Re: 2.6.38-11-generic re-introduced old bug on Gateway FX6850-51u

[Aleksandar Milivojevic]

On Mon, Aug 22, 2011 at 7:04 PM, Colin Watson  wrote:
> On Mon, Aug 22, 2011 at 11:36:07AM -0700, Aleksandar Milivojevic wrote:
>> Actually, thinking about it, it may also be switching between grub-pc
>> and grub-efi.  Originally, Ubuntu 11.04 installer put grub-efi on the
>> system.  On next kernel update (where halting/rebooting problem was
>> fixed), updater replaced grub-efi with grub-pc.
>
> This *really* shouldn't happen and is a bug that needs to be addressed,
> regardless of any of the other stuff here.  The logs in /var/log/apt/
> and possibly /var/log/dist-upgrade/ may help to figure out why the
> package management system felt it needed to do this.

These logs probably rotated some time ago on my system.  If updater
attempts to replace grub-efi with grub-pc on next kernel update, I'll
check them out and report.  Alternatively, Brian can check out logs on
his system, since it seems updater on his system is still trying to
force him into replacing grub-efi with grub-pc.

>> I'm not sure how and why would choice of boot loader (grub-pc vs.
>> grub-efi) affect shutting down the system, but it looks like it does
>> in some odd way!?
>
> Booting in a completely different firmware mode, which is the effect of
> switching between grub-pc (BIOS) and grub-efi (UEFI), could easily make
> a difference here.

Makes sense.  I managed to find a screenshot with kernel trace from
some time ago (this was likely 2.6.38-8, or whatever was the current
Ubuntu kernel on 2011-05-23 (the date on the file).  The call trace
shows few EFI-related (based on the name) functions being executed at
the time system crashed during either shutdown or reboot.  I uploaded
the screen shot here (if text looks unreadable, use Picasa's controlls
to zoom in): 
https://picasaweb.google.com/105136527709185448712/MiscAttachments?authkey=Gv1sRgCKiehIiMnpn_qwE#5643908739416731554

-- 
Ubuntu-devel-discuss mailing list
Ubuntu-devel-discuss@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss

Re: 2.6.38-11-generic re-introduced old bug on Gateway FX6850-51u

[Colin Watson]

On Mon, Aug 22, 2011 at 11:36:07AM -0700, Aleksandar Milivojevic wrote:
> Actually, thinking about it, it may also be switching between grub-pc
> and grub-efi.  Originally, Ubuntu 11.04 installer put grub-efi on the
> system.  On next kernel update (where halting/rebooting problem was
> fixed), updater replaced grub-efi with grub-pc.

This *really* shouldn't happen and is a bug that needs to be addressed,
regardless of any of the other stuff here.  The logs in /var/log/apt/
and possibly /var/log/dist-upgrade/ may help to figure out why the
package management system felt it needed to do this.

> I'm not sure how and why would choice of boot loader (grub-pc vs.
> grub-efi) affect shutting down the system, but it looks like it does
> in some odd way!?

Booting in a completely different firmware mode, which is the effect of
switching between grub-pc (BIOS) and grub-efi (UEFI), could easily make
a difference here.

-- 
Colin Watson   [cjwat...@ubuntu.com]

-- 
Ubuntu-devel-discuss mailing list
Ubuntu-devel-discuss@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss

Re: 2.6.38-11-generic re-introduced old bug on Gateway FX6850-51u

[Brian Vaughan]

> I'm bringing this up, since I think I was actually able to reboot the
> system without it hanging once or twice between applying last kernel
> upgrade, and switching back to grub-efi.
> 
> I'm not sure how and why would choice of boot loader (grub-pc vs.
> grub-efi) affect shutting down the system, but it looks like it does
> in some odd way!?
> 
I've been having a similar issue with my Gateway DX4850: sometimes it
shuts down or reboots cleanly, sometimes it hangs, after all the system
services have stopped. I can switch to console login prompts, but not
actually log in. Manually powering off the box is necessary. My system
is set up for dual booting, and the other operating system doesn't hang
when rebooting or shutting down, so I don't think it's a hardware issue,
per se.

When I installed Ubuntu on this system, it used grub-efi-amd64, and I've
been using that -- there was a system update I delayed, as it wanted to
replace grub-efi-amd64 with grub-pc, and that seemed like something to
avoid changing.


-- 
Ubuntu-devel-discuss mailing list
Ubuntu-devel-discuss@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss

Re: 2.6.38-11-generic re-introduced old bug on Gateway FX6850-51u

[Aleksandar Milivojevic]

On Mon, Aug 22, 2011 at 9:44 AM, Aleksandar Milivojevic
 wrote:
> It looks like 2.6.38-11 re-introduced old bug, where kernel would hang
> on halt or reboot.  The bug was present on the original kernel shipped
> with Ubuntu 11.04 (2.6.38-8, if my memory serves me right), and fixed
> in the very next kernel update.  However, with 2.6.38-11, the problem
> is back.
>
> In short, at the very end of shutdown sequence, after all file systems
> were unmounted, etc, kernel would simply hang.  User have to
> long-press power button to physically power cycle the system in order
> to recover.  In my case this was mostly an annoyance, but for other
> users it may be much more serious problem, as it makes system
> impossible to reboot remotely (for obvious reasons).
>
> My system is an Gateway FX6850-51u, with 3.4GHz Intel Core i7-2600
> Sandy Bridge quad-core processor, 8GB DDR3 RAM.  It has an Accer
> branded motherboard (no identifier of exact motherboard version
> printed anywhere on it).  Gateway's website has full system specs
> (http://us.gateway.com/gw/en/US/content/model/PT.GBN02.012).  The only
> difference from stock system as sold by Gateway is PCI-e WiFi card
> removed and replaced with an USB WiFi stick, and hard drive replaced
> with an SSD drive.
>
> I saw few older discussions on the mailing list mentioning same
> problem on some other systems as well (for example, Dell's).
>
> I'd provide a bit more info, however once the system is power cycled,
> there's not much left on the system itself.

Actually, thinking about it, it may also be switching between grub-pc
and grub-efi.  Originally, Ubuntu 11.04 installer put grub-efi on the
system.  On next kernel update (where halting/rebooting problem was
fixed), updater replaced grub-efi with grub-pc.  Now, while system was
not hanging on shutdown anymore, grub-pc introduced bunch of other
problems during booting up the system, making booting up the system
unreliable (installer partitioned and configured the system originally
as an EFI system, so switching back to grub-pc broke things, as it
doesn't seem to know about EFI).  Then, after last kernel upgrade, I
switched back to grub-efi.  Switching back to grub-efi fixed booting
up the system, error messages when running grub-update (or
grub-install) are gone, but now system hangs again on shutdown.

I'm bringing this up, since I think I was actually able to reboot the
system without it hanging once or twice between applying last kernel
upgrade, and switching back to grub-efi.

I'm not sure how and why would choice of boot loader (grub-pc vs.
grub-efi) affect shutting down the system, but it looks like it does
in some odd way!?

-- 
Ubuntu-devel-discuss mailing list
Ubuntu-devel-discuss@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss

2.6.38-11-generic re-introduced old bug on Gateway FX6850-51u

[Aleksandar Milivojevic]

It looks like 2.6.38-11 re-introduced old bug, where kernel would hang
on halt or reboot.  The bug was present on the original kernel shipped
with Ubuntu 11.04 (2.6.38-8, if my memory serves me right), and fixed
in the very next kernel update.  However, with 2.6.38-11, the problem
is back.

In short, at the very end of shutdown sequence, after all file systems
were unmounted, etc, kernel would simply hang.  User have to
long-press power button to physically power cycle the system in order
to recover.  In my case this was mostly an annoyance, but for other
users it may be much more serious problem, as it makes system
impossible to reboot remotely (for obvious reasons).

My system is an Gateway FX6850-51u, with 3.4GHz Intel Core i7-2600
Sandy Bridge quad-core processor, 8GB DDR3 RAM.  It has an Accer
branded motherboard (no identifier of exact motherboard version
printed anywhere on it).  Gateway's website has full system specs
(http://us.gateway.com/gw/en/US/content/model/PT.GBN02.012).  The only
difference from stock system as sold by Gateway is PCI-e WiFi card
removed and replaced with an USB WiFi stick, and hard drive replaced
with an SSD drive.

I saw few older discussions on the mailing list mentioning same
problem on some other systems as well (for example, Dell's).

I'd provide a bit more info, however once the system is power cycled,
there's not much left on the system itself.

-- 
Ubuntu-devel-discuss mailing list
Ubuntu-devel-discuss@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss

Re: fcron

[Micah Gersten]

On 08/22/2011 01:17 AM, Steven Sroka wrote:
>> On 22 August 2011 00:21, Jeremy Bicha  wrote:
>> On 22 August 2011 00:06, Steven Sroka  wrote:
>>> Can I request that fcron is updated in the universe repo? The version
>>> there is over 3 years old. The version listed on the fcron homepage
>>> [0] is from April 5th of last year.
>> Actually, fcron was removed from Debian. See the bug report:
>> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=636235
>>
>> So this isn't a simple software upgrade as there were bugs that
>> weren't able to be fixed by the Debian maintainer. Upstream isn't
>> really actively developing it either.
> I spoke to the author of fcron and he is still developing it. If there
> were problems would it not be wise to either try the newest version or
> drop it from the universe repo all together since it is still there?
>
>
If it's not updated before release, we should drop it.  However, as
Ubuntu doesn't have maintainers, anyone can update this.  At this point
it would need a feature freeze exception, but I think that should be
easy enough to get.   If someone would like to prepare an updated
package for 3.0.6, I'm happy to sponsor it once the freeze exception is
granted.  Please use Bug #591813
 for preparing the update.

Thanks,
Micah
-- 
Ubuntu-devel-discuss mailing list
Ubuntu-devel-discuss@lists.ubuntu.com
Modify settings or unsubscribe at: 
https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss