HA out of memory after 2023.4.0

So I’ve recently updated my Home assistant Core to 2023.4.0. Since the update I haven’t been able to start Home Assistant. When I look at the CLI when starting up it looks like it should work at first but then I get: Out of memory: Killed process xxxx (python3).
I made a backup of my Home assistant before updating to 2023.4.0 but now I can’t even restore it.
When I try restoring it via the CLI it just says: Processing … Done. Error: Unknown error, see supervisor.

Does anyone else have this problem or a way to solve/ restore my backup?

Thanks in advance

2 Likes

Same problem. Not sure if you’ve isolated what’s causing it?

1 Like

Same here :frowning: System gets frozen randomly. Sometimes in minutes, hours, days…
Hardware: Rasperry PI 4B-4 GB SDRAM
OS version: 10.1
Home Assistant Core: 2023.4.6

All these problems started after recent upgrade to OS 10.1
Isn’t 4GB enough to run it all using node-red and a few automations?

1 Like

should be. I found that turning off my zigbee integration solved it… might need to remove devices one by one to check what could be doing it?

1 Like

Txs for the tip @tome88.
The suggestion of turning off zigbee scares me! Most of my devices are Zigbee and have always been working seamlessly.
It is true that I am trying these days disabling other integrations, but at the end, sooner or later, the same error happens (Out of memory). Before that fatal error, the memory comsumption keeps around 19-21%.
:frowning:

1 Like

@cgonzalez you could create an automation to perhaps restart HA or your config based on a condition that memory use gets to above 80% or something like that. I find mine does it after a couple of weeks but apart from the error message on my VM it still runs absolutely fine

1 Like

Got the same issue.
Running a NUC with 16GB RAM.
Goes from 0 to 100% usage within ~4h and then kills python3 process and starts over.

image

1 Like

Same Problem here… Raspberrry Pi 4 with 4GB freezes after some time… I moved my Home Assistant installation to a new home server… Same Problem as described above ( after some time Out of memory)

1 Like

I also have the same kill memory and reboot problem, have you found a solution?

After updating HA and a bunch of integrations my curve flattened down quite a bit. Before it was 3-4 hour between restarts. Now it is ~30 hours.

Another update:
I’m doing updates of OS and integrations all the time so might be one of those. BUT, I also found out that one device had a (probably) quite talkative MQTT service turned on so I did shut that down roughly the same time the graph started to flat line at ~6% with 16GB available RAM.

Hi!
I have increased the swap memory to an indecent amount of 4GB, I am analyzing the use of memory, but:

  • No weird memory consumption is detected.
  • Swap memory is never used
  • System still gets frozen from time to time with the “Out of memory: killed process…”
    Why are not these memory requests properly managed by the OS generating the out of memory critial OS error?
    230705_memPerc
    Currently using Raspberrry Pi 4 with 4GB:
  • Home Assistant 2023.6.3
  • Supervisor 2023.06.4
  • Operating System 10.3
  • Frontend 20230608.0 - latest
1 Like

in the same BOAT with all of you.

Exact same problem. I noticed that a CLEAN reinstall without installing my backup solves the problem. I don’t want to have to reinstall my 60-70 devices one by one but that may be the only solution. I happened out of NOWHERE!

I finally solved the out of memory issue and from my point of view is an error in the message the OS is giving to me.
Since I moved the data to an external SSD drive I have not suffered any issue, so now I have a micro SD card (Reformatted) for the OS and an extra SSD drive for the data, so my conclusion is that my micro SD had some faulty sectors that made the OS collapse when requiring them, but in this case, the error message should have been different instead of the “OUT OF MEMORY” error I was getting.

And here I am trying to figure out how I am to describe the issue that I am having. Yep, out of memory as well on a Raspberry Pi 4b with 4GB of RAM. The odd thing is that my memory usage is stable. I use about 700-800mb and it is pretty constant. I never see it rise at all, not even when I check the history. The out of memory happens when I attempt to install a new version of HA or even upgrade an integration.

I’ve just upgraded to 2023.09.01, but I had to stop Deconz and Google Back up add-ons to ensure I don’t have any out of memories occurring when attempting the upgrade.

I’ve had to plug the Pi into a connected Tuya plug so that if this happens, I can cut power and start the Pi up again. Just seems that if the Pi consumes a lot of CPU, like during an update or a backup or something like that, this happens.

I really hope somebody has a solution for this. I use an SSD for both HA and data as I moved away from SD cards.

3 month update
I’ve had no memory issues at all the last couple of months.
Seems stable at 6-8% all the time.

I’m running HA on a Lenovo Tginkcenter , on an SSD and 8 Gb of ram.
After last update I have the error "Out of memory: Killed process 4559 (python3)
HA OS 11.0

I did a clean install of HA, but after a while of running and trying to restore a backup I have the same error.

Does anyone have a solution for this problem?

All of these issues could be totally unrelated, as any integration can misbehave. First check addons - they list their own memory. Try rebooting and then disabling some integrations. If it helps, it is one of them, otherwise enable them again, reboot again and proceed to disable the next few…

Hi, I’m running on a raspberry Pi4 with 500gb ssd. How do you increase the swap memory.