Nodiaque

I didn't say it's wrong, I said it's unsupported by the dev. When I open ticket in the past saying I used this container they closed the ticket saying its unsupported installation. And it's clearly documented in the installation documentation that it's unofficial and unsupported way. Since it's very easy to just have the 2 official container that replace that container and get the support that goes with it, and be able to use it 100% like the documentation, it just made sens to me to switch. But anyone is free to use the AIO container, I never said it was a bad container nor that it need to disappear or that the maker did a bad job.

That's why I switched to the official docker instead of an unofficial unsupported "all in one" Docker. This way, I just follow what the release say and that's it. No need to wait for a custom build and I can use 100% exactly as intended by Immich.

I did read on how to make the automation. One of the problem is a bit of lack of interest on my part. I'm no Github expert and when everything start to require custom yaml and such, I'm like bah, there's surely better way. Since my job doesn't use git, I don't learn it even if I code heavily and at my age, we tend to stop wasting time ;) I've yet to ask some friend who use git like a second nature to simply teach me. I know they will. I can say for sure that the only reason I was able to explain the docker build and such process is because I did it not too long ago for this specific Docker. I do know the rest properly, but I don't build docker often so I normally forget and need to read my notes again.

Yeah sorry I'm tired so translating my though in English is fleaky at best. In unraid, the "update available" is simply a docker webfront. What this does is it check the image in the template (where it says nodiaque/steamcmd:windrose for exemple). In this exemple, nodiaque/steamcmd:windrose mean: Repository user nodiaque Image name steamcmd Tag windrose Normally, in docker, you have image done in this way: Swag/swag:1.0.0 Swag/swag:latest Etc.. Now, what this does. When you select the app in the unraid store, it download the template (the Gui you use to customize settings when creating the app). This template in the end is like a docker-compose.yaml. When you click add/save/run, you can see it download the image and do a docker run command. Now, what happen in the background is periodically, unraid check the image docker have (in unraid cli, you can do "docker image list" and see all image downloaded) vs the docker hub registry (or ghcr.io depending on what Docker is used in the template). If the version on docker hub is more recent, it will show "update available". When you click on update, it does the command "docker pull image" where image is the information under the template for the docker hub registry (like nodiaque/steamcmd:windrose). Now, how is there a new version on docker hub? This is where someone need to create it. We do an image doing "docker build source tag" and then docker push to send the image to Docker hub. The docker build command require a source, and this is where you build the entire image and also, for us, update the server files from official windrose image. What I manage to do, with the source git of the other person, is automate the source file. I was doing it by hand but there's command in the dockerfile (used by the docker build command) to fetch file from another Docker directly. The problem now is to automate the docker build and docker push command when a new official windrose image is done. This is what the other person on git did, this is where I don't know how and need to learn a lot.

This is not what I'm talking about. To get what you are talking about is what I'm talking about. This update feature in unraid is simply a check of your Docker image version vs the current one with the tag you use. Unless I update the image this won't happen. It cannot be something in the image that check if something need to be updated and then it show that. It is a docker subsystem thing. The problem is updating that image. This is where automation is required else I have to do it manually each time. Edit: I might have badly explained myself but in the end, what need to be understood is I have no control on what the "apply update" do in unraid. This only trigger docker to download new image.

I know about the doc, I've read it many times. Using the CA is actually not supported, St least when I opened ticket since its not directly their container. It also says right at the start of the doc The Unraid template uses a community made image and is not officially supported by Immich This is for the aio. It is not official per immich, it's not even supported. Docker compose plugin is a way. Doing it manually with the other CA that are separate of each docker from the docker compose yaml is also supported since its the exact same docker as docker compose but with template. I got support on more then one occasion with using only the official docker, using compose or not. This AIO as 0 support from dev.

This is one way using this docker, that is all in one with the machine learning and doesn't use official image. The guide I provide some post behind is using all official container using unraid CA. It's replicating Docker Compose way of doing it without the plugin, all with CA. There's even a guide with all the setup for files which, since we are using official docker, we can now use all env variables properly. The choice is to the user in the end. I made the switch so I now have full support from Immich dev.

Someone said they downgrade when not in use? All slots are pcie 3.0 according to spec. It's also in the recommend slot from Dell to put gpu in

Restoring to a working version of your dB at the same version of the container... If you simply restore with the latest version, it's normal the issue is following.

Ah! Good to know

Yeah he's doing exactly the same as me but he automated it. I'm not very github / docker knowledable so I know nothing about automation. What he does is the automation check nightly if a new windroseserver/windroseserver docker image is built. If yes, import, extract /home/ue_user/app/ (where all the linux binary are) and then copy them into it's image. It's what I'm going but manually. We both use a debian-trixie image and we even have many things in comon on how our image are built. Maybe he looked at ich777 original work (or the one he based it off). In the end, it's not rocket science for that. He also start directly with debian-trixie-slim and build from there. while I start from a custom debian image ich777 keep updated (I have made mine but since I have nothing to keep everything updated automatically or simply notify, I still use his). Those "custom" image are tailored for Unraid thus better anyway (respect uid and guid as parameters, properly set permission on run, etc). edit: Restarting his docker (netdy tech blog) won't update your binary when looking at the code in github. It's a new docker image, so it's more then restart, you have to update the image, just like any normal docker image. edit 2: I did learn from his docker though. I didn't think I could copy files directly from another docker in a dockerfile. I'm currently testing it out. If it work, I'll have to learn automation for the next. Thing is now, how I find version since I was doing it manually. edit 3: Well it work, so it make updating the image much easier and faster. I read on the automation in github and my head hurt so I haven't got anywhere for now. I'm thinking I could do a fetch at the beginning of the image but this would take a long time at each boot. To be continued

Oh I don't contradict your information, I did found out the broadcom information about the LSI SAS 2008 which says it's PCIe Gen 2 (https://docs.broadcom.com/doc/12352283) So it's the plugin that is in the wrong then, cause it shows as 2.5gbps But the NVidia is downgraded to PCI gen 1 though LnkSta: Speed 2.5GT/s (downgraded), Width x16

I'm unsure if it's really an error, the whole page on amazon speak of gen 3, even the title say gen 3. It's been far too long for me to return. But shouldn't it show gen 2 and not gen 1?

So I got sold something that isn't what spec on the seller page if I understand correctly, cause they say it's a PCIe Gen 3 x8. But if it's gen 2, why does it connect at gen1 speed?

NVidia Quadro and SAS HBA

Here it is lspci.txt

The problem is the template, when you load it, doesn't load the default host path properly. It only put /mnt/user/appdata/redis instead of /mnt/user/appdata/redis/conf/. I don't see how changing the container path would fix that. Is there something behind the scene that prevent a container path /config to have a specific path name? I never saw that kind of problem on any other template, normally, path are working fine

Hello everyone, I'm looking at various plugin and information about my server. It is a Dell Precision 5820 with a Xeon W-2275. According to specs, it has all PCIE Gen 3 slots. But looking at lspci output and various other plugin like hbaviewer and GPUStats, they are reporting connected as Gen1 speed but Gen 3 available. Does anyone know how to enable it to be on Gen 3? Both card connected are Gen3. Thank you

That's why I made it a different container. Some might want to go to the Linux container instead of wine compatibility layer. I just really don't like how they did it. I never saw a game company release a docker and not use the distribution mechanic. I open a ticket with them asking to push the Linux server on steam like they did with Windows.

But why does it truncate the path in my XML? I never had that problem. I tried recreating the template and it still did that.

Is there some monitoring on that? It seems to be a place to upvote for feature, but right now it's a bug and because of that, my template is wrong although the xml isn't.

Ok, so I made a new template. Windrose-Linux. This docker is totally different then the current Windrose one. Windrose didn't release the Linux dedicated server into steamcmd. The only way to get the file are to extract them from their own docker. We could use their docker directly, but that docker doesn't run as 99:100 user which break unraid. So what I did is I extracted the docker, copied the server files, checked the dockerfile from the original docker and made it unraid compatible. Because of that, this docker doesn't auto-update with SteamCMD (it doesn't use SteamCMD at all). I will need to extract each time and create a new docker version, thus update won't be as fast for sure. Since I do not own the game, I haven't tested. According to log, everything seems to run fine. Migration should be as easy as starting the game, stopping, copying the config file and the save folder. DO NOT MAP DIRECTLY TO THE OTHER CONTAINER. You will have to create the config file (can be an empty file) before starting, else it will create a folder and crash. This docker contain all of the game, so it's bigger (~5gb) and you only have the savegame and the config file in appdata.

OK so I've checked that. What they release is a dedicated docker image built by them on repo windroseserver/windroseserver:latest. We could use that, problem is we cannot change the user. They run everything under the user ue_user and everything is under /home/ue_user/app. Because of that, if we set --user=99:100, it break. They didn't release anything to create our own linux server beside that. I'm trying to understand how the docker one they release work and see if it's simply running wine like we are doing. edit: Ok so far, I've manage to understand what they did. The problem is the source file aren't on steam. I could extract them directly from their own docker and have it run, problem will be each update will require a manual job (new image) which will mean having a maintainer. I'll continue investigating.

Easy fix? Restore your earlier working backup. Immich also do daily backup (unless you didn't enable backup) so you can simply restore immich backup.

Now what you can do is update to latest version of postgresql. The latest is ghcr.io/immich-app/postgres:18-vectorchord0.5.3-pgvector0.8.1. The version you currently have is a stepping stone because you had an older version. Do make a backup before. Easy way is stop the database, copy appdata and do the job. If something happen, just stop, change container info, copy back the backup and start. Also, check database log. If you are user other then postgres in your database, it could fail the update (some job after the database is ran using postgres user. I already opened a ticket on that). Easy way is simply to open a backup file using the FAQ procedure and run a restore database on a DB container (with updated configuration in immich container to for the database user). You can try without changing the username and see what happen. The database log will show you. Good thing with that is it's easy to restore.