After uninstalling flashmemory plugin with GUI for a test, reinstalling plugin again, i got nginx.service startup error, coming up on server console!

Quote from prtigger1

After that, i installed stable flashmemory plugin again with the GUI

Did you install the testing version of the plugin?

Quote from prtigger1

Any suggestions, what i can try now?

I would guess /var/log/nginx directory was lost.

sudo mkdir /var/log/nginx

sudo omv-salt deploy run nginx

Make sure to click sync in the flashmemory plugin.

Quote from prtigger1

Why this wasn't happening before...

I don't know. Maybe your system is taking longer to sync all of the files into ram on startup. This is one of the motivations for re-writing the plugin to use zram-config. Since zram-config uses overlay filesystems, it doesn't sync anything on startup.

Quote from prtigger1

because for me this solution (overriding the unit file) is

'hacking' a stable plugin.

The override is adjusting the folder2ram_startup starting position in systemd's starting list. It isn't hacking or even touching the plugin and is the only way to make a change that survives updates.

Quote from prtigger1

i would prefer a direct change of the unit file, but i find this at different

locations

That is not the way you should make changes. sudo systemctl edit folder2ram_startup.service will create the proper override file.

Quote from prtigger1

If i create the override config file, how

could i revert it back to the original state?

delete the override file.

rm /etc/systemd/system/folder2ram_startup.service.d/override.conf

systemctl daemon-reload

Quote from prtigger1

What about your new (testing) flashmemory plugin? How to handle the

default swap partition that omv is using

with my installation. What steps i have

to do, to get it running?

I don't understand what you are asking. You want to remove the swap partition that the OMV install creates?

Quote from prtigger1

Is there any problem with the swap partion use

nope. The swap partition and zram swap serve somewhat different purposes. swap partition is typically used for when you running low on ram. zram swap is for create compressed swap to keep things in ram that are there due to a swapiness setting.

Quote from prtigger1

Do i have to change something, before i can try your testing flashmemory version?

I didn't on any of my systems and they are all running the testing version. I did remove the zram swap (easy to do in the new plugin) because I didn't need it.

Quote from prtigger1

Or better staying with old plugin and doing the override config, because the new plugin is to much beta state right now?

zram-config isn't new or in beta. The plugin is basically just writing /etc/ztab. Doesn't take much testing to know that is working right. The beta part of the plugin was just to determine if the default settings are large enough. Since the settings are available to easily change in the plugin by the user, that isn't a huge deal either. And the fact the current folder2ram-based plugin is causing you problems, I don't what is the better option. If you really don't want to worry about it, use a good ssd or nvme and then you wouldn't have to use the plugin.

Quote from prtigger1

I´m not understanding the memory limits and disk sizes at this time, and what happend when sizes will exceed...

Each line is creating a ram disk for the directory mentioned. When it fills, it is the same as filling the old flashmemory or any directory.

Quote from prtigger1

... and how to cleanup growing logs on the system....

Depends on what it is. Most logs have logrotate setup or are sent to journald which automatically truncates. I think chatgpt would help with specific log questions.

Quote from prtigger1

Do i understand it correct, that zram swap is not using the swap partition, generated during OMV installation on the boot disk?

You can have more than one swap. zram is just creating another one. It does not use or interact with the install created one. You can just delete the swap line in the plugin. Then you don't have to worry about it.

Quote from prtigger1

I only try to understand the function.

I'm really not a good teacher. You can read about zram-config on the owner's page - https://github.com/ecdye/zram-config/

You can google about zram in the linux kernel to learn more about that.

Quote from prtigger1

I want to know, if this function is the zswap function i read about?

The plugin only uses zram-config. zram-config is only using zram. zram just creates block device. You can format it ext4 (which is what it is doing for files) or format it as swap. zram doesn't care. And while the plugin does create a zram device for swap, the plugin's main goal is to reduce file writes not swap writes. If a system is sized correctly, it shouldn't even use much swap.

Quote from prtigger1

Because swapping normally uses the swap partition on my small boot SSD i try to find a better solution

You could always move it to a file on a hard drive (if you have one) or plug in a usb stick that you don't care about.

Quote from prtigger1

because this is bad... Nobody of us can change the kernel programming and Trixie is on the run.... I guess, people with a lot of RAM don't mention this behavior..

Why is it bad? This is exactly why vm.swappiness exists. If you don't want the scheduler to swap out what is most likely unused ram, then disable it. Setting vm.swappiness to 0 should work just like what you were used to with the 6.1 kernel.

I honestly don't pay attention to my large server on 6.11 since it does have 128 GB of ram. I just set the swappiness to 0 to see how it acts. My RPi is my low mem system but it is still on the 6.6 kernel and using no swap.

OMV does have a swappiness environment variable
https://github.com/openmediava…loy/sysctl/nonrot.sls#L31