czarc.net> wrote:
Is this a bug (or a feature)?
that one is touchy. Anyone building modules for distribution might
want to keep the -devel packages around even if the kernels themselves
go away. But thats sort of a special case compared to the 90% usage
case.
As a user,even a user who occasionally needs to rebuild modules
locally i dont need the kernel-devel packages on the system if the
kernel is no longer there, so I'd want the -devel packages and any
associated module packages to disappear with the old kernels.
But there might be a need for a blacklist/whitelist for this plugin so
people can set specific wildcard lists to be ignored
ignore=kernel*-devel for example if you need to keep the devel
packages around without the kernels.
-jef
10:09 a.m.
On Fri, 2005-11-11 at 10:33 -0500, Gene C. wrote:
OK, I have updated yum and then updated the kernel packages on my
fc5dev
system. I notice that the old kernel-smp packages are removed so I only have
a couple of kernels install but the corresponding kernel-smp-devel packages
are NOT removed.
Is this a bug (or a feature)?
Hrmm.. the corresponding kernel-smp-devel should be getting removed as
well. Could you file in bugzilla so that I don't forget to look if I
don't get to it today?
Thanks,
Jeremy
10:34 a.m.
The installonlyn plugin is a bit too naive when (test) kernels are failing.
If I install a kernel that doesn't boot then the next kernel update will
wipe out my working kernel in favor of a newer kernel that may also not
work.
The simple solution would be to change the default number kept to be at
least one more but this is also likely to be a problem in the case of
failing test kernels.
A better solution would be to for installonlyn to only take action in the
case where the currently booted kernel (via uname -r) is the latest version.
darrell
10:40 a.m.
On Fri, 2005-11-11 at 08:34 -0800, darrell pfeifer wrote:
The installonlyn plugin is a bit too naive when (test) kernels are
failing.
If I install a kernel that doesn't boot then the next kernel update
will wipe out my working kernel in favor of a newer kernel that may
also not work.
The kernel you're booted into won't be removed. If the kernel is good
enough to install _new_ kernels then it should be reasonable to keep.
The simple solution would be to change the default number kept to be
at least one more but this is also likely to be a problem in the case
of failing test kernels.
A better solution would be to for installonlyn to only take action in
the case where the currently booted kernel (via uname -r) is the
latest version.
This is actually a very poor solution as it's very common for people
using the development tree to _not_ reboot every day. I generally ended
up with about 10 kernels of which I had maybe booted into 3 before
having to remove some to make space for more. This way, we're not
spending space on things that you haven't booted into
Jeremy
3:04 a.m.
This is actually a very poor solution as it's very common for
people
using the development tree to _not_ reboot every day. I generally ended
up with about 10 kernels of which I had maybe booted into 3 before
having to remove some to make space for more. This way, we're not
spending space on things that you haven't booted into
This should probably be different for the release compared to FC-devel?
regards,
Florian La Roche
10:41 a.m.
On Fri, Nov 11, 2005 at 08:34:29AM -0800, darrell pfeifer wrote:
A better solution would be to for installonlyn to only take action in
the
case where the currently booted kernel (via uname -r) is the latest version.
Or just to keep track of your kernels yourself, if you insist on using
rawhide.
10:42 a.m.
On 11/11/05, darrell pfeifer <darrellpf(a)gmail.com> wrote:
The installonlyn plugin is a bit too naive when (test) kernels are
failing.
If I install a kernel that doesn't boot then the next kernel update will
wipe out my working kernel in favor of a newer kernel that may also not
work.
It keeps 2 verions older than the "running" kernel by default right now.
-jef
9:13 p.m.
On Fri, 2005-11-11 at 11:42 -0500, Jeff Spaleta wrote:
On 11/11/05, darrell pfeifer <darrellpf(a)gmail.com> wrote:
> The installonlyn plugin is a bit too naive when (test) kernels are failing.
>
> If I install a kernel that doesn't boot then the next kernel update will
> wipe out my working kernel in favor of a newer kernel that may also not
> work.
It keeps 2 verions older than the "running" kernel by default right now.
It's only keeping one (plus the current) for me.
R.
--
"It's a fine line between denial and faith.
It's much better on my side"
6740
days inactive
6741
days old
8 comments
7 participants
participants (7)
-
darrell pfeifer -
Florian La Roche -
Gene C. -
Jef Spaleta -
Jeremy Katz -
Ralf Ertzinger -
Rodd Clarkson