#Yellow out of memory, HA dies.
1 messages · Page 1 of 1 (latest)
Will post some logs here.
Right now it's unresponsive, but I did capture this: ```
b6f5759f205c ghcr.io/home-assistant/yellow-homeassistant:2024.5.1 "/init" 26 hours ago Up 3 minutes homeassistant
cb880c1cc874 ghcr.io/home-assistant/aarch64-hassio-supervisor:latest "/init" 30 hours ago Up 4 minutes hassio_supervisor
The "up 3 minutes" is about as much as I get.
That's just a partial docker ps. Not that useful.
I get about a 5 minute window to grab any data, then have to wait for the restart.
CONTAINER ID NAME CPU % MEM USAGE / LIMIT MEM % NET I/O BLOCK I/O PIDS
5d6740ff7cd0 addon_a0d7b954_ssh 1.52% 53.17MiB / 1.844GiB 2.82% 0B / 0B 58.1MB / 26.1MB 20
f0e4bd7f1ce3 addon_core_mosquitto 0.17% 65.36MiB / 1.844GiB 3.46% 84.2kB / 3.67kB 55.9MB / 348kB 20
0d79dd283b9b hassio_multicast 0.00% 4.887MiB / 1.844GiB 0.26% 0B / 0B 4.83MB / 307kB 8
7d8aefab3797 hassio_audio 0.03% 27.39MiB / 1.844GiB 1.45% 90.6kB / 0B 22.2MB / 389kB 12
ec2b26b8276f hassio_dns 0.04% 23.38MiB / 1.844GiB 1.24% 102kB / 9.1kB 19.2MB / 315kB 17
c8c23e0e5a9a hassio_cli 0.00% 10.41MiB / 1.844GiB 0.55% 99.3kB / 3.58kB 8.7MB / 303kB 9
5f1d97327cd8 hassio_observer 0.00% 9.02MiB / 1.844GiB 0.48% 103kB / 0B 6.26MB / 0B 5
b6f5759f205c homeassistant 102.88% 658MiB / 1.844GiB 34.84% 0B / 0B 249MB / 565kB 32
cb880c1cc874 hassio_supervisor 0.11% 117.4MiB / 1.844GiB 6.22% 185kB / 101kB 70.7MB / 492kB 26
My current guess is that one of the integrations goes bananas. You can test with with ha core restart --safe-mode.
--safe-mode is relatively recent so hopefully you're up to date.
yes, up to date. The memory usage in my working RPI4 is stable, and about the same (1.3G for the ha container) so nothing is leaking memory.
I'll leave a top running and hopefully capture what it gets to before it dies.
You can try docker top homeassistant -faux.
The current memory usage is fine though. The CPU usage is not.
root 3045 97.3 62.0 1898152 1200380 ? Ssl 06:52 5:01 \_ python3 -m homeassistant --config /config
top - 16:59:28 up 8 min, 0 user, load average: 1.14, 0.93, 0.49
Tasks: 13 total, 1 running, 12 sleeping, 0 stopped, 0 zombie
%Cpu(s): 24.7 us, 5.3 sy, 0.0 ni, 69.8 id, 0.2 wa, 0.0 hi, 0.0 si, 0.0 st
MiB Mem : 1888.7 total, 50.4 free, 1800.1 used, 100.5 buff/cache
MiB Swap: 623.3 total, 120.3 free, 503.0 used. 88.7 avail Mem
Please share the whole output. Also try the safe mode as mentioned.
Discord barfs on more than 2000 chars.
Please use a code share site to share code or logs, for example:
- https://dpaste.org/ (select YAML for the language, and consider picking a longer expiry)
- http://pastie.org/ (select YAML for the language)
- https://paste.debian.net/ (you guessed it, select YAML as the language)
Please don't use Pastebin, since it can randomly add spaces to the main view. Please also don't share text as images since it makes it harder for people to help you. Remember that others may have colour blindness, impaired vision, etc.
Right now I think I'll just monitor my working system. If it stays up, which it has so far, I would consider the problem solved. Running in safe mode will prevent loading some integrations, so memory usage will reduce anyway, so it won't really prove much.
🤷
It's just died again (triggered by trying to access the UI). This was the last output of top: ```
top - 17:03:45 up 12 min, 0 user, load average: 50.91, 24.84, 9.88
Tasks: 13 total, 1 running, 12 sleeping, 0 stopped, 0 zombie
%Cpu(s): 0.9 us, 68.6 sy, 0.0 ni, 0.0 id, 28.7 wa, 0.0 hi, 1.9 si, 0.0 st
MiB Mem : 1888.7 total, 24.3 free, 1869.9 used, 44.7 buff/cache
MiB Swap: 623.3 total, 0.2 free, 623.1 used. 18.8 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
573 root 20 0 3984 640 384 R 3.1 0.0 0:03.04 top
463 root 20 0 6588 0 0 S 0.3 0.0 0:01.24 sshd
18 root 20 0 208 0 0 S 0.0 0.0 0:00.00 s6-supervise
27 root 20 0 208 0 0 S 0.0 0.0 0:00.00 s6-supervise
1 root 20 0 424 0 0 S 0.0 0.0 0:00.05 s6-svscan
29 root 20 0 208 0 0 S 0.0 0.0 0:00.00 s6-supervise
30 root 20 0 208 0 0 S 0.0 0.0 0:00.00 s6-supervise
38 root 20 0 196 0 0 S 0.0 0.0 0:00.00 s6-ipcserverd
20 root 20 0 192 0 0 S 0.0 0.0 0:00.00 s6-linux-init-s
28 root 20 0 208 0 0 S 0.0 0.0 0:00.00 s6-supervise
310 root 20 0 7068 0 0 S 0.0 0.0 0:00.18 ttyd
313 root 20 0 6252 384 256 S 0.0 0.0 0:00.09 sshd
465 root 20 0 6448 172 128 S 0.0 0.0 0:01.74 zsh
client_loop: send disconnect: Broken pipe
Notice it's out of swap, and has only 24+44MB of RAM left.
I'm running exactly the same configs on the Yellow and the RPI4.
The large swap usage suggests thrashing.
Just harking back to your suggestion that the RAM use is abnormal, how would you know that without more information? I did say I have few add-ons, but I do have quite a few integrations. Can't provide a list, since the only place to see those appears to be in the Repairs/overflow menu/Integrations startup times, which does not provide a copy-able format. I've seen several comments to the effect that 2GB is a bare minimum for HA, which isn't unreasonable, just that shipping the Yellow with only the bare minimum seems a bit silly, given that to upgrade requires a complete new CM4 module (which I now have on order - 8GB)