RPi-3+ TW 5.1.7 high ambient background load

English Install/Boot/Login
#1

I just updated 5.1.5 → 5.1.7, but I’ve seen the same issue on both installs: high ambient background load with no user tasks active:

hdtodd@Pi-6:~> uname -a
Linux Pi-6 5.1.7-1-default #1 SMP Tue Jun 4 07:56:54 UTC 2019 (55f2451) aarch64 aarch64 aarch64 GNU/Linux
hdtodd@Pi-6:~> uptime
 16:11:36  up   0:13,  1 user,  load average: 4.00, 3.82, 2.49
hdtodd@Pi-6:~>

If I run “top”, I see no culprit – nothing using more than 1% CPU and most using nothing. But performance is sluggish, so I think it’s more than just a load-calculation error.

Anyone else seeing this? Any idea what’s causing it and how to fix it?

#2

Hi
What about iostat, vmstat, iotop… not using swap?


iostat -x 5
vmstat 1
ps awwlx --sort=vsz (sorts by virtual sizes, low to high)
iotop

In your output, the 15 min time has not passed yet…

Install glances (python3-Glances), this should give the best overview but will add load…

What about disk i/o scheduler?


cat /sys/block/<disk_eg_mmcblk0>/queue/scheduler
#3

Thanks for the quick response, Malcolm!

After an hour …

Pi-6:/home/hdtodd # uptime
 17:08:41  up   1:11,  1 user,  load average: 4.00, 4.00, 4.00

Virtual sizes …

Pi-6:/home/hdtodd # ps awwlx --sort=-vsz
F   UID   PID  PPID PRI  NI    VSZ   RSS WCHAN  STAT TTY        TIME COMMAND
5   474   559     1  20   0 721304  3828 ep_pol Ssl  ?          0:00 /usr/sbin/nscd
5   502  1952  1951  20   0 171992  3212 do_sig S    ?          0:00 (sd-pam)
4     0     1     0  20   0 166620  9928 ep_pol Ss   ?          0:22 /usr/lib/systemd/systemd --switched-root --system --deserialize 32
4     0   383     1  20   0  86688  1324 do_sel Ss   ?          0:00 /usr/sbin/lvmetad -f
5   472  1600     1  20   0  78148  2104 do_sel S    ?          0:00 /usr/sbin/chronyd
4     0   367     1  20   0  49624 16336 ep_pol Ss   ?          0:02 /usr/lib/systemd/systemd-journald
4     0   394     1  20   0  29812  8176 ep_pol Ss   ?          0:03 /usr/lib/systemd/systemd-udevd
4     0  1591     1  20   0  29492  7816 ep_pol Ss   ?          0:00 /usr/sbin/cupsd -l
4     0  2113  1958  20   0  17708  5944 do_sys S    pts/0      0:00 sudo -s
5     0   530     1  16  -4  17672  1640 ep_pol S<sl ?          0:00 /sbin/auditd
4   502  1951     1  20   0  16868  8132 ep_pol Ss   ?          0:00 /usr/lib/systemd/systemd --user
4     0  1947  1634  20   0  16212  7532 do_sys Ss   ?          0:00 sshd: hdtodd [priv]
5   502  1957  1947  20   0  16212  4508 -      S    ?          0:00 sshd: hdtodd@pts/0
4     0   614     1  20   0  15092  5880 ep_pol Ss   ?          0:01 /usr/lib/systemd/systemd-logind
4   499   550     1  20   0  13548  5860 ep_pol Ss   ?          0:02 /usr/bin/dbus-daemon --system --address=systemd: --nofork --nopidfile --systemd-activation --syslog-only
4     0  1634     1  20   0  12784  6504 do_sel Ss   ?          0:00 /usr/sbin/sshd -D
4     0   664     1  20   0  12096  6944 do_sel Ss   ?          0:00 /usr/sbin/wpa_supplicant -c /etc/wpa_supplicant/wpa_supplicant.conf -u -t -f /var/log/wpa_supplicant.log
0   502  1958  1957  20   0  10396  6864 do_wai Ss   pts/0      0:00 -bash
4     0  2114  2113  20   0  10340  6808 do_wai S    pts/0      0:00 /bin/bash
4     0   660     1  20   0   9860  6036 do_sys SLs  ?          0:00 /usr/sbin/wickedd-nanny --systemd --foreground
4     0   655     1  20   0   9816  6252 do_sys SLs  ?          0:00 /usr/sbin/wickedd --systemd --foreground
4     0   551     1  20   0   9700  6096 do_sys SLs  ?          0:00 /usr/lib/wicked/bin/wickedd-dhcp6 --systemd --foreground
4     0   552     1  20   0   9700  6132 do_sys SLs  ?          0:00 /usr/lib/wicked/bin/wickedd-dhcp4 --systemd --foreground
4     0   556     1  20   0   9696  5852 do_sys SLs  ?          0:00 /usr/lib/wicked/bin/wickedd-auto4 --systemd --foreground
4     0  1658     1  20   0   7384  2444 hrtime Ss   ?          0:00 /usr/sbin/cron -n
0     0  2236  2114  20   0   7184  2884 -      R+   pts/0      0:00 ps awwlx --sort=-vsz
4     0  1597     1  20   0   2560   464 do_sel Ss+  tty1       0:00 /sbin/agetty -o -p -- \u --noclear tty1 linux
1     0     2     0  20   0      0     0 kthrea S    ?          0:00 [kthreadd]
(rest have VSZ = 0)

VMSTAT:

Pi-6:/home/hdtodd # vmstat 1
procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
 r  b   swpd   free   buff  cache   si   so    bi    bo   in   cs us sy id wa st
 0  0      0 525328  50456 291296    0    0    21     3 2046   34  1  0 99  0  0
 0  0      0 525328  50456 291296    0    0     0     0 8158  144  0  0 100  0  0
 0  0      0 525328  50456 291296    0    0     0     0 8329  316  0  0 100  0  0 ...

Not using MMC, using SDA:

Pi-6:/home/hdtodd # cat /sys/block/sda/queue/scheduler
[none] kyber bfq

IOTOP shows no activity.

And then, finally:

Pi-6:/home/hdtodd # uptime
 17:22:48  up   1:25,  1 user,  load average: 4.21, 4.14, 4.05
Pi-6:/home/hdtodd # iostat -x 5
Linux 5.1.7-1-default (Pi-6)     06/16/19     _aarch64_    (4 CPU)

avg-cpu:  %user   %nice %system %iowait  %steal   %idle
           0.75    0.26    0.38    0.04    0.00   98.56

Device            r/s     w/s     rkB/s     wkB/s   rrqm/s   wrqm/s  %rrqm  %wrqm r_await w_await aqu-sz rareq-sz wareq-sz  svctm  %util
sda              6.11    0.78     75.53     17.44     0.00     0.58   0.00  42.64    0.86    1.15   0.00    12.37    22.22   1.46   1.01


avg-cpu:  %user   %nice %system %iowait  %steal   %idle
           0.00    0.00    0.00    0.00    0.00  100.00

Device            r/s     w/s     rkB/s     wkB/s   rrqm/s   wrqm/s  %rrqm  %wrqm r_await w_await aqu-sz rareq-sz wareq-sz  svctm  %util
sda              0.00    0.00      0.00      0.00     0.00     0.00   0.00   0.00    0.00    0.00   0.00     0.00     0.00   0.00   0.00

Note 100% idle, but load >4 (so maybe it really is the load calculation!), but the system is very sluggish.

If I’m the only one seeing this, don’t worry about tracking it down: might get fixed in the next release. It just seemed odd, and I couldn’t find any sign of what was causing the problem … thought it might be an issue in the kernel.

David

#4

Hi
I’m guessing it’s a bug with uptime and load calculation with cpu’s…

I get


uptime
19:09:01  up   1:12,  1 user,  load average: 4.00, 4.00, 4.00

Which in IMHO should all be zero…

#5

Yes, I’m concluding the same thing. Over time, the “uptime” reported load levels out at 4 (on a quad-core), but idle is reported 100% by other tools. Still seems sluggish (going in via SSH), but I agree that it’s a bug in “uptime”.