[Bug?] Process usage, slow app response & possible battery drain

asked 2014-01-13 19:11:09 +0200

ranko_6 gravatar image

Hi,

Today, at work, I noticed slower app response. It took longer for them to start, and typing wasn't exactly the fastest either. But, since I was working, I didn't had chance to use it for more than few minutes so it didn't bother me. After work ended, I tried to text a friend, and it actually took me quite some time to do it.

Typing the code to unlock the phone took quite some time, sometimes it didn't even register that I pressed a number, typing a message too, though here it was with both text and numbers. Starting apps was slow too.

At that point, what I saw with "top" command, were 4 processes, each one taking about 45,6% of the processor, each one ran by user "nemo".

Since at that point I was still walking home, I figured I'd try to see if they finished doing whatever they were trying to do by the time I got home.

About 15 minutes later, still the same situation, so I decided to check what was occupying processor.

I ran "top" again, wrote down the PID's and decided to check which apps were they: (pid - cat /proc/{pid}/cmdline)

922 - /usr/bin/lipstick-plugindevtouch:/dev/input/event0:rotate=0-pluginevdevkeyboard:keymap=/usr/share/qt5/keymaps/boston.qmap--systemd

879 - /usr/bin/timed-qt5--systemd

933 - /usr/bin/msyncd

1139 - /usr/bin/contactsd

About two minutes into writing them down, two of them finally "finished", while two of them occupied around 89% (922 & 933). Few moments later, 933 finished and 922 was occupying about 92% of processor. Few moments later, 922 finished too and now my phone works "normal" again.

I have Sailfish OS Version 1.0.2.5.

Apps I have installed are following:

  • Maps
  • Calendar
  • Mail
  • Clock
  • Calculator
  • Documents
  • Notes
  • Media
  • Microsoft Exchange
  • Snow White
  • Tethering
  • chro.mono

I don't have Android support installed, nor any android apps. (Seemed annoying, most comments I read were that it either doesn't work, or works partially for most apps so I just figured I'd wait for native ones)

I didn't notice this behaviour before, so I'm not sure if it could be the reason, but difference between today and days before today was that I got tired of turning on/off mobile/wireless manually based on whether I'm in apartment or outside, so I turned both of them on.

Other difference was that I turned the touch screen vibration off since, well besides it probably draining a nice amount of battery by the time someone writes a normal message, it is annoying when it vibrates like I'm receiving ~20 messages when I'm trying to type "at work now, can't talk".

Thank you,

Ranko


Also, my phone just stopped responding to the point I had to reboot it... so I'm pretty sure this is some kind of a bug... guess I'm back at manually switching between mobile/internet and seeing if that helps...

edit retag flag offensive close delete

Comments

Same problem here, though I don't know how to check the usage like Ranko did. As I have not any ideas for solution, I have chosen to shut down my phone once in a day or two. That seems to help but isn't exactly ideal for long term.

statler ( 2014-01-13 19:25:56 +0200 )edit

I started to have similar problems today: just all of sudden my phone became sluggish and very unresponsive. Just waking up the device was really slow but reboot helped.

jpt ( 2014-01-13 22:13:24 +0200 )edit

This problem happens to me multiple times everyday. It seems it's triggered when starting to use mobile data. Sometimes the problem stops after a while and sometimes a reboot is needed. involved processes are always the same : connmand, system dbus, contactsd, msyncd, lipstick, timed-qt5

eber42 ( 2014-01-28 14:20:09 +0200 )edit

CPU is always at 100% and max speed (as shown by powertop). There is no significant I/O or swapping (vmstat). Funny experiment: killall -STOP connmand stops the high CPU usage (but it restarts after -CONT) seems like connmand is flooding dbus ? (but dbus-monitor shows nothing on both busses)

eber42 ( 2014-01-28 14:25:29 +0200 )edit

journalctl show nothing unusual (i think). I've got traces (strace, journalctl ...) available if someone wants to investigate And don't try the killall -STOP command above unless connected by USB + ssh

(and btw the comment box size is a bit small)

eber42 ( 2014-01-28 14:29:26 +0200 )edit