"Failed to stop xbmc-service.service: Unit xbmc-service.service not loaded." is the response i get, when i try to execute "systemctl stop xbmc-service".
But without the "-service" prefix it works.
So, i did the steps you mentioned, and the load is the same.
Roundabout 70% right now, 5 minutes after bootup.
Edit: this is the freshest log i have, maybe its useful:
http://xbmclogs.com/show.php?id=204351
(2014-05-18, 13:22)ken_s Wrote: [ -> ]Seeing the same as RaVeNV1, cpu 70-80% with no activity.
This is with a new xbmc folder.
You're measuring this with top (or maybe
bcmstat.sh)?
Because right now, with stock Confluence, I'm seeing under 20% (actually, with rbphdmi having switched off the HDMI, I'm currently seeing a total load of 14-15%).
(2014-05-18, 13:23)RaVeNV1 Wrote: [ -> ]But without the "-service" prefix it works.
Sorry duff instructions, wrote it from memory - yes, remove the -service - have updated the original post.
top says: 390 1 root R 306m 82.0 0 73.5 /usr/lib/xbmc/xbmc.bin --standalone -fs --lircdev /run/lirc/lircd
CPU load over 70%, but it switches between 60% and 80%, as i mentioned.
Can you paste the output from "pstree $(pidof xbmc.bin)" - it won't help much, but it's difficult to "see" what xbmc is doing unless it outputs something... pstree should show the threads belonging to the xbmc.bin process.
OpenELEC:~ # pstree $(pidof xbmc.bin)
xbmc.bin-+-2*[{AESink}]
|-{ActiveAE}
|-{AirPlayServer}
|-4*[{CECAdapter}]
|-{CRBPWorker}
|-{EventServer}
|-{HCEC Notify}
|-{HDispmanx Notif}
|-{HTV Notify}
|-{ILCS_HOST}
|-{JobWorker}
|-4*[{LanguageInvoker}]
|-{PeripBusCEC}
|-{PeripBusUSBUdev}
|-{TCPServer}
|-{VCHIQ completio}
|-2*[{libmicrohttpd}]
`-2*[{xbmc.bin}]
The only obvious difference between your threads and my threads is you have "4*[{CECAdapter}]" threads - I've disabled CEC in XBMC as my monitor has no CEC support.
What happens if you turn off/disable CEC - any change?
Edit: I've enabled the CEC adapter on my Pi and after restarting XBMC I also have the "4*[{CECAdapter}]" threads but after a few minutes load is hovering around 20% so that's not likely to be the reason unless libcec is going bananas on your system for some reason.
It remains the same.
xbmc.bin-+-2*[{AESink}]
|-{ActiveAE}
|-{AirPlayServer}
|-{CRBPWorker}
|-{EventServer}
|-{HCEC Notify}
|-{HDispmanx Notif}
|-{HTV Notify}
|-{ILCS_HOST}
|-{JobWorker}
|-4*[{LanguageInvoker}]
|-{PeripBusCEC}
|-{PeripBusUSBUdev}
|-{TCPServer}
|-{VCHIQ completio}
|-2*[{libmicrohttpd}]
`-2*[{xbmc.bin}]
Edit: Suddenly, there's no disk image of your build, i would try it with a clean and fresh install of your build.
I installed OE 4.0.1 first via disk image, and then updated to your build via OpenELEC Dev Update.
Maybe there was a failure, i dont know, i'm not familiar with this.
Will show the CPU usage of each thread.
(2014-05-18, 14:34)popcornmix Wrote: [ -> ]Will show the CPU usage of each thread.
Sadly, not on OpenELEC...
(maybe I can build "real" top somehow...)
OpenELEC:~ # top -H
top: invalid option -- 'H'
BusyBox v1.22.1 (2014-05-16 19:17:29 BST) multi-call binary.
Usage: top [-b] [-nCOUNT] [-dSECONDS] [-m]
Provide a view of process activity in real time.
Read the status of all processes from /proc each SECONDS
and display a screenful of them.
Keys:
N/M/P/T: show CPU usage, sort by pid/mem/cpu/time
S: show memory
R: reverse sort
H: toggle threads, 1: toggle SMP
Q,^C: exit
Options:
-b Batch mode
-n N Exit after N iterations
-d N Delay between updates
-m Same as 's' key
Code:
tc/eventlircd.d --socket=/run/lirc/lircd --release=_UP
358 1 root S 2092 0.5 0 0.0 /usr/bin/rpcbind
51 2 root SW< 0 0.0 0 0.0 [VCHIQr-0]
68 2 root SW 0 0.0 0 0.0 [mmcqd/0]
9 2 root SW 0 0.0 0 0.0 [rcu_preempt]
50 2 root SW< 0 0.0 0 0.0 [VCHIQ-0]
93 2 root SW< 0 0.0 0 0.0 [loop0]
21 2 root SW 0 0.0 0 0.0 [khubd]
87 2 root SW 0 0.0 0 0.0 [ext4lazyinit]
10 2 root SW 0 0.0 0 0.0 [rcuop/0]
183 2 root SW 0 0.0 0 0.0 [kworker/0:3]
3 2 root SW 0 0.0 0 0.0 [ksoftirqd/0]
85 2 root SW 0 0.0 0 0.0 [jbd2/mmcblk0p2-]
15 2 root SW 0 0.0 0 0.0 [watchdog/0]
61 2 root SW 0 0.0 0 0.0 [kworker/u2:2]
6 2 root SW 0 0.0 0 0.0 [kworker/u2:0]
22 2 root SW 0 0.0 0 0.0 [kworker/0:1]
2 0 root SW 0 0.0 0 0.0 [kthreadd]
4 2 root SW 0 0.0 0 0.0 [kworker/0:0]
5 2 root SW< 0 0.0 0 0.0 [kworker/0:0H]
7 2 root SW 0 0.0 0 0.0 [rcuc/0]
8 2 root SW 0 0.0 0 0.0 [rcub/0]
11 2 root SW 0 0.0 0 0.0 [rcu_sched]
12 2 root SW 0 0.0 0 0.0 [rcuos/0]
13 2 root SW 0 0.0 0 0.0 [rcu_bh]
14 2 root SW 0 0.0 0 0.0 [rcuob/0]
16 2 root SW< 0 0.0 0 0.0 [khelper]
17 2 root SW 0 0.0 0 0.0 [kdevtmpfs]
18 2 root SW< 0 0.0 0 0.0 [writeback]
19 2 root SW< 0 0.0 0 0.0 [bioset]
20 2 root SW< 0 0.0 0 0.0 [kblockd]
23 2 root SW< 0 0.0 0 0.0 [rpciod]
26 2 root SW 0 0.0 0 0.0 [kswapd0]
27 2 root SW 0 0.0 0 0.0 [fsnotify_mark]
28 2 root SW< 0 0.0 0 0.0 [nfsiod]
29 2 root SW< 0 0.0 0 0.0 [cifsiod]
30 2 root SW 0 0.0 0 0.0 [jfsIO]
31 2 root SW 0 0.0 0 0.0 [jfsCommit]
32 2 root SW 0 0.0 0 0.0 [jfsSync]
33 2 root SW< 0 0.0 0 0.0 [xfsalloc]
34 2 root SW< 0 0.0 0 0.0 [xfs_mru_cache]
35 2 root SW< 0 0.0 0 0.0 [xfslogd]
36 2 root SW< 0 0.0 0 0.0 [crypto]
52 2 root SW< 0 0.0 0 0.0 [VCHIQs-0]
53 2 root SW< 0 0.0 0 0.0 [iscsi_eh]
54 2 root SW< 0 0.0 0 0.0 [bcm2708_spi.0]
55 2 root SW< 0 0.0 0 0.0 [dwc_otg]
56 2 root SW< 0 0.0 0 0.0 [DWC Notificatio]
57 2 root SW 0 0.0 0 0.0 [kworker/u2:1]
58 2 root SW< 0 0.0 0 0.0 [ipv6_addrconf]
59 2 root SW< 0 0.0 0 0.0 [bioset]
60 2 root SW< 0 0.0 0 0.0 [deferwq]
86 2 root SW< 0 0.0 0 0.0 [ext4-rsv-conver]
135 2 root SW 0 0.0 0 0.0 [kworker/0:2]
255 2 root SW 0 0.0 0 0.0 [kworker/u2:3]
401 2 root SW 0 0.0 0 0.0 [VCHIQka-0]
OK, this is just a hack but should work... I've grabbed top from Raspbian:
Code:
cd /storage
mkdir fulltop
cd fulltop
wget http://www.nmacleod.com/public/fulltop.tar.gz -qO - | tar xzf -
Now "/storage/fulltop/fulltop.sh -H" should work on OpenELEC.
Fellas, please use [code] tags when posting formatted results... makes reading that bit easier!
Code:
OpenELEC:~/fulltop # /storage/fulltop/fulltop.sh -H
top - 12:57:32 up 9 min, 1 user, load average: 1.02, 1.00, 0.60
Threads: 100 total, 2 running, 98 sleeping, 0 stopped, 0 zombie
%Cpu(s): 55.0 us, 7.3 sy, 1.7 ni, 36.1 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem: 381356 total, 356820 used, 24536 free, 34420 buffers
KiB Swap: 0 total, 0 used, 0 free, 275284 cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
392 root 20 0 303m 39m 16m R 52.9 10.7 5:03.00 xbmc.bin
410 root 21 1 303m 39m 16m S 2.9 10.7 0:09.88 PeripBusCEC
51 root 1 -19 0 0 0 S 2.3 0.0 0:13.40 VCHIQr-0
468 root 20 0 2880 1236 924 R 1.6 0.3 0:00.29 top
409 root 20 0 303m 39m 16m S 1.3 10.7 0:11.87 CECAdapter
402 root 19 -1 303m 39m 16m S 1.0 10.7 0:04.10 VCHIQ completio
441 root 20 0 303m 39m 16m S 0.6 10.7 0:03.25 LanguageInvoker
9 root 20 0 0 0 0 S 0.3 0.0 0:01.83 rcu_preempt
50 root 1 -19 0 0 0 S 0.3 0.0 0:02.86 VCHIQ-0
183 root 20 0 0 0 0 S 0.3 0.0 0:00.29 kworker/0:3
202 avahi 20 0 2656 1500 1308 S 0.3 0.4 0:00.61 avahi-daemon
1 root 20 0 3744 2408 1544 S 0.0 0.6 0:02.67 systemd
2 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kthreadd
3 root 20 0 0 0 0 S 0.0 0.0 0:00.35 ksoftirqd/0
4 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kworker/0:0
5 root 0 -20 0 0 0 S 0.0 0.0 0:00.00 kworker/0:0H
7 root -2 0 0 0 0 S 0.0 0.0 0:00.00 rcuc/0
Thats the output.
I must mention, that I formatted the SD, installed OE 4.0.1 and then updated to your build.
So no addons, scripts etc. installed anymore. Just your "naked" build.
[code] tags would make that output a lot more readable... just saying (again).
and rather use pastebin.com to not dump this on our forum
Try:
Code:
/storage/fulltop/fulltop.sh -H -p $(pidof xbmc.bin)
and while top is running, press T (capital T!) so that processes and threads are sorted by cumulative time - this should result in the "busy" thread bubbling to the top...
(2014-05-18, 15:10)Martijn Wrote: [ -> ]and rather use pastebin.com to not dump this on our forum
For a relatively small number of lines I don't see the harm. For longer logs, you're absolutely correct.