Created attachment 839613 [details]
load graph
When I updated the new kernel I experienced weird load behaviour. The load kept
climbing and within 2 days I saw a load of 90 while no processes were really
active. One of the processes (a KVM virtual machine) which had some CPU usage
in top I killed but the load didn't really go down.
--
top - 07:48:10 up 1 day, 15:13, 8 users, load average: 89.73, 89.36, 88.64
Tasks: 371 total, 2 running, 369 sleeping, 0 stopped, 0 zombie
%Cpu(s): 4.4 us, 0.7 sy, 0.8 ni, 93.6 id, 0.5 wa, 0.0 hi, 0.1 si, 0.0 st
--
My process list without inactive processes was this:
--
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
6535 qemu 20 0 12.781g 8.198g 24108 S 25.17 26.11 277:43.92
qemu-system-x86
5613 qemu 20 0 2521260 1.689g 23968 S 3.311 5.381 135:35.22
qemu-system-x86
8066 qemu 20 0 3088768 2.190g 23936 S 2.649 6.977 49:15.60
qemu-system-x86
7859 qemu 20 0 3037224 2.188g 24064 S 1.987 6.969 56:20.74
qemu-system-x86
7979 qemu 20 0 1985716 1.189g 24116 S 1.656 3.789 187:19.17
qemu-system-x86
3318 icinga 20 0 2964888 36880 18684 S 0.993 0.112 13:03.10 icinga2
2486 influxdb 20 0 6815940 451780 49600 S 0.662 1.372 60:54.26 influxd
2498 redis 20 0 91632 24292 4716 S 0.662 0.074 13:44.69
redis-server
3532 postgres 20 0 1204624 46252 42568 S 0.662 0.141 5:11.47 postgres
29529 root 20 0 50272 4380 3452 R 0.662 0.013 0:00.08 top
29568 root 20 0 130144 4104 2736 S 0.662 0.012 0:00.02 sshd
2584 root 20 0 925036 60684 47496 S 0.331 0.184 2:40.19
fail2ban-server
2750 root 20 0 1275024 68076 16780 S 0.331 0.207 2:48.57
salt-minion
3048 named 20 0 872640 236812 7388 S 0.331 0.719 5:10.92 named
3415 icinga 20 0 692592 6660 4732 S 0.331 0.020 1:42.88 icinga2
26938 root 20 0 0 0 0 I 0.331 0.000 0:00.14
kworker/6:0-events
27676 root 20 0 0 0 0 I 0.331 0.000 0:00.08
kworker/u16:8-events_unbound
--
Not really troublesome I would say. I have attached a graph of the load of the
machine over the last 2 days. It clearly shows the moment the machine was
rebooted. I reverted back to the GM kernel (5.3.18-lp152.19-default) and had no
issues with the load anymore.
I am seeing this on a hardware machine (Intel(R) Atom(TM) CPU C3758 @ 2.20GHz).
Virtual machines don't seem to be that impacted. I had a load of 3 on one VM
with the 5.3.18-lp152.20.7-default kernel and 0,06 with the
5.3.18-lp152.19-default kernel. Not as high as the hardware machine but there
is something different.
Please let me know what more information you need to diagnose this issue.
Thanks in advance.