Need some help with cgroups configuration for qemu VMs

Help
,
21

I don’t think you need any of them.

1 Like
22

@ovi do you use parent/child setup?

Are there any lines like cgroup 'qemu.slice_100.scope' is called 'qemu.slice_100.scope' ?

23

I added the details to the issue QEMU metrics incorrect and crashing dashboard · Issue #9254 · netdata/netdata · GitHub

24

No I don’t use a parent/child setup and I can confirm that none of those previous edits are necessary as I have just tested.

Thanks everyone for your help, I’m a bit embarrassed as this could have been solved way faster had I known to look at the logs. It looked more like a missing feature than an error hence I didn’t check the logs :frowning:

After adding netdata to the www-data group I now see those lines in the log file:

2021-04-12 15:41:35: cgroup-name.sh: INFO: cgroup 'qemu.slice_102.scope' is called 'qemu.slice_102.scope'
2021-04-12 15:41:35: cgroup-name.sh: INFO: cgroup 'init.scope' is called 'init.scope'
2021-04-12 15:49:15: cgroup-name.sh: INFO: cgroup 'qemu.slice_100.scope' is called 'qemu_mail'
2021-04-12 15:49:15: cgroup-name.sh: INFO: cgroup 'qemu.slice_101.scope' is called 'qemu_host'
2021-04-12 15:49:15: cgroup-name.sh: INFO: cgroup 'qemu.slice_102.scope' is called 'qemu_apps'
2021-04-12 15:49:15: cgroup-name.sh: INFO: cgroup 'init.scope' is called 'init.scope'
2021-04-12 15:50:46: cgroup-name.sh: INFO: cgroup 'qemu.slice_100.scope' is called 'qemu_mail'
2021-04-12 15:50:46: cgroup-name.sh: INFO: cgroup 'qemu.slice_101.scope' is called 'qemu_host'
2021-04-12 15:50:46: cgroup-name.sh: INFO: cgroup 'qemu.slice_102.scope' is called 'qemu_apps'
2021-04-12 15:50:46: cgroup-name.sh: INFO: cgroup 'init.scope' is called 'init.scope'
2021-04-12 16:00:00: cgroup-name.sh: INFO: cgroup 'qemu.slice_100.scope' is called 'qemu_mail'
2021-04-12 16:00:00: cgroup-name.sh: INFO: cgroup 'qemu.slice_101.scope' is called 'qemu_host'
2021-04-12 16:00:00: cgroup-name.sh: INFO: cgroup 'qemu.slice_102.scope' is called 'qemu_apps'
2021-04-12 16:00:00: cgroup-name.sh: INFO: cgroup 'init.scope' is called 'init.scope'
2021-04-12 16:02:14: cgroup-name.sh: INFO: cgroup 'qemu.slice_100.scope' is called 'qemu_mail'
2021-04-12 16:02:14: cgroup-name.sh: INFO: cgroup 'qemu.slice_101.scope' is called 'qemu_host'
2021-04-12 16:02:14: cgroup-name.sh: INFO: cgroup 'qemu.slice_102.scope' is called 'qemu_apps'
2021-04-12 16:02:14: cgroup-name.sh: INFO: cgroup 'init.scope' is called 'init.scope'

Sorry for the delayed reply BUT:

You’ve reached the maximum number of replies a new user can create on their first day. Please wait 20 hours before trying again.

25

No, it is not that straightforward, you need to know what to look for in the error.log (grep cgroup-name).

No I don’t use a parent/child setup

What version are you using? We couldn’t reproduce the bug using current stable version with standalone netdata setup.

26

I had literally installed the latest netdata right before opening my thread here.

netdata -v
netdata v1.30.1-1-nightly

btw. I installed nightly by mistake, is there a way to easily go back to stable?
I basically installed netdata via your install curl script onto 3 servers and am only making those stats accessible via netdata cloud, no direct access.

###edit###
sorted with
bash <(curl -Ss https://my-netdata.io/kickstart.sh) --stable-channel --reinstall