initctl eating cpu on mako #181

Bug #1281061 reported by Alan Pope 🍺🐧🐱 🦄
12
This bug affects 2 people
Affects Status Importance Assigned to Milestone
upstart (Ubuntu)
Expired
Undecided
Unassigned

Bug Description

On Ubuntu image #181 on mako my phone is hot. I checked top and found initctl eating 96% of the cpu..

This may be related to bug 1281057 as my phone refuses to lock.

Unfortunately I couldn't get the bug report done on the phone because the battery died (unsurprisingly) mid way through reporting it. Managed to get this output from top though.

top - 10:34:05 up 1 day, 20:23, 1 user, load average: 4.19, 4.14, 3.82
Tasks: 244 total, 2 running, 242 sleeping, 0 stopped, 0 zombie
%Cpu(s): 49.1 us, 5.4 sy, 0.0 ni, 45.1 id, 0.4 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem: 1916236 total, 1855748 used, 60488 free, 18420 buffers
KiB Swap: 524284 total, 115420 used, 408864 free. 196160 cached Mem

  PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
 1981 phablet 20 0 679256 665780 1160 S 0.3 34.7 33:23.31 init
  586 message+ 20 0 280608 276924 712 S 0.0 14.5 26:18.81 dbus-daemon
    1 root 20 0 210636 206636 1056 S 0.0 10.8 12:16.86 init
 2211 phablet 20 0 553308 127668 77840 S 85.7 6.7 49:53.31 unity8
  598 root 20 0 88792 86892 584 R 96.1 4.5 433:39.94 initctl
 2318 phablet 20 0 296952 40432 33060 S 0.0 2.1 0:26.02 maliit-ser+
 1903 root 20 0 241404 27880 20500 S 13.0 1.5 10:27.68 unity-syst+
 9631 phablet 20 0 216224 19924 14828 S 0.0 1.0 0:01.78 telephony-+
 2350 phablet 20 0 55260 19540 2776 S 0.0 1.0 1:28.08 indicator-+
 1052 root 20 0 52248 19244 2252 S 0.0 1.0 22:39.85 upowerd
 2080 phablet 20 0 47300 12604 3396 S 0.0 0.7 1:43.27 mission-co+
 9368 phablet 20 0 80832 11872 9988 S 0.0 0.6 0:00.61 telephony-+
 1062 root 20 0 42916 11548 2544 S 0.0 0.6 84:53.39 polkitd
 2205 phablet 20 0 96980 11464 9280 S 0.0 0.6 0:05.66 hud-service
 9360 phablet 20 0 52164 9592 8120 S 0.0 0.5 0:00.21 telephony-+
 2234 phablet 20 0 134336 8716 5288 S 0.0 0.5 0:16.43 address-bo+
 2093 phablet 9 -11 203588 7304 4964 S 0.0 0.4 5:37.30 pulseaudio

Tags: avengers
Revision history for this message
James Hunt (jamesodhunt) wrote :

Hi Alan,

Is this reproducable? If so, a 'ps -efwww' would be useful so we can see what is actually running initctl and what it is trying to do. Also, if you can 'strace -o /tmp/initctl.strace -s 1024 -fFv -p pid-of-initctl' and attach /tmp/initctl.strace to the bug that would be very useful.

Changed in upstart (Ubuntu):
status: New → Incomplete
Revision history for this message
Alan Pope 🍺🐧🐱 🦄 (popey) wrote :

It seems to happen when left on for a while. This is a situation which only occurs over the weekend because I'm always rebooting my phone for testing new images during the week. I will try and capture more info next time.

Revision history for this message
Fabian Herb (fherb) wrote :

Happening right now for me. I have plugged in the phone to my computer (not the wall outlet adapter) but the phone is not charging, probably because the cpu takes the 500 mA alone. Among the top CPU users are initctl, powerd and unity8. Here's the output of ps -efwww, hope that helps.

Revision history for this message
Fabian Herb (fherb) wrote :

strace: command not found, sorry.

Even after killing initctl, unity8, upowerd and unity-system-co continue to take up excessive amounts of CPU time.

System Settings says I have 14.04 (r237), last updated on 2014-03-15.

Revision history for this message
Launchpad Janitor (janitor) wrote :

[Expired for upstart (Ubuntu) because there has been no activity for 60 days.]

Changed in upstart (Ubuntu):
status: Incomplete → Expired
Revision history for this message
Fabian Herb (fherb) wrote :

Happened again today.

Ubuntu 14.10 (r243), last updated 9/17/14.

Revision history for this message
James Hunt (jamesodhunt) wrote :

@Fabian - thanks for reporting but we need a lot more detail before we can get to the bottom of this (ps output, strace [you need to install it in dev mode], etc).

Revision history for this message
Fabian Herb (fherb) wrote :

I haven't experienced this issue on recent 14.09 images. However, now I have a proces called "init" (not the one with PID 1) that takes 200 megs of RAM and constantly about 15% of CPU:

top - 22:10:19 up 2 days, 11:13, 2 users, load average: 2,88, 3,48, 3,44
Tasks: 304 total, 3 running, 291 sleeping, 10 stopped, 0 zombie
%Cpu(s): 37,1 us, 62,4 sy, 0,0 ni, 0,0 id, 0,0 wa, 0,0 hi, 0,5 si, 0,0 st
KiB Mem: 1878572 total, 1811868 used, 66704 free, 143000 buffers
KiB Swap: 32764 total, 1696 used, 31068 free. 477768 cached Mem

  PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
 1453 phablet 20 0 215936 210984 1216 R 14,5 11,2 282:48.96 init
  766 message+ 20 0 18100 15588 756 S 4,2 0,8 139:55.43 dbus-daemon
 2765 phablet 20 0 552016 158796 59224 R 3,9 8,5 87:28.12 unity8
 1139 root 0 -20 7572 552 400 S 2,3 0,0 67:09.50 mpdecision
 1574 phablet 20 0 5344 2428 748 S 2,3 0,1 69:27.50 dbus-daemon
  988 root 20 0 9744 1260 796 S 1,9 0,1 47:23.86 systemd-udevd
    1 root 20 0 4412 2884 1112 S 1,3 0,2 31:58.40 init
 1691 phablet 20 0 5608 448 260 S 1,3 0,0 35:44.75 upstart-dbus-br
26167 phablet 20 0 6492 1396 860 R 1,3 0,1 0:00.42 top

This time, I have strace installed and recorded a few seconds of its output.

Do you think this is related, or should I open a new bug for this?

Revision history for this message
James Hunt (jamesodhunt) wrote :

Hi Fabian,

The problem you are having seem to be related to mtp-server respawning continuously. This may not be the same issue as originally reported, so please open a new bug against the mtp-server package.

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.