[gutsy] udevd hogs the processor and spams /var/log/messages

[Bogdan Butnaru]

Binary package hint: udev

Hello! I've updated from Feisty to Gutsy yesterday, and noticed a weird problem:

Right after boot, a "udevd" process (there are several in the System Monitor) keeps hogging the processor. (This means 100% pretty much all the time. However this is a dual-core machine, it remains responsive.)

I've looked in var/log, and at least "kern.log" and "messages" are spammed by ioctl errors (and I mean tens of megabytes). See below for snippets.

If I do a /etc/init.d/udev restart I get an "Loading additional hardware drivers..." and the scripts remains there (I killed it after a minute or so; the errant process kept hogging). However, "udev stop && udev start" worked, ie. it stopped the processor-hog and restarted udev; I still got an "error receiving uevent message: No buffer space available", but I suppose that's something wrong with a script.

After restarting udev, it doesn't hog the processor anymore nor does it spam the logs; everything seems to work OK until I reboot, when everything happens the same again.

This is an up-to-date Gutsy running on a Latitude D620 laptop (Centrino Duo processor).

$ uname -a
Linux cimer 2.6.22-6-generic #1 SMP Fri Jun 1 19:24:12 GMT 2007 i686 GNU/Linux

This is what /var/log/kern.log lists (megabytes of it):
Jun 8 14:19:47 cimer kernel: [ 543.856000] device-mapper: ioctl: error adding target to table
Jun 8 14:19:47 cimer kernel: [ 543.856000] device-mapper: table: 254:0: linear: dm-linear: Device lookup failed
Jun 8 14:19:47 cimer kernel: [ 543.856000] device-mapper: ioctl: error adding target to table
Jun 8 14:19:47 cimer kernel: [ 543.860000] device-mapper: table: 254:0: linear: dm-linear: Device lookup failed
Jun 8 14:19:47 cimer kernel: [ 543.860000] device-mapper: ioctl: error adding target to table
Jun 8 14:19:47 cimer kernel: [ 543.860000] device-mapper: table: 254:0: linear: dm-linear: Device lookup failed
Jun 8 14:19:47 cimer kernel: [ 543.860000] device-mapper: ioctl: error adding target to table
Jun 8 14:19:47 cimer kernel: [ 543.860000] device-mapper: table: 254:0: linear: dm-linear: Device lookup failed

And this is what /var/log/messages shows:
Jun 8 14:19:47 cimer kernel: [ 543.844000] device-mapper: ioctl: error adding target to table
Jun 8 14:19:47 cimer kernel: [ 543.848000] device-mapper: ioctl: error adding target to table
Jun 8 14:19:47 cimer last message repeated 6 times
Jun 8 14:19:47 cimer kernel: [ 543.852000] device-mapper: ioctl: error adding target to table
Jun 8 14:19:47 cimer last message repeated 7 times
Jun 8 14:19:47 cimer kernel: [ 543.856000] device-mapper: ioctl: error adding target to table
Jun 8 14:19:47 cimer last message repeated 8 times

[Justin Payne (neolithium82)]

I'm having the exact same problem with the identical kernel. Upgraded this morning on my Averatec 3270-EE1 Notebook and it seems to be absolutely mirrored to the above poster.

neo@ubuntu-dev:~$ uname -a
Linux ubuntu-dev 2.6.22-6-generic #1 SMP Fri Jun 1 19:24:12 GMT 2007 i686 GNU/Linux

Attached are parts from my kern.log and /messages

I haven't tried to boot into another kernel though, though I'll be trying that shortly. Perhaps I might need to simply recompile this existing one as well, but I'll have to test this stuff out later on.

Regards,
Justin

[François Tissandier (baloo)]

Same problem, on a Sony laptop.

[Dan Lawson (dlawson)]

I have the same from Feisty, with a custom built 2.6.21.5 kernel. Most of the .config options were defaults, with the exception of processor type (Set to Pentium M), and display adapters (enabled VESA).

[Andy Hauser (andy-ubuntu-bugzilla)]

Same after upgrade from feisty to gutsy. Laptop JVC MP-XP3210 (pentium III based celeron with 650 MHz).

[Andy Hauser (andy-ubuntu-bugzilla)]

As noted in http://forum.ubuntuusers.de/topic/96458/ removind evms fixed it for me.

[Stefan Schmid (schms)]

Hello! I've updated from Feisty to Gutsy yesterday, and noticed the same problem as described here.

[Octavio Alvarez (alvarezp)]

Same problem. Going back to kernel 2.6.20 makes the problem go away.

[Casey O'Donnell (chibiace)]

got it aswell. 2.6.20 does indeed make it go away :(

my laptop is dell inspiron 1300 with intel graphics and wireless.

[Alfredo Matas (amatas)]

I had the same problem but all of you can fix it removing all evms* packages of the system.

Regards.

[Luis Mondesi (lemsx1)]

Just wanted to say that I had the same problem and removing evms* packages fixed it for me as well.

Time for a real fix from Ubuntu developers

[Dan (dgar)]

Same here. dm-linear spam in logs and a load avg of 7+ just after bootup.

Removing 'evms-*' seems to make more sense than 'evms' which includes libdevmapper in the package selection.

[vilbara (vilmantas-baranauskas)]

I had the same problem but even worse. My "/home" is mounted on other partition and it was not mounted at all. I couldn't login with my normal account and I haven't set password for "root". CTRL+ALT+F1/Fn console was not usable because of device-mapper messages.

Removing "evms-*" helped.

[evets (itsmewhoelse)]

I had this problem when I first upgraded to Gutsy, then I removed "evms", and it went away.

Then we had an automatic upgrade come through, and it installed "evms-activate", and now I have the problem back. I can not remove or kill the evms-activate process, it's grayed out.

The only way to get my cpu usage back down, is to kill the udevd process. Not good. :)

The machine is a Toshiba R-15-S829 TABLET PC (convertible notebook). I'm running Gutsy 2.6.22-13-386 and Gnome 2.20.

[evets (itsmewhoelse)]

Just to be clear...evms-activate is the only evms I have.

I have attached an archive of all of my logs. 01 is while running, and 02 is after I killed udevd. There are erros generated, most are permissions denied, but evms and udevd are there.

[Marco Cimmino (cimmo)]

guys you are commenting a bug that was marked as duplicate, haven't read the line that appears ^^ when you comment?

"Remember, this bug report is a duplicate. Comment here only if you think the duplicate status is wrong."

[evets (itsmewhoelse)]

Thanks and sorry.