primeval_god

The plugin logs to the main system log.

I suggest uninstalling both the new and old swap plugins, rebooting, and then installing the new plugin. That should clear things out.

Probably not as the swap file is only used to swap out applications Actually I believe that when ram space is low any pages in ram (except maybe for some kernel stuff) are eligible to be written to swap. Theoretically if you have a swapfile active and are writing data to tmpfs those pages could make there way into the swapfile. That said i dont imagine performance would be that great, my experience with very large swapfiles (as big or bigger than system ram) hasnt been great. Might be worth trying though.

Reserved

Reserved

Swapfile (Unraid 6.9.0 and up) This plugin adds swapfile creation and management to unRAID. This is a fork of the original "Swap File Plugin for unRAID" by @theone. The goal of this fork is to revamp the plugin for unRAID 6.9.0 and add support for swapfile creation on BTRFS drives. Notes: You must remove the original Swap File Plugin before installing this plugin. This plugin no longer contains the feature to auto-update on startup. Dont forget to check the help text. Limitations: When using BTRFS swapfiles should only be placed on single disk BTRFS drives. BTRFS drive pools with RAID will not work correctly. When using a BTRFS disk the swapfile must be in a subvolume. The plugin will automatically create a subvolume if it is allowed to create the folder in which the swapfile will be placed. Better handling of this requirement will be added soon.

Ah yes you have found one of the classic stumbling blocks for new docker developers. This isnt an unRAID issue rather a non-intuitive specific of the Docker volume/bind mount system. You understood correctly however there are caveats that are not obvious. The VOLUME directive in a dockerfile tells docker that a volume needs to be mounted at the provided path. When the container is run if no volume flag is specified (-v) docker will automatically attach an anonymous volume to that path, alternatively you can use the -v flag to attach a named volume to your container. These volumes are managed by docker, stored somewhere in dockers system files (which on unRAID is in the docker image) and aren't easily accessible from outside. As the documentation suggests you can copy files to those volumes in your docker file and they will be placed into the volume at runtime. The caveat comes when we start talking about bind-mounts (which we typically use in unRAID). As an alternative to docker volumes you can use the -v flag to tell docker to mount a host directory in place of a volume (or really in place of any directory in a container). This is what you are doing when you use the path opton in unRAID's docker template. The issue you are seeing is that the files you added to the volume at build time are not made available when a folder is bind mounted into the container over that directory, this is just how docker does things. If you were to remove the path mapping in your unraid docker template and then try your test you would be able to see the yaml file which would be saved in an anonymous volume. As to how to achieve your original goal of pre-populating the config folder for your app, your best bet is to add those files to a different folder in your dockerfile and have a startup script in your container copy them over to your config volume on init if they are not already there.

They are not meant to be required, I believe that this plugin just hasn't been updated to account for swap file support in BTRFS.

No you definitely need to run mkswap on the file before using it. I didnt notice that the steps you had posted skipped that. Your missing steps are (note 1G for a 1GB swapfile).

I have had the same issue with the plugin before, particularly when using large swapfiles. Run swapon /mnt/cache-nvme/swap/swapfile

First step would probably be to go to the Nerdpack thread and request the addition of pixz. A less popular option, but one I always recommend because I think it should be more popular, would be to attempt to run it in docker. Normally I would suggest looking on docker hub for a pixz container, but I only see one available and it doesnt have a docker file listed. The your next best bet would be building your own docker image with pixz. Plenty of tutorials online of how to do it properly with a dockerfile, an easy start however would be to pull an alpine base image, run a console container, install pixz and then commit the image. Using pixz in unraid then should be as easy as running something like docker run --rm -v $(pwd):/mnt/:rw --workdir=/mnt/ yourimagename:tag pixz arguments Thats a relatively long command, but you could wrap it in a simple bash script or use an alias. I do something very similar with the crazymax/7zip container which I have script wrapped and aliased to "extract"

I have actually been toying around with a very similar idea myself for some time. In my case specifically however i was looking to have dockerman add this label net.unraid.dockerman.managed along with ones for the webui and icon to every docker container. The idea would be to have an option for unraid to display only containers labeled with the net.unraid.dockerman.managed label on the dashboard and an option on the docker page to have tabs for managed and unmanaged containers. This would allow me to hide any containers created by other sources (such as ephemeral containers) from the main interface. For getting the labels added to containers I believe adding the following lines to the xmlToCommand function in dockerman's Helpers.php was were i was looking // Add HOST_OS variable $Variables[] = 'HOST_OS="Unraid"' // Dockerman Control Labels $Labels[] = 'net.unraid.dockerman.managed'; $Labels[] = strlen($xml['WebUI']) ? 'net.unraid.dockerman.webui='.escapeshellarg($xml['WebUI']) : ''; $Labels[] = strlen($xml['Icon']) ? 'net.unraid.dockerman.icon='.escapeshellarg($xml['Icon']) : ''; @bonienl

A docker container can be connected to multiple docker networks, just not if i remember correctly anywhere in the unRAID web interface. You would use the command docker network connect from the terminal to connect a container to a second network. The annoying bit, again if i remember correctly, is that you would have to rerun that command everytime the container is re-created, such as when you use the update button in the web-gui.

So I am seeing something oddly similar. I just installed 6.9.1 yesterday and now i see this at the bottom if the "Array Operation" tab on the "Main" page. The thing is I dont have the s3 sleep plugin installed, and I havent for quite some time.

Why would you need to disable the docker service to change the appdata share settings?

Make sure that you have the search dockerhub setting enabled in Community Applications. Then use CA to search for the docker image you want to use. When CA returns no results, click the whale to continue on to results from dockerhub. Select the repo you are looking for and CA should open the template creator/editor. As for what you will need to add to the template i am not really sure. The docker hub page for the repo you listed doesnt give much for setup info, such as what ports to bind, what volumes are needed etc. Nor do i see a direct link to a github page that might have that info.

I would suggest building it into a docker container.

As most of the veterans here will tell you, except in the cases of a gaping security hole or potential array corruption, there is no immediately with the unRAID developers. Its more likely that they are unaware of the issue or that it is just to far down the priority queue. There are so few actual employees of Limetech that things tend to move slow and they dont monitor every forum post. It looks to me like the Bug report post is extremely new.

You can use the User.Scripts plugin to run a bash script on a schedule. So far as running youtube-dl specifically the best way is within a docker container. You would use a docker run command in your script (probably with the --rm option to delete the container when done) to launch a youtubedl container from your script. Depending on the image you choose you may need to override the container entrypoint to directly launch youtubedl and then your would pass your arguments in the docker cmd.

My suggestion is to build a docker image (using one of the official python base images), include your python script (run it from the entrypoint) and all of its dependencies. Then run the container from your user script.

While I dont know for sure how the docker isolation plays into things, i do know that underneath docker is using Cgroups to do its thing. Just an additional thought, while i know nothing about vmtouch, if it worked inside of a docker container that would enable using a version on top of another base os.

Theoretically would the script be effective if it was run within a docker container (specifically a long running container that periodically invoked the script rather than spinning up a new container for each invocation)? If so then might it be possible to use the per-container swappiness settings of docker to achieve the above?

Unfortunately so far as i am aware this is the best there is. When last i asked about how to get started creating plugins it was suggested that the best way is to look at the github repos of some of the more prolific plugin writers to see how they constructed their plugins.

Do you have all the dbengine configuration options uncommented? memory mode = dbengine page cache size = 200 dbengine disk space = 2048

It works in the netdata/netdata image. You also have bind mount /var/cache/netdata/dbengine/ to a folder in your appdata to ensure it persists across container upgrades.