[CLOSED] docker and whatpulse

So I have had this problem for a while now…

It seems whatpulse gets completely confused when I start to build docker images/containers.

Whatpulse will crash, I will try and kill it but it just goes to a zombie process.

I have looked and it seems to hogging the crap out of ksoftirqd. So it seems to be interrupt requests.

Has anyone else had this?
[hr]
prime example of this is as I write this post.

1.7 GiB + 2.1 GiB = 3.7 GiB whatpulse

PID USER PR NI VIRT RES %CPU %MEM TIME+ S COMMAND
3 root 20 0 0.0m 0.0m 44.0 0.0 1:33.39 S ksoftirqd/0
16986 ben 9 -11 611.8m 11.7m 18.7 0.1 0:13.96 S pulseaudio
14 root 20 0 0.0m 0.0m 14.7 0.0 0:34.19 S ksoftirqd/1

here is a strace of whatpulse

https://gist.github.com/81d1fdde737d608f6e72

Hi,

The only thing I can think of is that the client does not like the docker0 virtual interface. Could you try to turn off network stats and see if that fixes it?

I tried to ignore the docker adapter, but still caused the same problem and had to restart.

Now gone to settings and disabled network stats fully.

Is it possible to get this fixed :frowning: It seems quite a major bug!

nope, evening ignoring causes problems

627.6 MiB + 1.9 GiB = 2.5 GiB whatpulse

I don’t think ignoring would fix it as I assume the docker0 interface would create virtual interfaces for each container. Those virtual interfaces are probably the ones that the client can’t handle.

Could you confirm disabling the network counter completely stops the client from crashing when you run docker?

If so, you might want to leave it disabled when you work with it for now. I’ve put it on the todo, it might be an easy fix, it might not be.

I have disabled network counter but it still causes whatpulse to have huge memory problems + locks up my system.

For now I am not running whatpulse when doing docker work (which currently is most days :frowning: )

okay disabling network stats does seem to make a difference, I had to do a reboot.

I had same problem on my laptop too, so had to disable stats on that as well

Great, thanks for the feedback. I’ll put that in the help center for the time being.

when do you think you’ll get a fix for this??

This is a real major bug for me, as I want to track my whatpulse stuff still…

It keeps messing up my machine. If it disable/enable network, you have to reboot, otherwise it just causes crazy irq requests and sucks all your memory :frowning:

Not sure, the todo list is pretty long with more stuff with a higher priority then this one, as you are maybe 1 of 5 users who use this on their local machine. Not that you get anything from this answer, but I can’t give a proper one.

Huh, I have same problem (like Sc0rian).

Any changes for priority increase for this issue?

Slightly change of priority, as you’re the second to report this. :wink:

Not enough to give you a date of when it’ll be fixed though, sorry.

Hi,

Could you guys give me a bit more information about your OS (version, kernel version), docker version and network config? Trying to reproduce, not getting any crashing results.

so just had this today on latest version of whatpulse. Its really annoying :frowning:

Containers: 27
Running: 1
Paused: 0
Stopped: 26
Images: 146
Server Version: 1.12.3
Storage Driver: btrfs
Build Version: Btrfs v4.8.2
Library Version: 101
Logging Driver: json-file
Cgroup Driver: cgroupfs
Plugins:
Volume: local
Network: bridge host null overlay
Swarm: inactive
Runtimes: runc
Default Runtime: runc
Security Options: seccomp
Kernel Version: 4.8.6-1-ARCH
Operating System: Arch Linux
OSType: linux
Architecture: x86_64
CPUs: 12
Total Memory: 31.34 GiB
Name: skynet
ID: 4QAH:HHDX:3HWI:57CN:D5LX:OLBG:2N57:UIHA:UUZL:67BR:LKAG:B3XW
Docker Root Dir: /var/lib/docker
Debug Mode (client): false
Debug Mode (server): false
Registry: https://index.docker.io/v1/
Insecure Registries:
127.0.0.0/8

Do you have a link to any virtual images (VMware) with Arch Linux 4.8.6? The installation is horrendous and I’m not going to take 2 hours just to install an OS…

I doubt it’s a “Arch” thing, its probably a Linux thing. I would imagine you’d get same issue on a Ubuntu setup.

Anyway, you can try the vmware image here http://www.osboxes.org/arch-linux/ , but you’ll still need to setup a display manager etc…

It doesn’t always do it on the first time, but after 3-4 docker builds whatpulse normally crashes and ends up hogging the system. I can pkill the process and take control again.

It’s not a Linux thing, most (like Ubuntu) have a guided installer and allows you to get a UI along with the installation.

I’ll see if I can find an image with a proper setup (including UI), otherwise this will have to wait.