I/O issue when reading logs (Like supervised logs)

Hello,

I’m using a PI 4 2GB with a 120GB SSD over USB3.0.
The installation is a Raspbian Bullseye 64 bits with HomeAssistant Supervised Installation.

My problem is that when I look to the log of an add-on or the supervisor logs in https://hass.local:8123/hassio/system my I/O turns into 100% iowait.

It seems I can appears in other situation too, but I have som difficulties to debug this.

The problem is that a container make a huge READ for many minutes !

Any Idea how to solve this ?

Total DISK READ:       134.27 M/s | Total DISK WRITE:       159.23 K/s
Current DISK READ:     159.85 M/s | Current DISK WRITE:     116.05 K/s
    TID  PRIO  USER    DISK READ>  DISK WRITE  SWAPIN      IO    COMMAND
    813 ?sys root      129.20 M/s    0.00 B/s  0.00 % 85.94 % dockerd -H fd:// --containerd=/run/containerd/containerd.sock
   6881 ?sys root        4.14 M/s    0.00 B/s  0.00 % 26.61 % python3 -m homeassistant --config /config
Device            r/s     rkB/s   rrqm/s  %rrqm r_await rareq-sz     w/s     wkB/s   wrqm/s  %wrqm w_await wareq-sz     d/s     dkB/s   drqm/s  %drqm d_await dareq-sz     f/s f_await  aqu-sz  %util
sda            343.00 134848.00    33.00   8.78  128.87   393.14    1.00     60.00    14.00  93.33  184.00    60.00    0.00      0.00     0.00   0.00    0.00     0.00    0.00    0.00   44.39 100.00
❯ echo /proc/*/task/813(:h:h:t)
789
❯ ps -ef | grep 789
root         789       1  5 11:10 ?        00:00:50 /usr/bin/dockerd -H fd:// --containerd=/run/containerd/containerd.sock
❯ dpkg -l | grep docker
ii  docker-ce                            5:20.10.14~3-0~debian-bullseye   arm64        Docker: the open-source application container engine
ii  docker-ce-cli                        5:20.10.14~3-0~debian-bullseye   arm64        Docker CLI: the open-source application container engine
ii  docker-ce-rootless-extras            5:20.10.14~3-0~debian-bullseye   arm64        Rootless support for Docker.