udevd taking 100% CPU

[devr]

I have a pogoplug pro, ran pacman -Syu this morning, rebooted. Now udevd is taking 100% of cpu. Don't see anything useful in /var/log/, running the command "udevadm monitor" doesn't show anything.

From my pacman -Syu, this is the log:
$this->bbcode_second_pass_code('', '[2011-05-19 09:01] Running 'pacman -Syu'
[2011-05-19 09:01] synchronizing package lists
[2011-05-19 09:01] starting full system upgrade
[2011-05-19 09:02] Generating locales...
[2011-05-19 09:03] en_US.UTF-8...cannot map archive header: Invalid argument
[2011-05-19 09:03] upgraded glibc (2.13-4 -> 2.13-5)
[2011-05-19 09:03] upgraded attr (2.4.44-3 -> 2.4.46-1)
[2011-05-19 09:03] upgraded acl (2.2.49-2 -> 2.2.51-1)
[2011-05-19 09:03] upgraded binutils (2.21-4 -> 2.21-7)
[2011-05-19 09:03] upgraded udev (167-2 -> 168-1)
[2011-05-19 09:03] upgraded device-mapper (2.02.84-1 -> 2.02.85-1)
[2011-05-19 09:03] upgraded cryptsetup (1.2.0-1 -> 1.3.0-1)
[2011-05-19 09:03] upgraded gcc-libs (4.5.2-6 -> 4.6.0-4)
[2011-05-19 09:03] upgraded gmp (5.0.1-2 -> 5.0.2-1)
[2011-05-19 09:03] upgraded kbd (1.15.2-1 -> 1.15.3-1)
[2011-05-19 09:03] Notice
[2011-05-19 09:03] If you have a SheevaPlug, you will have to copy
[2011-05-19 09:03] /boot/uImage to the proper location, e.g. /dev/mmcblk0p1
[2011-05-19 09:03] after removing the old one.
[2011-05-19 09:03]
[2011-05-19 09:03] If you boot from USB, you do not need to do anything.
[2011-05-19 09:03]
[2011-05-19 09:03] This package is for all ARMv5te plug computers.
[2011-05-19 09:03] upgraded kernel26 (2.6.38.3-1.1 -> 2.6.38.6-2)
[2011-05-19 09:03] upgraded kernel26-headers (2.6.38.3-1.1 -> 2.6.38.6-2)
[2011-05-19 09:03] upgraded lvm2 (2.02.84-1 -> 2.02.85-1)
[2011-05-19 09:03] upgraded package-query (0.7-1 -> 0.8-1)
[2011-05-19 09:03] upgraded wget (1.12-5 -> 1.12-7)
[2011-05-19 09:07] Running 'pacman -S aoetools'
[2011-05-19 09:07] installed aoetools (32-1)
')

Did some quick searching on the internet, saw it could be a kernel issue.. but nothing certain, figured I would post here before digging more.

Thanks,

[kmihelich]

There's a udev update to the latest version that just dropped in ABS rsync, -Syu in about an hour to get it, reboot, and see if that fixes your problem.

[slycat]

any luck devr?

I rebooted my plug after backing up the rootfs and got the same issue. At first I thought it was my fstab definitions but i've commented out all my user-defined points. Also ran a pacman -Syu but still got the udevd running 100%. For the time being I've booted my plug, mounted my drives, and then killall udevd. After that I can run my apps without fail, but its just concerning why udev is running like an F40PH on a friday.

[devr]

I had some issues that ended up me accidently blowing away my plug. When doing the install I kept running in to the locale (reported in forum) issue. Was quite a coincidence, and I thought the locale was due to my earlier issues... End story.. reinstalled OS, dealt with locale issue, ran a pacman -Syu this morning.. udevd is still taking 100% cpu. Current fix is that I just killall udevd.

[slycat]

If anyone still gets this issue, run an update. A new udev package has been released and my plug no longer exhibits this power load.

[devr]

I just ran the update... a little better, but not by much.

Now udevd takes ~30% cpu and syslog-ng takes ~99%. the udevd process is logging like crazy, causing huge /var/log files (daemon.log, errors.log, everything.log). Log entries are:
$this->bbcode_second_pass_code('', 'Jun 2 09:56:18 blackie udevd[361]: unable to receive ctrl connection: Function not implemented
Jun 2 09:56:18 blackie udevd[361]: unable to receive ctrl connection: Function not implemented
Jun 2 09:56:18 blackie udevd[361]: unable to receive ctrl connection: Function not implemented
')

At work at the moment, so not able to google that and see if it is a configuration issue or what. Quick fix until i can poke around was "killall -9 udevd"

[WarheadsSE]

If you grabbed the latest udev, and you're running on the OXNAS kernel , try pacman -S udev-compat

[devr]

$this->bbcode_second_pass_code('', 'if you grabbed the latest udev, and you're running on the OXNAS kernel , try pacman -S udev-compat')

I did, I am, I did. No real help. After installing udev-compat, rebooted and the plug came back up doing the same thing.

to reconfirm, plug info is as follows:
$this->bbcode_second_pass_code('', ']# uname -a ; cat /proc/cpuinfo ; pacman -Q udev udev-compat
Linux blackie 2.6.31.6_SMP_820 #101 SMP Tue Apr 12 20:34:36 UTC 2011 armv6l ARMv6-compatible processor rev 5 (v6l) Oxsemi NAS GNU/Linux
Processor : ARMv6-compatible processor rev 5 (v6l)
processor : 0
BogoMIPS : 279.34

processor : 1
BogoMIPS : 279.34

Features : swp half thumb fastmult edsp java
CPU implementer : 0x41
CPU architecture: 7
CPU variant : 0x0
CPU part : 0xb02
CPU revision : 5

Hardware : Oxsemi NAS
Revision : 0000
Serial : 0000000000000000

udev 171-1
udev-compat 171-1
')

[WarheadsSE]

It was worth a shot

[devr]

The udev install said:
$this->bbcode_second_pass_code('', '[2011-06-02 09:48] ATTENTION UDEV:
[2011-06-02 09:48] ---------------
[2011-06-02 09:48] Kernel 2.6.32 or newer is now required.
[2011-06-02 09:48] OSS emulation modules are not loaded by default, add to rc.co
nf if needed.
')

Is that the issue, since I am on 2.6.31.6_SMP_820 ?

Also, took a work break, and only similar issue I saw was the following.