iowait time (wa) in lxc after PVE dist upgrade

drlight

Member
Dec 23, 2019
8
0
6
35
Hello. I've just found out, that iowait time (wa in top) parameter in the lxc containers remains zero value just after I've upgraded PVE from 5.3-8 to 5.4-13. Is it a bug or a feature? wa in the host system and in kvm virtual machines shows correctly.
See attached screenshot of one of my lxc containers before and after PVE upgrade between 12:45 and 13:15.Снимок.PNG
 
hi,
did you reboot after the upgrade?
 
privileged containers or unprivileged?

does it affect all containers?

do you see anything weird/interesting in journals or logs?
 
privileged containers or unprivileged?

does it affect all containers?

do you see anything weird/interesting in journals or logs?
Unprivileged. Yes, it affects all containers (Ubuntu 18.04).
Nothing unusual in log on host or container. The only error in log on the containers:
Dec 23 19:45:49 ns2 systemd[1]: Failed to reset devices.list on /user.slice: Operation not permitted
 
does this issue still happen if you install another distribution in a container, like debian or archlinux for example? maybe it's an ubuntu issue.

do you still see the io stats on the PVE GUI?
 
does this issue still happen if you install another distribution in a container, like debian or archlinux for example? maybe it's an ubuntu issue.

do you still see the io stats on the PVE GUI?
I'll try on the other distrib and post the result. thanks.
And yes, I still see io stats on the PVE GUI
 
does this issue still happen if you install another distribution in a container, like debian or archlinux for example? maybe it's an ubuntu issue.

do you still see the io stats on the PVE GUI?
I've just checked privileged and unprivileged fresh deployed containers on Archlinux, Debian 9, Ubuntu 16 and openSuse 15.1 on two nodes of cluster - wa is zero.
 
And yes, I still see io stats on the PVE GUI
that's good then.

i think the problem has to do with the way top tries to poll the io data. not very familiar with the internals, but if it uses some syscall to do it, that might be blocked.
 
that's good then.

i think the problem has to do with the way top tries to poll the io data. not very familiar with the internals, but if it uses some syscall to do it, that might be blocked.
So, how can I fix this behaviour? htop for example have no iowait display option on the lxc containers at all! top remains zero =((( It is very important parameter for monitoring...
I even tried to cat /proc/stat and see zero value at the 5th place (it shows the amount of time the CPU has spent performing different kinds of work. Time units are in USER_HZ (typically hundredths of a second))
 
Last edited:

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!