10:50 a.m.
Hello,
I feel like there are an increasing number of packages requiring a
system reboot. I'm wondering why. The following updates were installed
today, and required a full system reboot. I can't seem to find any
package in the list that I can conceivably see requiring a reboot, is it
that PK doesn't have the concept of X logout vs reboot? Is it a bug in
the packaging or PK or is there anything I can do/file to improve the
situation?
Dec 15 09:07:21 Updated: glib2-2.22.3-1.fc12.x86_64
Dec 15 09:07:23 Updated: mysql-libs-5.1.40-1.fc12.x86_64
Dec 15 09:07:23 Updated: gpm-libs-1.20.6-9.fc12.x86_64
Dec 15 09:07:25 Updated: mysql-5.1.40-1.fc12.x86_64
Dec 15 09:07:28 Updated: PyQt4-4.6.2-5.fc12.x86_64
Dec 15 09:07:32 Updated: mysql-server-5.1.40-1.fc12.x86_64
Dec 15 09:07:34 Updated: gpm-1.20.6-9.fc12.x86_64
Dec 15 09:07:37 Updated: 1:tk-8.5.7-3.fc12.x86_64
Dec 15 09:07:37 Updated: mpfr-2.4.1-5.fc12.x86_64
Dec 15 09:07:39 Updated: foomatic-4.0.3-5.fc12.x86_64
Dec 15 09:07:40 Updated: mysql-embedded-5.1.40-1.fc12.x86_64
Dec 15 09:07:41 Updated: glib2-2.22.3-1.fc12.i686
Dec 15 09:07:45 Updated: gtk2-2.18.4-3.fc12.x86_64
Dec 15 09:07:58 Updated: 1:gdm-2.28.1-25.fc12.x86_64
Dec 15 09:07:58 Updated: ibus-libs-1.2.0.20091204-2.fc12.x86_64
Dec 15 09:07:59 Updated: imsettings-libs-0.107.4-4.fc12.x86_64
Dec 15 09:08:02 Updated: glib2-devel-2.22.3-1.fc12.x86_64
Dec 15 09:08:07 Updated: yelp-2.28.1-1.fc12.x86_64
Dec 15 09:08:10 Updated: f-spot-0.6.1.5-1.fc12.x86_64
Dec 15 09:08:13 Updated: 1:xscreensaver-base-5.10-4.fc12.x86_64
Dec 15 09:08:13 Updated: 1:xscreensaver-gl-base-5.10-4.fc12.x86_64
Dec 15 09:08:16 Updated: gtk2-devel-2.18.4-3.fc12.x86_64
Dec 15 09:08:25 Updated: totem-2.28.4-1.fc12.x86_64
Dec 15 09:08:25 Updated: totem-nautilus-2.28.4-1.fc12.x86_64
Dec 15 09:08:27 Updated: 1:xscreensaver-gl-extras-5.10-4.fc12.x86_64
Dec 15 09:08:29 Updated: 1:xscreensaver-extras-5.10-4.fc12.x86_64
Dec 15 09:08:34 Updated: imsettings-0.107.4-4.fc12.x86_64
Dec 15 09:08:34 Updated: 1:gdm-user-switch-applet-2.28.1-25.fc12.x86_64
Dec 15 09:08:35 Updated: 1:gdm-plugin-fingerprint-2.28.1-25.fc12.x86_64
Dec 15 09:08:35 Updated: totem-mozplugin-2.28.4-1.fc12.x86_64
Dec 15 09:08:37 Updated: python-reportlab-2.3-1.fc12.x86_64
Dec 15 09:08:38 Updated: jna-3.2.4-1.fc12.x86_64
Dec 15 09:08:38 Updated: memcached-1.4.4-1.fc12.x86_64
Dec 15 09:08:38 Updated: less-436-3.fc12.x86_64
Dec 15 09:08:40 Updated: cscope-15.6-6.fc12.x86_64
Dec 15 09:08:40 Updated: xorg-x11-drv-mouse-1.5.0-1.fc12.x86_64
Dec 15 09:08:40 Updated: f-spot-screensaver-0.6.1.5-1.fc12.x86_64
Dec 15 09:08:46 Updated: gtk2-devel-docs-2.18.4-3.fc12.x86_64
Dec 15 09:08:46 Updated: gpm-devel-1.20.6-9.fc12.x86_64
Dec 15 09:08:46 Updated: liveusb-creator-3.9-1.fc12.noarch
Dec 15 09:08:48 Updated: mysql-devel-5.1.40-1.fc12.x86_64
Dec 15 09:08:53 Updated: etoys-4.0.2339-1.fc12.noarch
Dec 15 09:08:59 Updated: ibus-1.2.0.20091204-2.fc12.x86_64
Dec 15 09:08:59 Updated: ibus-gtk-1.2.0.20091204-2.fc12.x86_64
Wouldn't it be sufficient to logout? Is it a bug?
--
Nathanael
10:54 a.m.
On Tue, 15 Dec 2009, Nathanael D. Noblet wrote:
Hello,
I feel like there are an increasing number of packages requiring a system
reboot. I'm wondering why. The following updates were installed today, and
required a full system reboot. I can't seem to find any package in the list
that I can conceivably see requiring a reboot, is it that PK doesn't have the
concept of X logout vs reboot? Is it a bug in the packaging or PK or is there
anything I can do/file to improve the situation?
Dec 15 09:07:21 Updated: glib2-2.22.3-1.fc12.x86_64
Dec 15 09:07:23 Updated: mysql-libs-5.1.40-1.fc12.x86_64
Dec 15 09:07:23 Updated: gpm-libs-1.20.6-9.fc12.x86_64
Dec 15 09:07:25 Updated: mysql-5.1.40-1.fc12.x86_64
Dec 15 09:07:28 Updated: PyQt4-4.6.2-5.fc12.x86_64
Dec 15 09:07:32 Updated: mysql-server-5.1.40-1.fc12.x86_64
Dec 15 09:07:34 Updated: gpm-1.20.6-9.fc12.x86_64
Dec 15 09:07:37 Updated: 1:tk-8.5.7-3.fc12.x86_64
Dec 15 09:07:37 Updated: mpfr-2.4.1-5.fc12.x86_64
Dec 15 09:07:39 Updated: foomatic-4.0.3-5.fc12.x86_64
Dec 15 09:07:40 Updated: mysql-embedded-5.1.40-1.fc12.x86_64
Dec 15 09:07:41 Updated: glib2-2.22.3-1.fc12.i686
Dec 15 09:07:45 Updated: gtk2-2.18.4-3.fc12.x86_64
Dec 15 09:07:58 Updated: 1:gdm-2.28.1-25.fc12.x86_64
Dec 15 09:07:58 Updated: ibus-libs-1.2.0.20091204-2.fc12.x86_64
Dec 15 09:07:59 Updated: imsettings-libs-0.107.4-4.fc12.x86_64
Dec 15 09:08:02 Updated: glib2-devel-2.22.3-1.fc12.x86_64
Dec 15 09:08:07 Updated: yelp-2.28.1-1.fc12.x86_64
Dec 15 09:08:10 Updated: f-spot-0.6.1.5-1.fc12.x86_64
Dec 15 09:08:13 Updated: 1:xscreensaver-base-5.10-4.fc12.x86_64
Dec 15 09:08:13 Updated: 1:xscreensaver-gl-base-5.10-4.fc12.x86_64
Dec 15 09:08:16 Updated: gtk2-devel-2.18.4-3.fc12.x86_64
Dec 15 09:08:25 Updated: totem-2.28.4-1.fc12.x86_64
Dec 15 09:08:25 Updated: totem-nautilus-2.28.4-1.fc12.x86_64
Dec 15 09:08:27 Updated: 1:xscreensaver-gl-extras-5.10-4.fc12.x86_64
Dec 15 09:08:29 Updated: 1:xscreensaver-extras-5.10-4.fc12.x86_64
Dec 15 09:08:34 Updated: imsettings-0.107.4-4.fc12.x86_64
Dec 15 09:08:34 Updated: 1:gdm-user-switch-applet-2.28.1-25.fc12.x86_64
Dec 15 09:08:35 Updated: 1:gdm-plugin-fingerprint-2.28.1-25.fc12.x86_64
Dec 15 09:08:35 Updated: totem-mozplugin-2.28.4-1.fc12.x86_64
Dec 15 09:08:37 Updated: python-reportlab-2.3-1.fc12.x86_64
Dec 15 09:08:38 Updated: jna-3.2.4-1.fc12.x86_64
Dec 15 09:08:38 Updated: memcached-1.4.4-1.fc12.x86_64
Dec 15 09:08:38 Updated: less-436-3.fc12.x86_64
Dec 15 09:08:40 Updated: cscope-15.6-6.fc12.x86_64
Dec 15 09:08:40 Updated: xorg-x11-drv-mouse-1.5.0-1.fc12.x86_64
Dec 15 09:08:40 Updated: f-spot-screensaver-0.6.1.5-1.fc12.x86_64
Dec 15 09:08:46 Updated: gtk2-devel-docs-2.18.4-3.fc12.x86_64
Dec 15 09:08:46 Updated: gpm-devel-1.20.6-9.fc12.x86_64
Dec 15 09:08:46 Updated: liveusb-creator-3.9-1.fc12.noarch
Dec 15 09:08:48 Updated: mysql-devel-5.1.40-1.fc12.x86_64
Dec 15 09:08:53 Updated: etoys-4.0.2339-1.fc12.noarch
Dec 15 09:08:59 Updated: ibus-1.2.0.20091204-2.fc12.x86_64
Dec 15 09:08:59 Updated: ibus-gtk-1.2.0.20091204-2.fc12.x86_64
Wouldn't it be sufficient to logout? Is it a bug?
Does gdm entirely restart when you logout? I don't believe so. I suspect
you get the same result by killing X then going back to that runlevel but
for many many many users a reboot is going to be less error-prone.
-sv
11 a.m.
On 12/15/2009 09:54 AM, Seth Vidal wrote:
Does gdm entirely restart when you logout? I don't believe so. I
suspect
you get the same result by killing X then going back to that runlevel
but for many many many users a reboot is going to be less error-prone.
Isn't there gdm-restart for that purpose? I don't really know, but I'm
just confused as to why a program that lets me login requires a reboot...
I *really* don't want to sound whiny or anything like that, or be one of
those that compare us to windows... but one of my favorite things from
years ago was that I only had to reboot with a new kernel. Now I feel
like I reboot every update. I mean, even the ibus stuff was stating I
needed a reboot. As far as I know that is used for alternative language
input, which I don't use, fair enough it doesn't know that. But what
about it needs a reboot?
I'm also curious why gdm is still running once I've logged in. I see the
user-switch stuff but I'm just wondering. I mean rebooting isn't the end
of the world but man it sure happens a lot now
11:08 a.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Dne 15.12.2009 18:00, Nathanael D. Noblet napsal(a):
On 12/15/2009 09:54 AM, Seth Vidal wrote:
> Does gdm entirely restart when you logout? I don't believe so. I suspect
> you get the same result by killing X then going back to that runlevel
> but for many many many users a reboot is going to be less error-prone.
Isn't there gdm-restart for that purpose? I don't really know, but I'm
just confused as to why a program that lets me login requires a reboot...
I *really* don't want to sound whiny or anything like that, or be one of
those that compare us to windows... but one of my favorite things from
years ago was that I only had to reboot with a new kernel. Now I feel
like I reboot every update. I mean, even the ibus stuff was stating I
needed a reboot. As far as I know that is used for alternative language
input, which I don't use, fair enough it doesn't know that. But what
about it needs a reboot?
I'm also curious why gdm is still running once I've logged in. I see the
user-switch stuff but I'm just wondering. I mean rebooting isn't the end
of the world but man it sure happens a lot now
+1
- --
Nikola Pajkovsky <npajkovs(a)redhat.com> .~.
Base Operating Systems Brno /V\
// \\
Jabber: nikis(a)isgeek.info /( )\
Mobile: +420 777 895 064 ^`~'^
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEcBAEBAgAGBQJLJ8J+AAoJED4+/Vgo+H2XtUsIAIzOIdfL5OVS92bLJvFe9Vny
RCGKnlnG2TkEztpNZPu0AyU8nNOxAmH5+w+uAmvEpD6jmh5F7QlrSVTJLSamL76G
fjWmeuz+4EreSSKwhemlbr7p6vB/CxPkIcivOVYK4OscI7LwJGAQTK74661QvoWr
9BXAP+Wd3KaNere3Ckg4iamSwBDLjSf2fVmVUJjGL4CYnX/mJVnWm7V7VnHy0Rgr
lvM9fBUiJAuhOEH5xYScdTdmHmcefDfia1PmlMPVmG7mHlrIxXRF04KhyIJlwirK
nEj6mXQk2m0M+KO6zBoV78ROgY4xYk/RmyYbT8icOeojH12pJgo1KZWQ+89iPkI=
=Qe52
-----END PGP SIGNATURE-----
11:38 a.m.
On Tue, 15 Dec 2009, Nathanael D. Noblet wrote:
On 12/15/2009 09:54 AM, Seth Vidal wrote:
> Does gdm entirely restart when you logout? I don't believe so. I suspect
> you get the same result by killing X then going back to that runlevel
> but for many many many users a reboot is going to be less error-prone.
Isn't there gdm-restart for that purpose? I don't really know, but I'm just
confused as to why a program that lets me login requires a reboot...
I *really* don't want to sound whiny or anything like that, or be one of
those that compare us to windows... but one of my favorite things from years
ago was that I only had to reboot with a new kernel. Now I feel like I reboot
every update. I mean, even the ibus stuff was stating I needed a reboot. As
far as I know that is used for alternative language input, which I don't use,
fair enough it doesn't know that. But what about it needs a reboot?
I'm also curious why gdm is still running once I've logged in. I see the
user-switch stuff but I'm just wondering. I mean rebooting isn't the end of
the world but man it sure happens a lot now
I don't have a good answer. You might want to ask on the
fedora-desktop-list and/or in a bug for that component. I was just trying
to explain the specific behavior you saw.
Now, having said that - how would you feel if the updater stopped you
before it ran and said "you're running an app I'm trying to update, please
close the app so I can update it". Would that be a pain or ok?
-sv
11:40 a.m.
2009/12/15 Seth Vidal <skvidal(a)fedoraproject.org>:
Now, having said that - how would you feel if the updater stopped you
before
it ran and said "you're running an app I'm trying to update, please close
the app so I can update it". Would that be a pain or ok?
That's exactly the PackageKit functionality I've added for packages
like firefox, that explode internally when they get updated. The trick
it to offer to close them down, and bring them back when done.
Richard.
11:42 a.m.
On Tue, 15 Dec 2009, Richard Hughes wrote:
2009/12/15 Seth Vidal <skvidal(a)fedoraproject.org>:
> Now, having said that - how would you feel if the updater stopped you before
> it ran and said "you're running an app I'm trying to update, please
close
> the app so I can update it". Would that be a pain or ok?
That's exactly the PackageKit functionality I've added for packages
like firefox, that explode internally when they get updated. The trick
it to offer to close them down, and bring them back when done.
this is what colin and I talked about at fudcon in toronto. I just added
some code to yum so it returns to you a list of all pkgs on the system
that own a file that is currently open/used in a running process.
should make that part of your lookup easier.
YumBase.rpmdb.return_running_packages()
-sv
7:47 p.m.
On 12/15/2009 12:42 PM, Seth Vidal wrote:
this is what colin and I talked about at fudcon in toronto. I just
added
some code to yum so it returns to you a list of all pkgs on the system
that own a file that is currently open/used in a running process.
should make that part of your lookup easier.
YumBase.rpmdb.return_running_packages()
This seems to be compounding the problem rather than try9ing to avoid
it all together.
As an alternate suggestion:
It may make a lot more sense, to instead install apps in a shadow
directory (named by version for example) with the usual spot being a
link to the right place. When installing a new one simply install it in
the shadow area and flip the link. The existing running app will see
home base in the same place and have zero problems.
The last part is a clean up phase which could be deferred to reboot
or perhaps something a little more clever.
firefox will work perfectly if done this way for example.
gene
8:24 p.m.
On 12/16/2009 03:51 AM, Richard Hughes wrote:
2009/12/16 Mail Lists <lists(a)sapience.com>:
> The last part is a clean up phase which could be deferred to reboot
> or perhaps something a little more clever.
The devil is in the detail :)
Richard.
Yes but how are:
(a) Kill app - install - restart
(b) install in separate are - on reboot
(app is now dead, flip link to new install - delete old)
Any different?
We have ways now of running things as the system is going down or
coming up ... don't we ?
11:43 a.m.
On Tue, Dec 15, 2009 at 5:40 PM, Richard Hughes <hughsient(a)gmail.com> wrote:
2009/12/15 Seth Vidal <skvidal(a)fedoraproject.org>:
> Now, having said that - how would you feel if the updater stopped you before
> it ran and said "you're running an app I'm trying to update, please
close
> the app so I can update it". Would that be a pain or ok?
That's exactly the PackageKit functionality I've added for packages
like firefox, that explode internally when they get updated. The trick
it to offer to close them down, and bring them back when done.
This exists? Can you point me to the code?
12:01 p.m.
-----Original Message-----
From: fedora-devel-list-bounces(a)redhat.com [mailto:fedora-devel-list-
bounces(a)redhat.com] On Behalf Of Colin Walters
Sent: Tuesday, December 15, 2009 11:44
To: Development discussions related to Fedora
Subject: Re: packages requiring me to reboot...
On Tue, Dec 15, 2009 at 5:40 PM, Richard Hughes <hughsient(a)gmail.com>
wrote:
> 2009/12/15 Seth Vidal <skvidal(a)fedoraproject.org>:
>> Now, having said that - how would you feel if the updater stopped you
before
>> it ran and said "you're running an app I'm trying to update,
please
close
>> the app so I can update it". Would that be a pain or ok?
>
> That's exactly the PackageKit functionality I've added for packages
> like firefox, that explode internally when they get updated. The trick
> it to offer to close them down, and bring them back when done.
>
This exists? Can you point me to the code?
--
I am not sure what the argument is, factually there are packages that have
files open, locks and etc. that need to be shutdown to update, if they are
running and you replace the executable, doesn't mean that the memory image
is replaced. It is quicker and simpler to just reboot, also the list of the
packages that cause you to reboot is probably longer than the ones that are
flagged. I think that a reboot should be made whether necessary or not,
clears up a lot of grief.
fedora-devel-list mailing list
fedora-devel-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
11:19 p.m.
On Tue, 2009-12-15 at 12:01 -0600, Otto Haliburton wrote:
I am not sure what the argument is, factually there are packages that
have
files open, locks and etc. that need to be shutdown to update, if they are
running and you replace the executable, doesn't mean that the memory image
is replaced. It is quicker and simpler to just reboot, also the list of the
packages that cause you to reboot is probably longer than the ones that are
flagged. I think that a reboot should be made whether necessary or not,
clears up a lot of grief.
Wow. That's twice today that the suggestion of forcing the user to
reboot whether they like it or not has been made. What a long way we've
come since the early days :) Personally, I think a dialog box is pretty
simple for most users (no forcing, just persuasion), and allows those
who know what they're doing to ignore it. And obviously reducing the
number of times you ask to just absolute necessity is a win-win.
But please, no more forced reboot suggestions. This isn't MSFT.
Jon.
2:50 a.m.
2009/12/15 Colin Walters <walters(a)verbum.org>:
This exists? Can you point me to the code?
I only finished this just this morning.
It's just been pushed to git master. You want to see this commit
http://cgit.freedesktop.org/packagekit/commit/?id=66d3fc26054abd528ee1801...
for the juicy config bits.
The UI isn't very pretty at the moment (it just fails with an update
error) but I'll work on something a little bit more user friendly.
Richard.
9:51 a.m.
On Wed, 2009-12-16 at 08:50 +0000, Richard Hughes wrote:
2009/12/15 Colin Walters <walters(a)verbum.org>:
> This exists? Can you point me to the code?
I only finished this just this morning.
It's just been pushed to git master. You want to see this commit
http://cgit.freedesktop.org/packagekit/commit/?id=66d3fc26054abd528ee1801...
for the juicy config bits.
Looking at 3cb32ad40af3a38456e09baf1b29b046d82c587e, AIUI the commit
with the code bits in it, I'm pretty sure you are now requiring
filelists to be downloaded for all updates.
The UI isn't very pretty at the moment (it just fails with an
update
error) but I'll work on something a little bit more user friendly.
How do you plan on restarting firefox? Or you just planning to kill()
and get the user to restart?
--
James Antill - james(a)fedoraproject.org
http://yum.baseurl.org/wiki/releases
http://yum.baseurl.org/wiki/whatsnew/3.2.25
http://yum.baseurl.org/wiki/YumMultipleMachineCaching
11:05 a.m.
On Thu, Dec 17, 2009 at 10:51 AM, James Antill <james(a)fedoraproject.org> wrote:
How do you plan on restarting firefox? Or you just planning to kill()
and get the user to restart?
Trying to send a close button event to the app's windows is probably
our best short-term approach; slightly longer term, we could have apps
expose a standard interface for Quit (e.g. over dbus).
Knowing how to restart is trickier, (though we could use the
window-to-app mapping system we need for gnome 3 anyways) but it's
also very simple in this case if we require that packages contain at
most one .desktop file.
2:21 p.m.
On Thu, Dec 17, 2009 at 12:05:15PM -0500, Colin Walters wrote:
On Thu, Dec 17, 2009 at 10:51 AM, James Antill
<james(a)fedoraproject.org> wrote:
>
> How do you plan on restarting firefox? Or you just planning to kill()
> and get the user to restart?
Trying to send a close button event to the app's windows is probably
our best short-term approach;
That would fail pretty badly in the case of Firefox with multiple
windows open. Firefox's session support can restore multiple tabs to
multiple windows if you quit it and restart, but if you have two windows
open, close one, then close the second causing the browser to quit, then
on restart the session that is restored will only contain the tabs from
the second window. The logic goes that the first window was no longer
part of the session when the app quit. Simply killing the process is
actually less disruptive.
Ewan
2:52 p.m.
On Thu, Dec 17, 2009 at 3:21 PM, Ewan Mac Mahon <ewan(a)macmahon.me.uk> wrote:
That would fail pretty badly in the case of Firefox with multiple
windows open.
True; there's nothing stopping us from adding something
Firefox-specific as a short term measure, since how it does session
saving is fairly unique right now (ideally we add a nice API for this
to GTK+ which then has to be mirrored in XUL). Killing the process
though has the downside of triggering the "Well, that was
embarassing..." which is arguably a Firefox bug though.
Anyways, "the perfect is the enemy of the good", etc.
10:42 a.m.
tor 2009-12-17 klockan 20:21 +0000 skrev Ewan Mac Mahon:
Simply killing the process is
actually less disruptive.
Yes. Please do not send close events to my Firefox windows.
Do keep in mind that Firefox is pretty much broken and might not even be
able to show a "Do you want to restart?" dialog by the time the RPM
transaction is rolling, so if it's going to involve user interaction,
that will have to happen in the PackageKit GUI.
Another option is to make Firefox notice internally how broken it is and
try to restart itself. It would need to somehow wait until the
transaction is done, though.
/abo
12:22 p.m.
On Thu, 2009-12-17 at 10:51 -0500, James Antill wrote:
> The UI isn't very pretty at the moment (it just fails with
an update
> error) but I'll work on something a little bit more user friendly.
How do you plan on restarting firefox? Or you just planning to kill()
and get the user to restart?
If we're just talking about Firefox (i.e. not the general case), then it
has its own 'restart Firefox' hook you might be able to access. It's
used, for instance, when you enable or disable an extension. I'm not
sure if you can poke it from an external app easily, though.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net
11:40 a.m.
On Tue, 2009-12-15 at 10:00 -0700, Nathanael D. Noblet wrote:
On 12/15/2009 09:54 AM, Seth Vidal wrote:
> Does gdm entirely restart when you logout? I don't believe so. I suspect
> you get the same result by killing X then going back to that runlevel
> but for many many many users a reboot is going to be less error-prone.
Isn't there gdm-restart for that purpose? I don't really know, but I'm
just confused as to why a program that lets me login requires a reboot...
I *really* don't want to sound whiny or anything like that, or be one of
those that compare us to windows... but one of my favorite things from
years ago was that I only had to reboot with a new kernel. Now I feel
like I reboot every update. I mean, even the ibus stuff was stating I
needed a reboot. As far as I know that is used for alternative language
input, which I don't use, fair enough it doesn't know that. But what
about it needs a reboot?
Because often you cannot tell if a reboot is required or not. Consider a
shared library which gets updated. If the old version of the library
contains a security bug and you have already say ten apps running using
that library, then after the update of the library the apps will still
use the old buggy library. Only after closing and restarting the
particular apps will help. From a novice point of view you cannot expect
that he/she knows what app to close, so a restart is probably the
easiest and safest way.
Another problem may raise up if you want to update daemons. During the
update the daemon shouldn't be restarted because it might be a critical
service I'm using right now. Therefore, having a convenient way for
novices is just to say "please restart". For all others, just check what
gets updated and then decide on your own if you really need to restart
the whole system or only some apps/services and then klick on "hide this
icon" ;-)
Coming back to your fist post:
Wouldn't it be sufficient to logout?
In some cases I would say so, but not in all.
cheers,
Stefan
12:29 p.m.
Nathanael D. Noblet (nathanael(a)gnat.ca) said:
I'm also curious why gdm is still running once I've logged
in.
When you start a display manager, you start an X server; the display
manager then draws on this. Then, when you log in, you have to
stat an user session, as the authenticated user (which has a connection
to the X server, so it can know when it goes away.)
You also have to tell the init daemon which process it's supposed to
be tracking, so it can respawn it when it exits.
Having that process be the gdm daemon (which forks and execs both
the X server and the user session) is arguably a lot simpler than
trying to architect it such that the daemon goes away entirely and
init then ends up tracking either the X serve or the user session.
Bill
4:15 p.m.
On Tue, 15 Dec 2009 13:29:25 -0500
Bill Nottingham <notting(a)redhat.com> wrote:
Nathanael D. Noblet (nathanael(a)gnat.ca) said:
> I'm also curious why gdm is still running once I've logged in.
When you start a display manager, you start an X server; the display
manager then draws on this. Then, when you log in, you have to
stat an user session, as the authenticated user (which has a connection
to the X server, so it can know when it goes away.)
You also have to tell the init daemon which process it's supposed to
be tracking, so it can respawn it when it exits.
Having that process be the gdm daemon (which forks and execs both
the X server and the user session) is arguably a lot simpler than
trying to architect it such that the daemon goes away entirely and
init then ends up tracking either the X serve or the user session.
Bill
Isn't GDM just doing a wait(2) on the user-session?
Clark
11:41 a.m.
On 15/12/09 16:50, Nathanael D. Noblet wrote:
Hello,
I feel like there are an increasing number of packages requiring a
system reboot. I'm wondering why. The following updates were installed
today, and required a full system reboot. I can't seem to find any
package in the list that I can conceivably see requiring a reboot, is it
that PK doesn't have the concept of X logout vs reboot? Is it a bug in
the packaging or PK or is there anything I can do/file to improve the
situation?
Personally,
I just update, ignore the "restart".
Shut-down going to bed.
Or keep the updates,
till bedtime.
--
Regards,
Frank Murphy
UTF_8 Encoded.
5:24 p.m.
Am 2009-12-15 17:50, schrieb Nathanael D. Noblet:
Hello,
I feel like there are an increasing number of packages requiring a
system reboot. I'm wondering why. The following updates were installed
today, and required a full system reboot. I can't seem to find any
package in the list that I can conceivably see requiring a reboot, is it
that PK doesn't have the concept of X logout vs reboot? Is it a bug in
the packaging or PK or is there anything I can do/file to improve the
situation?
Dec 15 09:07:21 Updated: glib2-2.22.3-1.fc12.x86_64
Dec 15 09:07:23 Updated: mysql-libs-5.1.40-1.fc12.x86_64
Dec 15 09:07:23 Updated: gpm-libs-1.20.6-9.fc12.x86_64
Dec 15 09:07:25 Updated: mysql-5.1.40-1.fc12.x86_64
Dec 15 09:07:28 Updated: PyQt4-4.6.2-5.fc12.x86_64
Dec 15 09:07:32 Updated: mysql-server-5.1.40-1.fc12.x86_64
Dec 15 09:07:34 Updated: gpm-1.20.6-9.fc12.x86_64
Dec 15 09:07:37 Updated: 1:tk-8.5.7-3.fc12.x86_64
Dec 15 09:07:37 Updated: mpfr-2.4.1-5.fc12.x86_64
Dec 15 09:07:39 Updated: foomatic-4.0.3-5.fc12.x86_64
Dec 15 09:07:40 Updated: mysql-embedded-5.1.40-1.fc12.x86_64
Dec 15 09:07:41 Updated: glib2-2.22.3-1.fc12.i686
Dec 15 09:07:45 Updated: gtk2-2.18.4-3.fc12.x86_64
Dec 15 09:07:58 Updated: 1:gdm-2.28.1-25.fc12.x86_64
Dec 15 09:07:58 Updated: ibus-libs-1.2.0.20091204-2.fc12.x86_64
Dec 15 09:07:59 Updated: imsettings-libs-0.107.4-4.fc12.x86_64
Dec 15 09:08:02 Updated: glib2-devel-2.22.3-1.fc12.x86_64
Dec 15 09:08:07 Updated: yelp-2.28.1-1.fc12.x86_64
Dec 15 09:08:10 Updated: f-spot-0.6.1.5-1.fc12.x86_64
Dec 15 09:08:13 Updated: 1:xscreensaver-base-5.10-4.fc12.x86_64
Dec 15 09:08:13 Updated: 1:xscreensaver-gl-base-5.10-4.fc12.x86_64
Dec 15 09:08:16 Updated: gtk2-devel-2.18.4-3.fc12.x86_64
Dec 15 09:08:25 Updated: totem-2.28.4-1.fc12.x86_64
Dec 15 09:08:25 Updated: totem-nautilus-2.28.4-1.fc12.x86_64
Dec 15 09:08:27 Updated: 1:xscreensaver-gl-extras-5.10-4.fc12.x86_64
Dec 15 09:08:29 Updated: 1:xscreensaver-extras-5.10-4.fc12.x86_64
Dec 15 09:08:34 Updated: imsettings-0.107.4-4.fc12.x86_64
Dec 15 09:08:34 Updated: 1:gdm-user-switch-applet-2.28.1-25.fc12.x86_64
Dec 15 09:08:35 Updated: 1:gdm-plugin-fingerprint-2.28.1-25.fc12.x86_64
Dec 15 09:08:35 Updated: totem-mozplugin-2.28.4-1.fc12.x86_64
Dec 15 09:08:37 Updated: python-reportlab-2.3-1.fc12.x86_64
Dec 15 09:08:38 Updated: jna-3.2.4-1.fc12.x86_64
Dec 15 09:08:38 Updated: memcached-1.4.4-1.fc12.x86_64
Dec 15 09:08:38 Updated: less-436-3.fc12.x86_64
Dec 15 09:08:40 Updated: cscope-15.6-6.fc12.x86_64
Dec 15 09:08:40 Updated: xorg-x11-drv-mouse-1.5.0-1.fc12.x86_64
Dec 15 09:08:40 Updated: f-spot-screensaver-0.6.1.5-1.fc12.x86_64
Dec 15 09:08:46 Updated: gtk2-devel-docs-2.18.4-3.fc12.x86_64
Dec 15 09:08:46 Updated: gpm-devel-1.20.6-9.fc12.x86_64
Dec 15 09:08:46 Updated: liveusb-creator-3.9-1.fc12.noarch
Dec 15 09:08:48 Updated: mysql-devel-5.1.40-1.fc12.x86_64
Dec 15 09:08:53 Updated: etoys-4.0.2339-1.fc12.noarch
Dec 15 09:08:59 Updated: ibus-1.2.0.20091204-2.fc12.x86_64
Dec 15 09:08:59 Updated: ibus-gtk-1.2.0.20091204-2.fc12.x86_64
Wouldn't it be sufficient to logout? Is it a bug?
I'd like for the icon that reboots to be more difficult to click on. Or
to confirm. Or anything.
10:18 a.m.
So again today, I see some updates two of which require a full system
reboot.
nfs-utils and ibus-rawcode. My system seriously needs to be shut down
for those to be properly updated? This is what I don't get. nfs-utils
never got a system reboot before, it doesn't get one on RHEL/Centos
boxes... What requires a reboot here? Again, I don't want the tone of
this email to come off as anger, rude or whatever, mainly I'm wondering
why so many packages require a reboot, why isn't nfs-utils just
restarting any services it has or that depend on it if needed? Is that
not reliable?
10:43 a.m.
On Wed, 16 Dec 2009, Nathanael D. Noblet wrote:
So again today, I see some updates two of which require a full system
reboot.
nfs-utils and ibus-rawcode. My system seriously needs to be shut down for
those to be properly updated? This is what I don't get. nfs-utils never got a
system reboot before, it doesn't get one on RHEL/Centos boxes... What
requires a reboot here? Again, I don't want the tone of this email to come
off as anger, rude or whatever, mainly I'm wondering why so many packages
require a reboot, why isn't nfs-utils just restarting any services it has or
that depend on it if needed? Is that not reliable?
you're an experienced user? You're comfortable knowing what does and what
does not require a reboot? Then why are you using PK?
Disable pk and do the updates directly via yum.
Bam - no more requests to reboot.
-sv
10:48 a.m.
On 12/16/2009 11:43 AM, Seth Vidal wrote:
you're an experienced user? You're comfortable knowing what
does and
what does not require a reboot? Then why are you using PK?
Disable pk and do the updates directly via yum.
Bam - no more requests to reboot.
I get what you're saying, and it's kindof a fair point, but there's also
some utility to having the system automatically, proactively notify you
of updates.
--
Peter
For some reason it has always seemed to me that the term software
engineering contains some very optimistic assumptions about the
nature of reality.
10:51 a.m.
On Wed, 16 Dec 2009, Peter Jones wrote:
On 12/16/2009 11:43 AM, Seth Vidal wrote:
> you're an experienced user? You're comfortable knowing what does and
> what does not require a reboot? Then why are you using PK?
>
> Disable pk and do the updates directly via yum.
>
> Bam - no more requests to reboot.
I get what you're saying, and it's kindof a fair point, but there's also
some utility to having the system automatically, proactively notify you
of updates.
And you can do that. Just don't have pk DO the update.
There are lots of ways to get notifications of updates not using PK in the
system.
And again, we're not talking about for the default everyday user.
we're talking about the experienced user who is comfortable knowing what
does and does not need a reboot.
All I'm saying is - we've not taken away any option, the experienced user
can do what they want.
-sv
10:54 a.m.
Once upon a time, Seth Vidal <skvidal(a)fedoraproject.org> said:
we're talking about the experienced user who is comfortable
knowing what
does and does not need a reboot.
It seems though that there is a problem with how the "needs a reboot"
option is set (and if that is the case, it should be addressed). For
example, in the nfs-utils case, what happened to having the %post
scriptlet do "service foo condrestart"? Is it impossible to restart the
daemons?
--
Chris Adams <cmadams(a)hiwaay.net>
Systems and Network Administrator - HiWAAY Internet Services
I don't speak for anybody but myself - that's enough trouble.
10:56 a.m.
On Wed, 16 Dec 2009, Chris Adams wrote:
Once upon a time, Seth Vidal <skvidal(a)fedoraproject.org> said:
> we're talking about the experienced user who is comfortable knowing what
> does and does not need a reboot.
It seems though that there is a problem with how the "needs a reboot"
option is set (and if that is the case, it should be addressed). For
example, in the nfs-utils case, what happened to having the %post
scriptlet do "service foo condrestart"? Is it impossible to restart the
daemons?
well nfs restarts have never been likely to work if something is mounted
iirc.
but I'm not the nfs expert here - maybe steve dickson can better answer
this.
-sv
11:08 a.m.
On 12/16/2009 09:51 AM, Seth Vidal wrote:
On Wed, 16 Dec 2009, Peter Jones wrote:
> On 12/16/2009 11:43 AM, Seth Vidal wrote:
>> you're an experienced user? You're comfortable knowing what does and
>> what does not require a reboot? Then why are you using PK?
>>
>> Disable pk and do the updates directly via yum.
>>
>> Bam - no more requests to reboot.
>
> I get what you're saying, and it's kindof a fair point, but there's also
> some utility to having the system automatically, proactively notify you
> of updates.
And you can do that. Just don't have pk DO the update.
There are lots of ways to get notifications of updates not using PK in
the system.
And again, we're not talking about for the default everyday user.
we're talking about the experienced user who is comfortable knowing what
does and does not need a reboot.
All I'm saying is - we've not taken away any option, the experienced
user can do what they want.
yeah, I totally get what you mean. I just feel like there are more and
more reboot requests because that is easier. Obviously I know when/if I
need to reboot based on what I'm running. However I'm questioning the
number of packages requesting a reboot I guess.
Maybe this is a feature that needs to be addressed in the rpm layer or
something so that upgrades can have multiple effects with regards to
needing a reboot. I'm not sure how PK gets the request to reboot from a
package, but I'm wondering about it. For example, why aren't some of the
packages simply a 'log out of X', or 'restart app', or ??? PK could
provide that information. However, as it stands, if firefox is updated,
I would (under the current way this seems to work) fully expect it to
ask for my system to reboot, instead of closing FF and starting it
again. I'm not sure if it does or not, but it seems like a package
basically has complex upgrade issues, so we reboot. Are there other tags
packages can have other than reboot? Should there be? etc etc..
I am an advanced user, and manage a handful of servers and workstations,
so yes I don't have to reboot. I'm just wondering about the reboot
'feature' usage patterns I'm seeing.
--
Nathanael
11:11 a.m.
On Wed, 16 Dec 2009, Nathanael D. Noblet wrote:
Maybe this is a feature that needs to be addressed in the rpm layer
or
something so that upgrades can have multiple effects with regards to needing
a reboot. I'm not sure how PK gets the request to reboot from a package, but
I'm wondering about it.
It doesn't get it from the pkg. It uses the updateinfo.xml metadata that
is generated by our update processing system that is called 'bodhi'.
You can see this data using the yum-security plugin.
seems like a package basically has complex upgrade issues, so we
reboot. Are
there other tags packages can have other than reboot? Should there be? etc
etc..
No.
I am an advanced user, and manage a handful of servers and
workstations, so
yes I don't have to reboot. I'm just wondering about the reboot 'feature'
usage patterns I'm seeing.
And again. PK is not designed for you. The 'reboot often' solution is not
FOR you.
I said this earlier on another subject but you shouldn't be shocked that
camels are slow swimmers.
-sv
11:16 a.m.
On 12/16/2009 10:11 AM, Seth Vidal wrote:
On Wed, 16 Dec 2009, Nathanael D. Noblet wrote:
> Maybe this is a feature that needs to be addressed in the rpm layer or
> something so that upgrades can have multiple effects with regards to
> needing a reboot. I'm not sure how PK gets the request to reboot from
> a package, but I'm wondering about it.
It doesn't get it from the pkg. It uses the updateinfo.xml metadata that
is generated by our update processing system that is called 'bodhi'.
You can see this data using the yum-security plugin.
Cool.
> seems like a package basically has complex upgrade issues, so we
> reboot. Are there other tags packages can have other than reboot?
> Should there be? etc etc..
No.
The reason for this is that PKs target audience is not someone like me,
and as such no need to provide different messages per package?
> I am an advanced user, and manage a handful of servers and
> workstations, so yes I don't have to reboot. I'm just wondering about
> the reboot 'feature' usage patterns I'm seeing.
And again. PK is not designed for you. The 'reboot often' solution is
not FOR you.
I said this earlier on another subject but you shouldn't be shocked that
camels are slow swimmers.
So basically, PK is designed for the non-experienced users, as such
everything it does is dumbed down, and experienced users should just
ignore it, using other tools to keep their system up to date.
So one last question then, in the case of nfs-utils, (ignoring for now
any nfs specific restart/condrestart issues). The packaging guidlines
will continue to require that a post update script does what is sensible
for an update, and not just depend on the admin rebooting their server?
--
Nathanael
11:28 a.m.
On Wed, 16 Dec 2009, Nathanael D. Noblet wrote:
>> seems like a package basically has complex upgrade issues, so
we
>> reboot. Are there other tags packages can have other than reboot?
>> Should there be? etc etc..
>
> No.
The reason for this is that PKs target audience is not someone like me, and
as such no need to provide different messages per package?
No, the reason for this is there is not more to go on, yet. I would love
to require more detailed info on the update including if it is an
important/trivial/security/packaging/upstream-update or what not fix.
Hands are needed to help advance this. Care to lend one?
>
> And again. PK is not designed for you. The 'reboot often' solution is
> not FOR you.
>
> I said this earlier on another subject but you shouldn't be shocked that
> camels are slow swimmers.
So basically, PK is designed for the non-experienced users, as such
everything it does is dumbed down, and experienced users should just ignore
it, using other tools to keep their system up to date.
If what the experienced user wants to do is not something that PK can do
or can be configured to do then, yes, disable it and move along.
Hell, same thing is true of yum. If you really know what you're doing and
yum is in your way then stop using it.
So one last question then, in the case of nfs-utils, (ignoring for
now any
nfs specific restart/condrestart issues). The packaging guidlines will
continue to require that a post update script does what is sensible for an
update, and not just depend on the admin rebooting their server?
the post scripts do what is sensible, on many occasions restarting the
daemon will not ensure that the new sw is in use and in other occasions
there is no graceful way to restart.
so your options are:
1. don't restart but ask the user to
2. restart and drop whatever connections are active.
neither are great.
-sv
11:36 a.m.
On 12/16/2009 10:28 AM, Seth Vidal wrote:
On Wed, 16 Dec 2009, Nathanael D. Noblet wrote:
>>> seems like a package basically has complex upgrade issues, so we
>>> reboot. Are there other tags packages can have other than reboot?
>>> Should there be? etc etc..
>>
>> No.
>
> The reason for this is that PKs target audience is not someone like
> me, and as such no need to provide different messages per package?
No, the reason for this is there is not more to go on, yet. I would love
to require more detailed info on the update including if it is an
important/trivial/security/packaging/upstream-update or what not fix.
Hands are needed to help advance this. Care to lend one?
Yes. I'm attempting to become more involved. I've submitted my first
package, and am going through the review process. That doesn't help in
this particular case, but I am not complaining for the sake of
complaining. I want to help. I fully realize what I use daily for work
is the result of many people like you who build this stuff. Thus my
desire to become part of it.
What can I do here?
the post scripts do what is sensible, on many occasions restarting
the
daemon will not ensure that the new sw is in use and in other occasions
there is no graceful way to restart.
so your options are:
1. don't restart but ask the user to
2. restart and drop whatever connections are active.
neither are great.
For sure.
11:38 a.m.
On Wed, 16 Dec 2009, Nathanael D. Noblet wrote:
>
>
> Hands are needed to help advance this. Care to lend one?
Yes. I'm attempting to become more involved. I've submitted my first package,
and am going through the review process. That doesn't help in this particular
case, but I am not complaining for the sake of complaining. I want to help. I
fully realize what I use daily for work is the result of many people like you
who build this stuff. Thus my desire to become part of it.
What can I do here?
How much python do you know? We need some time spent on the updateinfo.xml
and what information we provide there and tying this in with what info is
required from packagers submitting updates to their pkgs.
Another good angle to approach is to talk to the folks in fedora-qa and
see where they can use a hand.
-sv
11:41 a.m.
On 12/16/2009 10:38 AM, Seth Vidal wrote:
On Wed, 16 Dec 2009, Nathanael D. Noblet wrote:
>>
>>
>> Hands are needed to help advance this. Care to lend one?
>
> Yes. I'm attempting to become more involved. I've submitted my first
> package, and am going through the review process. That doesn't help in
> this particular case, but I am not complaining for the sake of
> complaining. I want to help. I fully realize what I use daily for work
> is the result of many people like you who build this stuff. Thus my
> desire to become part of it.
>
> What can I do here?
How much python do you know? We need some time spent on the
updateinfo.xml and what information we provide there and tying this in
with what info is required from packagers submitting updates to their pkgs.
Unfortunately very very little. I can program in C/C++ and PHP. I've
used python years ago. However I can learn if I have to.
Another good angle to approach is to talk to the folks in fedora-qa
and
see where they can use a hand.
Will do.
4:33 a.m.
2009/12/16 Nathanael D. Noblet <nathanael(a)gnat.ca>:
So basically, PK is designed for the non-experienced users, as such
everything it does is dumbed down, and experienced users should just ignore
it, using other tools to keep their system up to date.
See http://www.packagekit.org/pk-profiles.html
Richard.
11:16 a.m.
Am 2009-12-16 17:51, schrieb Seth Vidal:
On Wed, 16 Dec 2009, Peter Jones wrote:
> On 12/16/2009 11:43 AM, Seth Vidal wrote:
>> you're an experienced user? You're comfortable knowing what does and
>> what does not require a reboot? Then why are you using PK?
>>
>> Disable pk and do the updates directly via yum.
>>
>> Bam - no more requests to reboot.
>
> I get what you're saying, and it's kindof a fair point, but there's also
> some utility to having the system automatically, proactively notify you
> of updates.
And you can do that. Just don't have pk DO the update.
There are lots of ways to get notifications of updates not using PK in
the system.
And again, we're not talking about for the default everyday user.
we're talking about the experienced user who is comfortable knowing what
does and does not need a reboot.
All I'm saying is - we've not taken away any option, the experienced
user can do what they want.
-sv
True, but the default should be sensible.
11:21 a.m.
On Wed, 16 Dec 2009, nodata wrote:
>
> we're talking about the experienced user who is comfortable knowing what
> does and does not need a reboot.
>
> All I'm saying is - we've not taken away any option, the experienced
> user can do what they want.
>
> -sv
>
True, but the default should be sensible.
And the default is sensible for the inexperienced user:
Don't try to explain to the user how to restart the apps individually,
tell them to bounce the box and it will be the right version when it comes
back.
-sv
11:29 a.m.
Am 2009-12-16 18:21, schrieb Seth Vidal:
On Wed, 16 Dec 2009, nodata wrote:
>>
>> we're talking about the experienced user who is comfortable knowing what
>> does and does not need a reboot.
>>
>> All I'm saying is - we've not taken away any option, the experienced
>> user can do what they want.
>>
>> -sv
>>
>
> True, but the default should be sensible.
And the default is sensible for the inexperienced user:
Don't try to explain to the user how to restart the apps individually,
tell them to bounce the box and it will be the right version when it
comes back.
-sv
On the other hand I think requiring more reboots than Windows is a bad
thing...
11:34 a.m.
On Wed, 16 Dec 2009, nodata wrote:
Am 2009-12-16 18:21, schrieb Seth Vidal:
>
>
> On Wed, 16 Dec 2009, nodata wrote:
>
>>>
>>> we're talking about the experienced user who is comfortable knowing what
>>> does and does not need a reboot.
>>>
>>> All I'm saying is - we've not taken away any option, the experienced
>>> user can do what they want.
>>>
>>> -sv
>>>
>>
>> True, but the default should be sensible.
>
> And the default is sensible for the inexperienced user:
>
> Don't try to explain to the user how to restart the apps individually,
> tell them to bounce the box and it will be the right version when it
> comes back.
>
> -sv
>
On the other hand I think requiring more reboots than Windows is a bad
thing...
Then I can think of a couple of solutions to this problem:
1. Have fewer update pushes per release - this is something I'm actively
advocating and I think is possible
2. Match up more updates to a specific running app so we can see if the
reboot is really necessary at all. - something else I've wrriten some code
in support of.
How would you like to help in these goals?
-sv
10:05 p.m.
On 12/16/2009 06:34 PM, Seth Vidal wrote:
On Wed, 16 Dec 2009, nodata wrote:
> Am 2009-12-16 18:21, schrieb Seth Vidal:
>>
>>
>> On Wed, 16 Dec 2009, nodata wrote:
>>
>>>>
>>>> we're talking about the experienced user who is comfortable knowing
>>>> what
>>>> does and does not need a reboot.
>>>>
>>>> All I'm saying is - we've not taken away any option, the
experienced
>>>> user can do what they want.
>>>>
>>>> -sv
>>>>
>>>
>>> True, but the default should be sensible.
>>
>> And the default is sensible for the inexperienced user:
>>
>> Don't try to explain to the user how to restart the apps individually,
>> tell them to bounce the box and it will be the right version when it
>> comes back.
>>
>> -sv
>>
>
> On the other hand I think requiring more reboots than Windows is a bad
> thing...
Then I can think of a couple of solutions to this problem:
1. Have fewer update pushes per release - this is something I'm actively
advocating and I think is possible
Depends on what you actually have in mind.
Simply letting update pile up would seem a silly idea to me, it
contradicts Fedora's goal and removes what makes Fedora "attractive".
Letting pile up "updates, which require a reboot, but are not addressig
real bugs", could be applicable.
2. Match up more updates to a specific running app so we can see if
the
reboot is really necessary at all. - something else I've wrriten some
code in support of.
Yes, this would be helpful - But only in case of non-bugfix
updates.
Bug-fix updates should be pushed ASAP, IMO.
3. Having better tools to avoid reboots.
(Consider daemons, servers).
4. Maintainers to be more careful/reluctant/conservative, when
considering to update packages, which require a reboot.
Ralf
2:43 p.m.
----- Original Message -----
From: "nodata" <lsof(a)nodata.co.uk>
To: "Development discussions related to Fedora"
<fedora-devel-list(a)redhat.com>
Sent: Wednesday, December 16, 2009 11:29 AM
Subject: Re: packages requiring me to reboot...
Am 2009-12-16 18:21, schrieb Seth Vidal:
>
>
> On Wed, 16 Dec 2009, nodata wrote:
>
>>>
>>> we're talking about the experienced user who is comfortable knowing
>>> what
>>> does and does not need a reboot.
>>>
>>> All I'm saying is - we've not taken away any option, the experienced
>>> user can do what they want.
>>>
>>> -sv
>>>
>>
>> True, but the default should be sensible.
>
> And the default is sensible for the inexperienced user:
>
> Don't try to explain to the user how to restart the apps individually,
> tell them to bounce the box and it will be the right version when it
> comes back.
>
> -sv
>
On the other hand I think requiring more reboots than Windows is a bad
thing...
windows update will automatically reboot your system when it automatically
updates it
windows tried the optional stuff but now almost every case it requires a
restart.
--
fedora-devel-list mailing list
fedora-devel-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
3:30 p.m.
On 12/16/2009 03:43 PM, Otto Haliburton wrote:
windows update will automatically reboot your system when it
automatically
updates it windows tried the optional stuff but now almost every case it requires a
restart.
Even when it asks, it does that with a modal focus-grabbing dialog
window. I was once working on an important server, and was about to
click OK on one of my dialog boxes, when a windows update confirmation
popped up conveniently located so that my click went to IT, rebooting
the server. The queen was not amused.
HOWEVER, just because Windows suck doesn't mean that Linux should suck
too. My own preference would be a more discriminating dialog
that offers three possibilities: 'do nothing', 'bounce the
service/application' and 'reboot'.
3:38 p.m.
On Wed, 2009-12-16 at 16:30 -0500, Przemek Klosowski wrote:
HOWEVER, just because Windows suck doesn't mean that Linux should
suck
too. My own preference would be a more discriminating dialog
that offers three possibilities: 'do nothing', 'bounce the
service/application' and 'reboot'.
Yup, +1
Jon.
6:35 a.m.
New subject: packages requiring me to reboot - semi-off -topic response
On 12/17/2009 08:38 AM, Jon Masters wrote:
On Wed, 2009-12-16 at 16:30 -0500, Przemek Klosowski wrote:
> too. My own preference would be a more discriminating dialog
> that offers three possibilities: 'do nothing', 'bounce the
> service/application' and 'reboot'.
Yup, +1
Bounce the application would be really cool for eg firefox, thunderbird,
that continue to work, but some things sort of don't (like new popups
dont open and so forth). In doing so the application would need to
return to the identical locations, arrangements that it was bounced from.
And really, the above shouldn't be difficult for a computer to do:
- have window, pos,size, doc open, caret position and so on, even those
pesky web forms with unsubmitted data and so on. Make it part of the
freedesktop standards - an app forcibly closed or ?disappears shall
reopen as if nothing changed ! (although firefox multi-tab and working
out which tab not to reopen after a crash is a good game).
Make open documents files etc, be always stored immediately on change
with snapshots at each save to disk, few minutes of operation etc. Make
it like paper, but actually better, you know: what is written on paper
doesn't mysteriously disappear when an update or a power failure, or app
crash occurs, nor does it disappear from where you left it.
I think the above scenario for user apps would seem to be a reasonable
goal. On the other hand, services are not so clear cut. If I'm an
external user logged into the web service, filling a form, I don't
expect even momentary downtime to cause me to lose information I'm
entering, or corrupt a file I have open / editing on network share
(though see the works better than paper description).
In the first example, would it make sense for the web server to start
all new processes using the new updated code, while existing users stay
connected to the existing instance, until these timeout/logout and no
user's connection is using the old code.
When a reboot is really needed, a PK dialog could inform of the need,
and ask when that could occur (do at 4:30 am tueday morning)
ps: did gnome desktop regain the feature of autologin and rerun apps on
desktop capability of pre f-10 ?
4:35 p.m.
-----Original Message-----
From: fedora-devel-list-bounces(a)redhat.com [mailto:fedora-devel-list-
bounces(a)redhat.com] On Behalf Of Przemek Klosowski
Sent: Wednesday, December 16, 2009 15:30
To: fedora-devel-list(a)redhat.com
Subject: Re: packages requiring me to reboot...
On 12/16/2009 03:43 PM, Otto Haliburton wrote:
> windows update will automatically reboot your system when it
automatically
> updates it windows tried the optional stuff but now almost every case it
requires a
> restart.
Even when it asks, it does that with a modal focus-grabbing dialog
window. I was once working on an important server, and was about to
click OK on one of my dialog boxes, when a windows update confirmation
popped up conveniently located so that my click went to IT, rebooting
the server. The queen was not amused.
HOWEVER, just because Windows suck doesn't mean that Linux should suck
too. My own preference would be a more discriminating dialog
that offers three possibilities: 'do nothing', 'bounce the
service/application' and 'reboot'.
There is always the circle jerk
argument, well linux reboots more than
windows, no, windows abandoned the mandatory restart to end up reinstating
it. Then, because windows restarts after update sucks, so linux doesn't
need to suck like windows. Rebooting after updates is not a trivial matter,
if you are a expert don't use PackageKit. Even tho it notifies you of
updates doesn't mean you need to use it, there is synaptic, yum extender,
yum, etc. none of those require you to reboot, also I don't think that
PackageKit mandates a reboot, just asks and you can cancel out of it, I
don't remember now, but I think so I know it flags the packages that cause a
restart. Anyway nothing will be solved in this thread, for those of us that
have been burned by failing to restart, and those that think that rebooting
after updates is trivial, will continue to do the things that are required
for us to be satisfied in a update.
--
fedora-devel-list mailing list
fedora-devel-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
5:09 a.m.
On Wed, Dec 16, 2009 at 9:43 PM, Otto Haliburton
<ottohaliburton(a)tx.rr.com> wrote:
----- Original Message ----- From: "nodata" <lsof(a)nodata.co.uk>
To: "Development discussions related to Fedora"
<fedora-devel-list(a)redhat.com>
Sent: Wednesday, December 16, 2009 11:29 AM
Subject: Re: packages requiring me to reboot...
> Am 2009-12-16 18:21, schrieb Seth Vidal:
>>
>>
>> On Wed, 16 Dec 2009, nodata wrote:
>>
>>>>
>>>> we're talking about the experienced user who is comfortable knowing
>>>> what
>>>> does and does not need a reboot.
>>>>
>>>> All I'm saying is - we've not taken away any option, the
experienced
>>>> user can do what they want.
>>>>
>>>> -sv
>>>>
>>>
>>> True, but the default should be sensible.
>>
>> And the default is sensible for the inexperienced user:
>>
>> Don't try to explain to the user how to restart the apps individually,
>> tell them to bounce the box and it will be the right version when it
>> comes back.
>>
>> -sv
>>
>
> On the other hand I think requiring more reboots than Windows is a bad
> thing...
>
windows update will automatically reboot your system when it automatically
updates it
windows tried the optional stuff but now almost every case it requires a
restart.
I don't like the term "experienced user" and I never feel comfortable
adding myself to that group but anyway,
- I don't want Windows to automatically reboot so I disable the
automatic Windows Update on the machines I'm using.
- I don't want my Fedora to reboot automatically so I disable and
remove PackageKit on the machines I'm using.
There isn't that much I could say about the times Fedora ask for a
reboot but at least I think it's kind of "unfair" to compare it with
an OS which pushes their updates just once a month.
Just my € 0.02.
--
With kind regards / Met vriendelijke groet,
Eelko Berkenpies
http://blog.berkenpies.nl/
8:02 a.m.
-----Original Message-----
From: fedora-devel-list-bounces(a)redhat.com [mailto:fedora-devel-list-
bounces(a)redhat.com] On Behalf Of Eelko Berkenpies
Sent: Thursday, December 17, 2009 05:09
To: Development discussions related to Fedora
Subject: Re: packages requiring me to reboot...
On Wed, Dec 16, 2009 at 9:43 PM, Otto Haliburton
<ottohaliburton(a)tx.rr.com> wrote:
>
> ----- Original Message ----- From: "nodata" <lsof(a)nodata.co.uk>
> To: "Development discussions related to Fedora"
> <fedora-devel-list(a)redhat.com>
> Sent: Wednesday, December 16, 2009 11:29 AM
> Subject: Re: packages requiring me to reboot...
>
>
>> Am 2009-12-16 18:21, schrieb Seth Vidal:
>>>
>>>
>>> On Wed, 16 Dec 2009, nodata wrote:
>>>
>>>>>
>>>>> we're talking about the experienced user who is comfortable
knowing
>>>>> what
>>>>> does and does not need a reboot.
>>>>>
>>>>> All I'm saying is - we've not taken away any option, the
experienced
>>>>> user can do what they want.
>>>>>
>>>>> -sv
>>>>>
>>>>
>>>> True, but the default should be sensible.
>>>
>>> And the default is sensible for the inexperienced user:
>>>
>>> Don't try to explain to the user how to restart the apps individually,
>>> tell them to bounce the box and it will be the right version when it
>>> comes back.
>>>
>>> -sv
>>>
>>
>> On the other hand I think requiring more reboots than Windows is a bad
>> thing...
>>
> windows update will automatically reboot your system when it
automatically
> updates it
> windows tried the optional stuff but now almost every case it requires a
> restart.
I don't like the term "experienced user" and I never feel comfortable
adding myself to that group but anyway,
- I don't want Windows to automatically reboot so I disable the
automatic Windows Update on the machines I'm using.
- I don't want my Fedora to reboot automatically so I disable and
remove PackageKit on the machines I'm using.
There isn't that much I could say about the times Fedora ask for a
reboot but at least I think it's kind of "unfair" to compare it with
an OS which pushes their updates just once a month.
Just my € 0.02.
first of all PackageKit does not do mandatory reboots. If you
hadn't
disabled it you would know that. In fact the people that are complaining
don't seem to have any idea why reboots are necessary. You need to get a
grip on file processing, cache, and other processes that speeds up execution
then you will know why it is not trivial. i.e. you kill a task that is in
the process of writing data to a file after you update it. What happens????
--
With kind regards / Met vriendelijke groet,
Eelko Berkenpies
http://blog.berkenpies.nl/
--
fedora-devel-list mailing list
fedora-devel-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
3:08 p.m.
Am 2009-12-17 15:02, schrieb Otto Haliburton:
> -----Original Message-----
> From: fedora-devel-list-bounces(a)redhat.com [mailto:fedora-devel-list-
> bounces(a)redhat.com] On Behalf Of Eelko Berkenpies
> Sent: Thursday, December 17, 2009 05:09
> To: Development discussions related to Fedora
> Subject: Re: packages requiring me to reboot...
>
> On Wed, Dec 16, 2009 at 9:43 PM, Otto Haliburton
> <ottohaliburton(a)tx.rr.com> wrote:
>>
>> ----- Original Message ----- From: "nodata"<lsof(a)nodata.co.uk>
>> To: "Development discussions related to Fedora"
>> <fedora-devel-list(a)redhat.com>
>> Sent: Wednesday, December 16, 2009 11:29 AM
>> Subject: Re: packages requiring me to reboot...
>>
>>
>>> Am 2009-12-16 18:21, schrieb Seth Vidal:
>>>>
>>>>
>>>> On Wed, 16 Dec 2009, nodata wrote:
>>>>
>>>>>>
>>>>>> we're talking about the experienced user who is comfortable
knowing
>>>>>> what
>>>>>> does and does not need a reboot.
>>>>>>
>>>>>> All I'm saying is - we've not taken away any option, the
experienced
>>>>>> user can do what they want.
>>>>>>
>>>>>> -sv
>>>>>>
>>>>>
>>>>> True, but the default should be sensible.
>>>>
>>>> And the default is sensible for the inexperienced user:
>>>>
>>>> Don't try to explain to the user how to restart the apps
individually,
>>>> tell them to bounce the box and it will be the right version when it
>>>> comes back.
>>>>
>>>> -sv
>>>>
>>>
>>> On the other hand I think requiring more reboots than Windows is a bad
>>> thing...
>>>
>> windows update will automatically reboot your system when it
> automatically
>> updates it
>> windows tried the optional stuff but now almost every case it requires a
>> restart.
>
> I don't like the term "experienced user" and I never feel comfortable
> adding myself to that group but anyway,
>
> - I don't want Windows to automatically reboot so I disable the
> automatic Windows Update on the machines I'm using.
>
> - I don't want my Fedora to reboot automatically so I disable and
> remove PackageKit on the machines I'm using.
>
> There isn't that much I could say about the times Fedora ask for a
> reboot but at least I think it's kind of "unfair" to compare it with
> an OS which pushes their updates just once a month.
>
> Just my € 0.02.
first of all PackageKit does not do mandatory reboots. If you hadn't
disabled it you would know that. In fact the people that are complaining
don't seem to have any idea why reboots are necessary. You need to get a
grip on file processing, cache, and other processes that speeds up execution
then you will know why it is not trivial. i.e. you kill a task that is in
the process of writing data to a file after you update it. What happens????
>
> --
> With kind regards / Met vriendelijke groet,
>
> Eelko Berkenpies
> http://blog.berkenpies.nl/
>
> --
> fedora-devel-list mailing list
> fedora-devel-list(a)redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-devel-list
I wish mailing list discussions were point-for-point for-and-against. It
would be so much easier.
Here is my point: Windows requires a reboot less often than Linux. Argue
all you want, it's true.
Linux has quicker security updates than Linux. That's an advantage.
ksplice can patch a running kernel...
3:17 p.m.
-----Original Message-----
From: fedora-devel-list-bounces(a)redhat.com [mailto:fedora-devel-list-
bounces(a)redhat.com] On Behalf Of nodata
Sent: Thursday, December 17, 2009 15:09
To: Development discussions related to Fedora
Subject: Re: packages requiring me to reboot...
Am 2009-12-17 15:02, schrieb Otto Haliburton:
>
>
>> -----Original Message-----
>> From: fedora-devel-list-bounces(a)redhat.com [mailto:fedora-devel-list-
>> bounces(a)redhat.com] On Behalf Of Eelko Berkenpies
>> Sent: Thursday, December 17, 2009 05:09
>> To: Development discussions related to Fedora
>> Subject: Re: packages requiring me to reboot...
>>
>> On Wed, Dec 16, 2009 at 9:43 PM, Otto Haliburton
>> <ottohaliburton(a)tx.rr.com> wrote:
>>>
>>> ----- Original Message ----- From:
"nodata"<lsof(a)nodata.co.uk>
>>> To: "Development discussions related to Fedora"
>>> <fedora-devel-list(a)redhat.com>
>>> Sent: Wednesday, December 16, 2009 11:29 AM
>>> Subject: Re: packages requiring me to reboot...
>>>
>>>
>>>> Am 2009-12-16 18:21, schrieb Seth Vidal:
>>>>>
>>>>>
>>>>> On Wed, 16 Dec 2009, nodata wrote:
>>>>>
>>>>>>>
>>>>>>> we're talking about the experienced user who is
comfortable
knowing
>>>>>>> what
>>>>>>> does and does not need a reboot.
>>>>>>>
>>>>>>> All I'm saying is - we've not taken away any option,
the
experienced
>>>>>>> user can do what they want.
>>>>>>>
>>>>>>> -sv
>>>>>>>
>>>>>>
>>>>>> True, but the default should be sensible.
>>>>>
>>>>> And the default is sensible for the inexperienced user:
>>>>>
>>>>> Don't try to explain to the user how to restart the apps
individually,
>>>>> tell them to bounce the box and it will be the right version when
it
>>>>> comes back.
>>>>>
>>>>> -sv
>>>>>
>>>>
>>>> On the other hand I think requiring more reboots than Windows is a
bad
>>>> thing...
>>>>
>>> windows update will automatically reboot your system when it
>> automatically
>>> updates it
>>> windows tried the optional stuff but now almost every case it requires
a
>>> restart.
>>
>> I don't like the term "experienced user" and I never feel
comfortable
>> adding myself to that group but anyway,
>>
>> - I don't want Windows to automatically reboot so I disable the
>> automatic Windows Update on the machines I'm using.
>>
>> - I don't want my Fedora to reboot automatically so I disable and
>> remove PackageKit on the machines I'm using.
>>
>> There isn't that much I could say about the times Fedora ask for a
>> reboot but at least I think it's kind of "unfair" to compare it
with
>> an OS which pushes their updates just once a month.
>>
>> Just my € 0.02.
> first of all PackageKit does not do mandatory reboots. If you hadn't
> disabled it you would know that. In fact the people that are
complaining
> don't seem to have any idea why reboots are necessary. You need to get
a
> grip on file processing, cache, and other processes that speeds up
execution
> then you will know why it is not trivial. i.e. you kill a task that is
in
> the process of writing data to a file after you update it. What
happens????
>>
>> --
>> With kind regards / Met vriendelijke groet,
>>
>> Eelko Berkenpies
>> http://blog.berkenpies.nl/
>>
>> --
>> fedora-devel-list mailing list
>> fedora-devel-list(a)redhat.com
>> https://www.redhat.com/mailman/listinfo/fedora-devel-list
>
>
I wish mailing list discussions were point-for-point for-and-against. It
would be so much easier.
Here is my point: Windows requires a reboot less often than Linux. Argue
all you want, it's true.
I have three computers running windows XP Windows
Vista and Windows Vista
Premium 24 hours per day, I have 2 computers dedicated to Linux so I know
what I am talking about. Windows is a commercial system, it gets paid for
what it produces and so it would be nice for them to boot less, but now they
reboot on every update and that is once a week generally on Tuesday.
Linux has quicker security updates than Linux. That's an advantage.
ksplice can patch a running kernel...
If you really want that then you can design and use it yourself. I don't
believe that anyone wants to patch a running Kernel especially without
testing and not be able to recover the old kernel. Are you really thinking
and considering the reasons for a reboot, #1 is simplicity!!!
--
fedora-devel-list mailing list
fedora-devel-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
3:23 p.m.
-----Original Message-----
From: fedora-devel-list-bounces(a)redhat.com [mailto:fedora-devel-list-
bounces(a)redhat.com] On Behalf Of nodata
Sent: Thursday, December 17, 2009 15:09
To: Development discussions related to Fedora
Subject: Re: packages requiring me to reboot...
Am 2009-12-17 15:02, schrieb Otto Haliburton:
>
>
>> -----Original Message-----
>> From: fedora-devel-list-bounces(a)redhat.com [mailto:fedora-devel-list-
>> bounces(a)redhat.com] On Behalf Of Eelko Berkenpies
>> Sent: Thursday, December 17, 2009 05:09
>> To: Development discussions related to Fedora
>> Subject: Re: packages requiring me to reboot...
>>
>> On Wed, Dec 16, 2009 at 9:43 PM, Otto Haliburton
>> <ottohaliburton(a)tx.rr.com> wrote:
>>>
>>> ----- Original Message ----- From:
"nodata"<lsof(a)nodata.co.uk>
>>> To: "Development discussions related to Fedora"
>>> <fedora-devel-list(a)redhat.com>
>>> Sent: Wednesday, December 16, 2009 11:29 AM
>>> Subject: Re: packages requiring me to reboot...
>>>
>>>
>>>> Am 2009-12-16 18:21, schrieb Seth Vidal:
>>>>>
>>>>>
>>>>> On Wed, 16 Dec 2009, nodata wrote:
>>>>>
>>>>>>>
>>>>>>> we're talking about the experienced user who is
comfortable
knowing
>>>>>>> what
>>>>>>> does and does not need a reboot.
>>>>>>>
>>>>>>> All I'm saying is - we've not taken away any option,
the
experienced
>>>>>>> user can do what they want.
>>>>>>>
>>>>>>> -sv
>>>>>>>
>>>>>>
>>>>>> True, but the default should be sensible.
>>>>>
>>>>> And the default is sensible for the inexperienced user:
>>>>>
>>>>> Don't try to explain to the user how to restart the apps
individually,
>>>>> tell them to bounce the box and it will be the right version when
it
>>>>> comes back.
>>>>>
>>>>> -sv
>>>>>
>>>>
>>>> On the other hand I think requiring more reboots than Windows is a
bad
>>>> thing...
>>>>
>>> windows update will automatically reboot your system when it
>> automatically
>>> updates it
>>> windows tried the optional stuff but now almost every case it requires
a
>>> restart.
>>
>> I don't like the term "experienced user" and I never feel
comfortable
>> adding myself to that group but anyway,
>>
>> - I don't want Windows to automatically reboot so I disable the
>> automatic Windows Update on the machines I'm using.
>>
>> - I don't want my Fedora to reboot automatically so I disable and
>> remove PackageKit on the machines I'm using.
>>
>> There isn't that much I could say about the times Fedora ask for a
>> reboot but at least I think it's kind of "unfair" to compare it
with
>> an OS which pushes their updates just once a month.
>>
>> Just my € 0.02.
> first of all PackageKit does not do mandatory reboots. If you hadn't
> disabled it you would know that. In fact the people that are
complaining
> don't seem to have any idea why reboots are necessary. You need to get
a
> grip on file processing, cache, and other processes that speeds up
execution
> then you will know why it is not trivial. i.e. you kill a task that is
in
> the process of writing data to a file after you update it. What
happens????
>>
>> --
>> With kind regards / Met vriendelijke groet,
>>
>> Eelko Berkenpies
>> http://blog.berkenpies.nl/
>>
>> --
>> fedora-devel-list mailing list
>> fedora-devel-list(a)redhat.com
>> https://www.redhat.com/mailman/listinfo/fedora-devel-list
>
>
I wish mailing list discussions were point-for-point for-and-against. It
would be so much easier.
Here is my point: Windows requires a reboot less often than Linux. Argue
all you want, it's true.
Linux has quicker security updates than Linux. That's an advantage.
ksplice can patch a running kernel...
LOL, you have to be joking with all the stuff
that blows up around here you
would trust someone patch you kernel while it was running and especially if
like me you run a custom built kernel. I applaud your trust and guts if you
will stand for that.
--
fedora-devel-list mailing list
fedora-devel-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
5:25 p.m.
On Thu, 2009-12-17 at 22:08 +0100, nodata wrote:
Here is my point: Windows requires a reboot less often than Linux.
Argue
all you want, it's true.
It's entirely false, because Linux *never* requires a reboot. Fedora
(not Linux, you are generalizing too far) *advises* reboots, it never
requires them. Windows forces you to reboot - literally, you cannot
prevent it from rebooting when it decides you have to.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net
4:08 a.m.
tor 2009-12-17 klockan 22:08 +0100 skrev nodata:
I wish mailing list discussions were point-for-point for-and-against.
It
would be so much easier.
You can easily accomplish this by quoting only the necessary text of the
parent post and commenting between the lines.
My Evolution preview window shows 38 lines of text. When _all_ of those
are quoted text, the person who sent that mail did something wrong. (I'm
not picking on you specifically, I think it's a general problem on this
list.)
Here is my point: Windows requires a reboot less often than Linux.
Argue
all you want, it's true.
Like has been pointed out before, Fedora does not force a reboot at all.
All it does is show an icon in the corner, telling me I need to log out
or I need to reboot. It could gain more options, like "need to restart
Firefox".
Linux has quicker security updates than Linux. That's an
advantage.
ksplice can patch a running kernel...
I want suspend and resume with a new kernel. Also, ponies. :)
/abo
5:01 p.m.
On Wed, Dec 16, 2009 at 11:43 AM, Seth Vidal <skvidal(a)fedoraproject.org> wrote:
you're an experienced user? You're comfortable knowing what
does and what
does not require a reboot? Then why are you using PK?
Disable pk and do the updates directly via yum.
Bam - no more requests to reboot.
This is a completely bogus rationale but one I commonly hear on this list.
I, and many other fedora users would be quite *capable* of running our
systems with any help of a distribution, we could go and fetch from
source and do all the integration ourselves...
...but we'd actually like to get some work done using our computers
and don't want to burn our lives away playing master-of-my-own-distro,
though we're willing to spend some time contributing to a shared
effort to build a good distribution for many.
In exchange for not having to personally micro-manage things, we're
willing to tolerate some things being configured in violation of our
own preferences or aesthetics, or even a few things being outright
broken, but that doesn't mean that it's not important for it to work
right.
Yes, I'm quite capable of executing some big manual process or
changing packagekit to behave like I want. But every such action has
costs, it takes time and effort which usually has to be repeated every
upgrade. The non-standard configuration carries the risk of triggering
bugs in other system components, breaking the upgrade process, etc.
The gratuitous reboots are harmful to all users. They diminish a
significant advantage our systems can have compared to alternatives
like Microsoft Windows. They discourage the reporting of bugs in
applications… "System acting weird? Just restart!". When triggered at
inconvenient times they can cause significant harm by interrupting
people's work.
Yes— users with more expertise are more likely to complain about this,
but thats not reason to dismiss the issue. If there were truly a
disconnect here betweens the needs of the novices and those of the
expert users you could argue favouring the novices, but that just
isn't applicable here.
5:08 p.m.
On Wed, Dec 16, 2009 at 2:01 PM, Gregory Maxwell <gmaxwell(a)gmail.com> wrote:
Yes— users with more expertise are more likely to complain about
this,
but thats not reason to dismiss the issue. If there were truly a
disconnect here betweens the needs of the novices and those of the
expert users you could argue favouring the novices, but that just
isn't applicable here.
Uhm. am I missing something. Aren't we talking about reboot requests
that PK is spawning and I can choose to cancel in the UI interaction
because I know better instead of mandatory reboots?
-jef
1 a.m.
Am 2009-12-17 00:08, schrieb Jeff Spaleta:
On Wed, Dec 16, 2009 at 2:01 PM, Gregory
Maxwell<gmaxwell(a)gmail.com> wrote:
> Yes— users with more expertise are more likely to complain about this,
> but thats not reason to dismiss the issue. If there were truly a
> disconnect here betweens the needs of the novices and those of the
> expert users you could argue favouring the novices, but that just
> isn't applicable here.
Uhm. am I missing something. Aren't we talking about reboot requests
that PK is spawning and I can choose to cancel in the UI interaction
because I know better instead of mandatory reboots?
-jef
No, we're talking about requiring fewer reboots for normal users.
Prompting a user like this teaches them to ignore recommendations. This
isn't a good thing.
3:36 a.m.
-----Original Message-----
From: fedora-devel-list-bounces(a)redhat.com [mailto:fedora-devel-list-
bounces(a)redhat.com] On Behalf Of nodata
Sent: Thursday, December 17, 2009 01:01
To: Development discussions related to Fedora
Subject: Re: packages requiring me to reboot...
Am 2009-12-17 00:08, schrieb Jeff Spaleta:
> On Wed, Dec 16, 2009 at 2:01 PM, Gregory Maxwell<gmaxwell(a)gmail.com>
wrote:
>> Yes- users with more expertise are more likely to complain about this,
>> but thats not reason to dismiss the issue. If there were truly a
>> disconnect here betweens the needs of the novices and those of the
>> expert users you could argue favouring the novices, but that just
>> isn't applicable here.
>
> Uhm. am I missing something. Aren't we talking about reboot requests
> that PK is spawning and I can choose to cancel in the UI interaction
> because I know better instead of mandatory reboots?
>
> -jef
>
No, we're talking about requiring fewer reboots for normal users.
Prompting a user like this teaches them to ignore recommendations. This
isn't a good thing.
there are no mandatory reboots in PackageKit, you are notified what packages
will cause a request to reboot and you can exit the process without
rebooting!!!!!! Or you can remove the packages from the update processes
and install when convenient for you.
--
fedora-devel-list mailing list
fedora-devel-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
12:50 p.m.
Am 2009-12-17 10:36, schrieb Otto Haliburton:
> -----Original Message-----
> From: fedora-devel-list-bounces(a)redhat.com [mailto:fedora-devel-list-
> bounces(a)redhat.com] On Behalf Of nodata
> Sent: Thursday, December 17, 2009 01:01
> To: Development discussions related to Fedora
> Subject: Re: packages requiring me to reboot...
>
> Am 2009-12-17 00:08, schrieb Jeff Spaleta:
>> On Wed, Dec 16, 2009 at 2:01 PM, Gregory Maxwell<gmaxwell(a)gmail.com>
> wrote:
>>> Yes- users with more expertise are more likely to complain about this,
>>> but thats not reason to dismiss the issue. If there were truly a
>>> disconnect here betweens the needs of the novices and those of the
>>> expert users you could argue favouring the novices, but that just
>>> isn't applicable here.
>>
>> Uhm. am I missing something. Aren't we talking about reboot requests
>> that PK is spawning and I can choose to cancel in the UI interaction
>> because I know better instead of mandatory reboots?
>>
>> -jef
>>
>
> No, we're talking about requiring fewer reboots for normal users.
>
> Prompting a user like this teaches them to ignore recommendations. This
> isn't a good thing.
>
there are no mandatory reboots in PackageKit, you are notified what packages
will cause a request to reboot and you can exit the process without
rebooting!!!!!! Or you can remove the packages from the update processes
and install when convenient for you.
yep. but all of that assumes I know what I am doing, and the people that
this is aimed at don't. windows requires fewer reboots now.
1:05 p.m.
On 12/17/2009 01:50 PM, nodata wrote:
yep. but all of that assumes I know what I am doing, and the people
that
this is aimed at don't. windows requires fewer reboots now.
+1, and remember that they have an advantage right off the bat:
- much fewer subsystems (Windows and a couple of tiny apps, vs. Linux's
entire universe of applications)
- patch model that pushes large patch sets at long intervals rather than
frequent fine-grained patches.
I think Linux has to have a better heuristic as to when a reboot is
necessary. Actually, any event that breaks the user's work flow is as
bad: X crash/logoff is as disruptive as a reboot, unless we had a way to
restore the application state in the way Firefox or Emacs or OpenOffice
recover from crashes (restarting, opening the windows where they were
and recovering the content).
1:55 p.m.
-----Original Message-----
From: fedora-devel-list-bounces(a)redhat.com [mailto:fedora-devel-list-
bounces(a)redhat.com] On Behalf Of Przemek Klosowski
Sent: Thursday, December 17, 2009 13:05
To: fedora-devel-list(a)redhat.com
Subject: Re: packages requiring me to reboot...
On 12/17/2009 01:50 PM, nodata wrote:
> yep. but all of that assumes I know what I am doing, and the people that
> this is aimed at don't. windows requires fewer reboots now.
+1, and remember that they have an advantage right off the bat:
- much fewer subsystems (Windows and a couple of tiny apps, vs. Linux's
entire universe of applications)
- patch model that pushes large patch sets at long intervals rather than
frequent fine-grained patches.
I think Linux has to have a better heuristic as to when a reboot is
necessary. Actually, any event that breaks the user's work flow is as
bad: X crash/logoff is as disruptive as a reboot, unless we had a way to
restore the application state in the way Firefox or Emacs or OpenOffice
recover from crashes (restarting, opening the windows where they were
and recovering the content).
I am now getting a picture of why this subject is so hard to grasp, everyone
who is for fewer reboots address the issue on a task by task bases, while
those who support reboots think of it on a system bases. Windows now
restarts each time a patch occurs, at the current time I can't think of any
patch in the last 3 months which hasn't required a reboot. Another reason is
that some of the windows operating systems are coming to their end of life
cycle, and windows is choosing to do a lot of patches in one update, but
believe me, it is startling to come in the next morning to see your computer
restarted with a message that a update was performed and you were restarted
and that occurs with all of the windows operating systems that are currently
supported. Linux has many applications that are running and on each update
you can have as many as many as 65(the last update) task being updated each
week and how you will avoid reboots will be amazing to me. I think then
people will complain that it takes 6 hours to update now and it use to take
30 minutes. At the present time you have apps that run across logoff and
login, so trying to get into starting and stopping task in update situations
will be a nightmare, but if there are some ambitious people out there go at
it, there are only 5000 apps that need to be updated to solve the problem.
--
fedora-devel-list mailing list
fedora-devel-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
2:07 a.m.
Otto Haliburton wrote:
Windows now
restarts each time a patch occurs, at the current time I can't think of any
patch in the last 3 months which hasn't required a reboot. Another reason is
that some of the windows operating systems are coming to their end of life
cycle, and windows is choosing to do a lot of patches in one update, but
believe me, it is startling to come in the next morning to see your computer
restarted with a message that a update was performed and you were restarted
and that occurs with all of the windows operating systems that are currently
supported.
Please be consistent. Windows also doesn't *have* to auto restart.
My
systems are set to manual update of predownloaded files. Besides, since
my systems are dual (or more) boot, they wouldn't auto reboot into the
running version of windows. Would be very annoying.
4:11 a.m.
-----Original Message-----
From: fedora-devel-list-bounces(a)redhat.com [mailto:fedora-devel-list-
bounces(a)redhat.com] On Behalf Of shmuel siegel
Sent: Friday, December 18, 2009 02:07
To: Development discussions related to Fedora
Subject: Re: packages requiring me to reboot...
Otto Haliburton wrote:
>
> Windows now
> restarts each time a patch occurs, at the current time I can't think of
any
> patch in the last 3 months which hasn't required a reboot. Another
reason is
> that some of the windows operating systems are coming to their end of
life
> cycle, and windows is choosing to do a lot of patches in one update, but
> believe me, it is startling to come in the next morning to see your
computer
> restarted with a message that a update was performed and you were
restarted
> and that occurs with all of the windows operating systems that are
currently
> supported.
Please be consistent. Windows also doesn't *have* to auto restart. My
systems are set to manual update of predownloaded files. Besides, since
my systems are dual (or more) boot, they wouldn't auto reboot into the
running version of windows. Would be very annoying.
you don't follow the list very well, and obviously didn't read the post
that
this replies to so don't go around calling people inconsistent. Windows
forces you to reboot and there is no mandatory reboots in Linux and windows
does force you to restart, don't restart when it ask you and see what
happens, it will not continue without insisting that you restart. Where do
you get this crap? With windows you cannot avoid a restart. And there is no
auto-restart in windows it is auto-update and with auto-update you have no
choice it will restart for you. Try it then tell me about it.
--
fedora-devel-list mailing list
fedora-devel-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
4:28 a.m.
On Fri, 2009-12-18 at 04:11 -0600, Otto Haliburton wrote:
you don't follow the list very well, and obviously didn't
read the post that
this replies to so don't go around calling people inconsistent. Windows
forces you to reboot and there is no mandatory reboots in Linux and windows
does force you to restart, don't restart when it ask you and see what
happens, it will not continue without insisting that you restart. Where do
you get this crap? With windows you cannot avoid a restart. And there is no
auto-restart in windows it is auto-update and with auto-update you have no
choice it will restart for you. Try it then tell me about it.
You're arguing from different starting points, and there's no need to be
uncivil about it. Shmuel's point is that you can disable automatic
updates on Windows (which is quite true), and hence you won't then be
forced to restart until you manually choose to install the updates.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net
6:48 p.m.
----- Original Message -----
From: "Gregory Maxwell" <gmaxwell(a)gmail.com>
To: "Development discussions related to Fedora"
<fedora-devel-list(a)redhat.com>
Sent: Wednesday, December 16, 2009 5:01 PM
Subject: Re: packages requiring me to reboot...
On Wed, Dec 16, 2009 at 11:43 AM, Seth Vidal
<skvidal(a)fedoraproject.org>
wrote:
> you're an experienced user? You're comfortable knowing what does and what
> does not require a reboot? Then why are you using PK?
>
> Disable pk and do the updates directly via yum.
>
> Bam - no more requests to reboot.
This is a completely bogus rationale but one I commonly hear on this list.
I, and many other fedora users would be quite *capable* of running our
systems with any help of a distribution, we could go and fetch from
source and do all the integration ourselves...
that my friend is the open source concept. The users test, develop, &
create new ideas
for development.
...but we'd actually like to get some work done using our
computers
and don't want to burn our lives away playing master-of-my-own-distro,
though we're willing to spend some time contributing to a shared
effort to build a good distribution for many.
then go commercial, you pays your money you get your dues
In exchange for not having to personally micro-manage things,
we're
willing to tolerate some things being configured in violation of our
own preferences or aesthetics, or even a few things being outright
broken, but that doesn't mean that it's not important for it to work
right.
how much do you pay for Fedora, as I remember I think it is free!!!
Yes, I'm quite capable of executing some big manual process or
changing packagekit to behave like I want. But every such action has
costs, it takes time and effort which usually has to be repeated every
upgrade. The non-standard configuration carries the risk of triggering
bugs in other system components, breaking the upgrade process, etc.
as I said go commercial!!
The gratuitous reboots are harmful to all users. They diminish a
significant advantage our systems can have compared to alternatives
like Microsoft Windows. They discourage the reporting of bugs in
applications… "System acting weird? Just restart!". When triggered at
inconvenient times they can cause significant harm by interrupting
people's work.
If that is so, then you and them are wasting your time with open source
Yes— users with more expertise are more likely to complain about
this,
but thats not reason to dismiss the issue. If there were truly a
disconnect here betweens the needs of the novices and those of the
expert users you could argue favouring the novices, but that just
isn't applicable here.
every one contributes. It's free!!!! that is why you need to be more
tolerant.
--
fedora-devel-list mailing list
fedora-devel-list(a)redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
5249
days inactive
5253
days old
66 comments
27 participants
participants (27)
-
Adam Williamson -
Alexander Boström -
Bill Nottingham -
Chris Adams -
Clark Williams -
Colin Walters -
Development discussions related to Fedora -
Eelko Berkenpies -
Ewan Mac Mahon -
Frank Murphy -
Genes MailLists -
Gregory Maxwell -
James Antill -
Jef Spaleta -
Jon Masters -
Jon Masters -
Nathanael D. Noblet -
Nikola Pajkovsky -
nodata -
ottos -
Peter Jones -
Przemek Klosowski -
Ralf Corsepius -
Richard Hughes -
Seth Vidal -
shmuel siegel -
Stefan Schulze Frielinghaus