can't update acpi-support package on lucid running of an external hdd

Bug #567182 reported by filip
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
acpi-support (Ubuntu)
Fix Released
High
Steve Langasek
Lucid
Fix Released
High
Steve Langasek
hdparm (Ubuntu)
Fix Released
High
Colin Watson
Lucid
Fix Released
High
Colin Watson

Bug Description

Binary package hint: acpi-support

I am running lucid 64-bit from my external hdd (freecom mobile xxs).
When i try to update the package it tries to disable powermanagement.
This takes relative long and after this my disk is read only.
Which triggers errors (could not enummerate,...) and of course the package doesn't install further.

I've tried booting with noacpi as bootparameter but it doesn't work.
Also disabling acpi-support with sysv-rc-conf in every runlevel didn't work.

I also had problems with errors at boot before but this is resolved with the new hdparm package.

I still have resets of the disk late in the bootproces, but this triggers no errors. Maybe this is related.

on the same hardware i can update the package when running lucid from my usb-stick (cruzer).

I suspect there is something wrong with the powermanagement of the disk. But haven't found a solution.
Is there something i can do to test?

Revision history for this message
Steve Langasek (vorlon) wrote :

The problem with your disk being marked read-only is probably related to pm-powersave invoking hdparm (via /usr/lib/pm-utils/power.d/95hdparm-apm).

If you boot with the 'nohdparm' option, does this solve the problem?

The acpi-support part of this bug is straightforward - we should not be calling 'invoke-rc.d acpi-support start' in the package postinst, the init script is only meant to be used for setting the initial power management policy at boot time.

Changed in acpi-support (Ubuntu):
status: New → Triaged
importance: Undecided → High
Revision history for this message
Steve Langasek (vorlon) wrote :

(and if the new hdparm package fixed your boot time problems, chances are it fixes the problem when upgrading acpi-support too - bug #515023.)

Revision history for this message
filip (bandit-s-fw) wrote :

The new hdparm package doesn't fix all my problems. I still get errors late at boot just before gnome start appearing.
And also it is impossible to update the acpi-support package.

However, booting with the nohdparm argument did the trick. I have been able to update the acpi-support with now problems and the resets late at boot are gone.
Thank you for mentioning the nohdparm argument. I've been looking for a solution for 5 months.
With karmic i also had these kind of problems.

What are the consequences of booting with nohdparm?

Revision history for this message
Steve Langasek (vorlon) wrote :

> The new hdparm package doesn't fix all my problems. I still get errors late
> at boot just before gnome start appearing.

What are these errors?

> And also it is impossible to update the acpi-support package.

Oh - yes, it seems the fix for bug #515023 only applies when called from udev; so when called to reapply hdparm settings on a power event, or from the (incorrect) acpi-support maintainer script, the problem still exists. Marking this bug as affecting hdparm as well.

> What are the consequences of booting with nohdparm?

Power management settings won't be applied to any of your hard drives with this setting.

Changed in hdparm (Ubuntu Lucid):
importance: Undecided → High
status: New → Triaged
Changed in acpi-support (Ubuntu Lucid):
assignee: nobody → Steve Langasek (vorlon)
milestone: none → ubuntu-10.04
Revision history for this message
filip (bandit-s-fw) wrote :

I mistakenly wrote errors. It must be resets. late in the bootproces i got 3 times the same reset message.

The resets are of the form like these:
.....
reset high speed USB device using ehci_hcd and address 4
reset high speed USB device using ehci_hcd and address 4
reset high speed USB device using ehci_hcd and address 4
....

(copied from another bug report as i am not on my home pc)

but with the nohdparm argument at boot these resets are gone.

Steve Langasek (vorlon)
Changed in acpi-support (Ubuntu Lucid):
status: Triaged → Fix Committed
Colin Watson (cjwatson)
Changed in hdparm (Ubuntu Lucid):
status: Triaged → Fix Committed
milestone: none → ubuntu-10.04
assignee: nobody → Colin Watson (cjwatson)
Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package hdparm - 9.15-1ubuntu9

---------------
hdparm (9.15-1ubuntu9) lucid; urgency=low

  * Fetch ID_PATH from 'udevadm info' if it isn't already set in the
    environment (LP: #567182).
  * Factor Firewire/USB check out into a new hdparm_try_apm function.
  * debian/95hdparm-apm: Skip Firewire and USB devices.
 -- Colin Watson <email address hidden> Thu, 22 Apr 2010 17:35:22 +0100

Changed in hdparm (Ubuntu Lucid):
status: Fix Committed → Fix Released
Revision history for this message
Launchpad Janitor (janitor) wrote :

This bug was fixed in the package acpi-support - 0.135

---------------
acpi-support (0.135) lucid; urgency=low

  * debian/rules: don't call the init script in the postinst, the script
    exists only to ensure the power policy is applied on boot so we should
    keep our hands off on upgrade. LP: #567182
 -- Steve Langasek <email address hidden> Fri, 23 Apr 2010 04:31:14 -0700

Changed in acpi-support (Ubuntu Lucid):
status: Fix Committed → Fix Released
Revision history for this message
filip (bandit-s-fw) wrote :

just booted with the new packages without the hdparm argument and everythings is ok.
Thank you for fixing this before the release. Many people will benefit from it.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.