9.10 goes into a sleep of death whenever power switches to battery.

[suguru]

Binary package hint: acpi-support

Ubuntu 9.10 has this problem but not Kubuntu 9.10. It is a big problem now but not in previous releases. Once I got this kernel panic which you can see in the enclosed dmesg immediately after restart.

[Steve Langasek (vorlon)]

Thank you for taking the time to report this bug and help to improve Ubuntu.

Can you please explain what you mean by "sleep of death"? What exactly do you see? Is it reproducible if you switch to a text console before disconnecting power?

The acpi-support package is shared between Ubuntu and Kubuntu, so is unlikely to be the source of the bug; but we need more information to be able to narrow down the nature of this problem.

The dmesg information seems unlikely to be related, since this happens post-reboot and apparently does not cause the same "sleep of death" behavior (since the problem was successfully logged).

[suguru (oouc)]

By sleep of death I mean that there is disk activity, cpu goes into hi gear, then the run led shuts down, the sleep led comes on and nothing works except holding down the power button until the system shuts down. This happened when I first loaded 9.10 from the DVD in the magazine Ubuntu User. So I tried it with the CD from UDS-L and got the same behaviour. So I tried the Kubuntu 9.10 from UDS-L and it ran fine except it took Jonathan Riddle 1/2 an hour to get on the WiFi there. When I shut down another developer spent over a 1/2 hour and could not get Kubuntu 9.10 to connect to the WiFi again although he was looking at the ./bash_history of the previous success. The kernel team told me there was a fix in
https://wiki.ubuntu.com/DebuggingKernelSuspendHibernateResume
or
https://wiki.ubuntu.com/DebuggingKernelSuspend
so I went to both places but could not solve the problem.

Next I did a search in Launchpad and found about 90 things marked "solved" for this problem. After too much wasted time I found one that said "poke around in" url.... The user said thanks, I solved it and he did not mention anything about how he solved. This struck me as the very poor communication which I see so often which seems to me to be the greatest problem holding back the adoption of Linux.

So I went back to 8.04 to get some urgent things done. At night, I again loaded the 32 bit 9.10 from Ubuntu User. Then I loaded the kubuntu-desktop. Then I upgraded kubuntu and ubuntu.

The problem still occurs only in Ubuntu 9.10 not in Kubuntu 9.10. It also happens when the only window I have open in Gnome is the terminal and I disconnect the power adapter. I also rebooted without the power adaptor. As soon as I plugged the power adaptor in I again got the sleep of death.

Another thing I did was change all of the Gnome settings so that it would suspend everything to disk rather than ram. But this did not really make much difference other than it took longer to sleep into death. Sometimes when I hit the power button briefly in an effort to resume, I will see activity on the hard disk and I nearly always remember the cpu changing from 800 to 1.6 Ghz.

My laptop computer is from China. I bought it in June 2006 from Fry's with a 5 year replacement guarantee. But Fry's makes excuses instead of honoring their $199 extended warranty. It is an AMD Turion 64 with ATI graphics and Rt2500pci wifi. It may have been made from the same place that made Compaq. The female plug accepting the power adapter is worn out so it switches from power to battery with every breath on or movement of the power cord. This caused the last half of my UDS-L session to be without a computer since I was running Gnome because I really dislike every package having to be something new and special that must begin with the letter K instead of using what was proven many years ago and continually improved since them.

Any help you give will be appreciated and with proper communications will likely help many in the third world who cannot buy a new laptop every 2 or 3 years.

[suguru (oouc)]

Instead of Gnome or Kubuntu, I booted into xterm. No power or shutdown problems when I unplugged and plugged in the power adaptor.

[Mark Spohr (mspohr)]

I am having a similar problem with 9.10 (clean install) on Dell 700m laptop. Power management worked great with 8.10 but with 9.10 whenever I plug or unplug the laptop power adapter, the machine goes into sleep or hibernate (whichever I have set in the power options). This happens regardless of the state of charge of the battery and it happens on both plugging and unplugging the power adapter.
The first message in dmesg after the unplug is:
[ 287.600663] b44: eth0: powering down PHY
This seems to be initiating the sleep or hibernate action.
I had 8.10 installed on this machine before and never had a problem like this. Power management worked perfectly with sleep and hibernate and I could plug and unplug the power adapter without causing a sleep or hibernate event.

[Steve Langasek (vorlon)]

Mark,

The problem you describe is unrelated to the bug described by the bug submitter. Please file a separate bug report.

Please include in your report the output of the 'acpi_listen' command left running across a suspend/resume cycle.

[Steve Langasek (vorlon)]

Reassigning this bug to gnome-power-manager, since it's GNOME-specific. suguru, if you disable *all* suspend/hibernate settings in gnome-power-manager (setting them all to 'blank screen'), do you still experience this problem?

[David Tombs (dgtombs)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. You reported this bug a while ago and there hasn't been any activity in it recently. We were wondering if this is still an issue for you. Thanks in advance.

[suguru (oouc)]

I tore apart this laptop and am running 10.04 on another laptop. So, last I knew it was a problem but not on every machine.

[suguru (oouc)]

I tore apart this laptop and am running 10.04 on another laptop. So, last I knew it was a problem but not on every machine.

If it is not a problem for anyone else, I won't put the laptop back together and disable all suspend/hibernate settings in gnome-power-manager by setting them to blank screen.

[David Tombs (dgtombs)]

OK, I am closing this bug since you can no longer reproduce the issue. Thanks for the report!