udevd causes high cpu-usage

Bug #130818 reported by Doomi
14
This bug affects 2 people
Affects Status Importance Assigned to Milestone
udev (Ubuntu)
New
Undecided
Unassigned

Bug Description

It all looks like this Bug https://bugs.launchpad.net/ubuntu/+source/udev/+bug/89693, but its on gutsy, of course.

Revision history for this message
Bruce Miller (brm0423) wrote :

I have exactly the same problem on a freshly-upgraded Kubuntu Gutsy Gibbon Tribe 4 (upgraded today from Feisty using update-manager -d).

I also have the same device-mapper message as bug #89693 scrolling past. The device number on my error message, however, is 254 rather than 253:
device-mapper: table: 254:1: linear: dm-linear: Device lookup failed

Revision history for this message
Bruce Miller (brm0423) wrote :

Further to my comment of two days ago, I have since noted that, in these conditions, udevd does not recognize the addition of a hot-pluggable device to the system. The case in point is powering on a Plextor PX-716UF DVD writer which is connected through a USB connection.

Not even restarting udevd will convince it to recognize the Plextor. A full system reboot does, however, allow me to see and to use the drive.

Revision history for this message
Christopher Lees (leezer3) wrote :

I'm seeing this too.
I believe its related to DMRaid & drives detection- I'm getting massive CPU usage and constant hard-drive churning.
I say this because my fstab was re-written on the upgrade to Gutsy (From Feisty- Could this be part of the problem?), to exclude everything except the boot drive. Leaving the fstab unaltered keeps udevd in check, but obviously I need my other drives.
The other two drives are a 320gb LVM which I use for keeping my media on and a 160gb IDE drive- I would guess that theres a loop somewhere in the detection of some types of special drivers.

Revision history for this message
Christopher Lees (leezer3) wrote :

Sorry, error in my case is:
device-mapper: table: 254:6: linear: dm-linear: Device lookup failed

Revision history for this message
Doomi (dasletzteabendmahl) wrote :

my errror is:
device-mapper: table: 254:0: linear: dm-linear: Device lookup failed

and i upgrade from feisty

sorry for the few information i gave in my first post.

Revision history for this message
Bruce Miller (brm0423) wrote :

Since my comments of 2007-08-18, I have wiped the version of gutsy which I had upgraded from Feisty and replaced it with a fresh install of Kubuntu Gutsy Tribe 4. I had some bother installing the nVidia restricted driver and had to reboot the system twice. This udevd problem never appeared. Late last night, however, I added the Gnome and XFCE desktops using "dpkg --set-selections" and the list of packages from the previous setup. This udevd problem is usually evident, not just on system startup, but also on shutdown. I saw nothing, however, when I shut down the system in the wee hours of this morning.

When I, however, next started the system later this morning, the udevd problem was back again. I therefore expect something within the >1400 packages which were installed late last night has triggered the problem again. While I recognize, this is only partially reducing the size of the haystack in which to hunt for the needle, I attach all the package installation logs from yesterday. The key log is the dselect log from 20070821T0141 EDT (GMT -0400).

Revision history for this message
Christopher Lees (leezer3) wrote :

The kernel upgrade to 2.6.22 springs out to me from those logs. Have a look through this bug-
https://bugs.launchpad.net/ubuntu/+source/evms/+bug/115616

Looks to me like a dupe of the above?

Revision history for this message
Bruce Miller (brm0423) wrote :

@Christopher Lees
Thank you for helping me to track down this bug. Yes, the solution was in the logs, but no, it was not the upgrade to 2.6.22. Had it been, the bug would have bitten on the first reboot with the new kernel and not after several reboots. The culprit --- and Bug #115616 demonstrates this amply --- was the package "evms" (Enterprise Volume Management System) which was installed in the download which began to be logged at 0131 local time. It was only on a restart after this download that the bug bit. I have removed the package evms, restarted the machine and am mercifully free once more of this debilitating bug :-)

I am not going to mark this bug report immediately as a dupe of Bug #115616. I did not open this bug and it is only right to give other subscribers a chance to state whether their problems are also caused by the evms package. If anyone reports an ongoing problem BUT does not have evms on their system, then we have a separate problem requiring a separate bug report.

Revision history for this message
Matthew East (mdke) wrote :

Yes, looks like a duplicate. Marking as such

Revision history for this message
Jorge Juan (jjchico) wrote :

Same problem in my upgrade on 2007-09-20. Solved by uninstalling evms, so I am pretty sure it is a duplicate of #115616.

Revision history for this message
Ayan (ayanm) wrote :

Same with me. Uninstalling evm solves the problem.

--Ayan

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.