so its more from a performance standpoint then? ok... well im not a demanding person as far as needing blazing file access. im happy that it works and can play videos and store pictures. its just a personal thing. if i was using it in a work enviroment, it might be a different story, but for that stuff i keep it all local on this system.
thanks again for the replies
thanks for the reply ryeco. any reason you would run rain on a pi? to be more specific, im using 1 of the radxa penta hats, and i have 4 1TB ssds attached to it, in raid 5, as i said above. granted, its only been up for about a month or so, but so far I've had no issues.
I will look at omv regen as well. though, i think my intent was to simply make clones of the sd card periodically with another pi altogether. the 1st time i cloned the sd, i put the sd in a USB sd reader, and plugged it into an older pi4 running a desktop. I used the card clone tool on the pi desktop which was super easy.
im not putting anything on the raid that i dont already have double backed up anywhere, so if it failed, it wouldnt be the end of the world. but im just putting alot of effort into organizing the files, so that would be a long process to redo if the raid died. so i guess even if i used an sd clone that was a year old, it wouldnt matter... it would still boot up and bring to back to a time when OMV was alive and the raid was accessible.
Hi all. Iv recently setup a 4 disk raid5 through OMV. Iv spent the last several days upload alot of stuff to it and im wondering what the best way is to ensure the raid array is accesible should there be a system crash, where the pi/sd suffer a failure, but the actual raid array and disks are all fine.
i think the obvious answer is just clone the whole sd card, which i have already done. but im wondering if there is a more concise way to just backup OMV, or just backup the settings that describe the raid configuration? I only though of this because I believe I saw a video of someone setting up TruNas after transfering all his discs to a new case, and then stepping up to a different/newer version of TruNas. He explained how he'd backuped up the raid settings before hand, and then resotred them, and once loading up the settings, the raid was back intact and all this data was still accessible.
Thanks
I since moved away from OMV Compose and started using Dockge because it seemed it was easier to follow along with alot of other instructions about getting docker containers going. Probably because of situations like this where OMV wanted to do things different than most other explanations and I didnt have the knowledge or experience to follow along or make the necessary adaptations. Maybe it was a useless move... i dont know. But as i recall the solution in the case of Compose was putting the .env variables in the Global Environment files in the Compose tab (it has its own button in the top menu). Then in the Immich compose I removed references to .env specifically and just listed the variables which were then located in the Global Env. Without looking I believe thats what was described in the thread i linked above.
EG.. (just snippets to show the idea...)
Compose:
services:
immich-server:
container_name: immich_server
image: ghcr.io/immich-app/immich-server:${IMMICH_VERSION:-release}
command: [ "start.sh", "immich" ]
volumes:
- ${UPLOAD_LOCATION}:/usr/src/app/upload
- /etc/localtime:/etc/localtime:ro
environment:
DB_HOSTNAME: ${DB_HOSTNAME}
DB_PASSWORD: ${DB_PASSWORD}
DB_USERNAME: ${DB_USERNAME}
DB_DATABASE_NAME: ${DB_DATABASE_NAME}Global Env:
#Immich variables
DB_HOSTNAME=immich_postgres
DB_PASSWORD=MYPASSWORD
DB_USERNAME=postgres
DB_DATABASE_NAME=immichCan i ask a question about this as well? Haschmi's asnwer above seems like it will do the trick if I launch both containers, but before i do, im not sure on 1 point.
I initially setup piHole, and had it working for a couple days, but decided to take it down to do some tweaking, and figure out unbound. I had added the piHole IP address to my routes DNS. Again, this worked. Looking at examples of how to setup unbound, including the above, it looks like unbound gets its own IP address, which is referenced by piHole. Iv seen this be a 127.0.0.1, or something like 192.168.1.251 (which is the IP i would use in my application, as iv reserved the last few IPs in the range).
So up to this point it all makes sense, but what im not clear on is if the router DNS keeps the piHole IP? Or do i change the router to use the unBound IP? From what im read and seen, its seems like the router looks at piHole, and piHole looks at unBound which acts on its own, and then piHole gives the info back to the router.
Long story short... do i keep the router set to the piHole DNS? or switch to the unbound DNS?
thanks
Just for info, the openmediavault-k8s Kubernetes plugin now also offers a one-click installation of immich.
Thanks for the tip. When i started reading about all this a couple weeks ago, i did read about Kubernetes, but since I seemed to see much more related to doing things with Docker, I chose to try and stick with that and not make it even more confusing for myself by throwing in other peices of software.
All this said and done, im not even sure i like how Immich works now that iv tried it I initially installed Photoprism because it was a plugin, and worked immediately, but I didnt love the interface. I figured Id give Immich a try since it seemed more inline withe Google Photos. Im mainly doing this as a solution for my wife's massive photo hoarding problem, and she's familiar with Google, so Immich seemed like a good alternative to try. So, perhaps Photoprism needs another look.
Im the same way regarding helping people who know less than me in the feilds i study very thoroughly. Im not a good teacher. But im not putting in no effort. No effort would be bringing my setup to *TechSupportBusinessNameHere* and asking them to make it work while I pop out for a coffee.
Re: Docker or Immich forum. I put this here, in the Docker subforum, because it was a problem related to Docker as far as I could tell. If i was trying to do this and was not using OMV at all, yes, I probably would have asked on a completely different forum. Either way, It seemed the most appropraite place on here to ask. If you think it should be in another subforum, please feel free to move it where it makes the most sense.
I see you added the resolved tag. Thanks. I realized this morning I had fogotten to do that.
Not to flog this furhter, but a few additional thoughts...
Its not a lack of wanting to read, or learn. I did go reading, and did a bunch of searching and reading through other posts, here and elsewhere, of other people having issues with Immich, and how they got around the problems. Reading through other people's similar situations is more likely to be of benefit to someone in my situation because the chances of landing on the answer more quickly are higher then jumping headlong into the complete inner workings of Docker. Simply poiting to the manual is very much a needle in a haystack situation because i dont even know what im looking for.
Had you said something like "The error messages seems to indicate the problem is related to _____, so check page _____". At least thats a starting point and im headed in the right direction.
Im sure everyone here appreciates the work you have put in, but remember not everyone who joins a forum is joining to be an authorotative voice of help. Personally, most forums I join are ones where I need help, not wher I have knowledge to share. People join to BE helped, because they dont have the time or desire to become intimately knowlablge with the inner workings of a particular subject. To use the car analogy again, we all drive a car, and some of us know some bacisc like changing the oil, maybe the spark plugs, but far fewer know enough to tear apart and rebuilt a transmission. We are happy to enjoy the benefits of having a car without realy knowing exactly how it works and how to fix it if theres a problem. So we turn to others who do know in those times, with the hope of being helped, and not pointed towards self help. So like a car, Im trying to enjoy the benefits of OMV and Immich, without knowing the complete details of how and why it all works.
Anyway, again, thank you. Im sure i will be back in another thread with another issue 
But as always, i will be looking for the answer myself first, before asking.
Thanks for the reply, but I think thats akin to bringing my car to a mechanic and them throwing the service manual in my face. Im sure this is old hat for you, and any other mod on this forum, but alot of users here are , like myself, probably very unfamiliar with linux and we are just tyring to muddle through, not become power users. We're here for help, not redirection to self help.
Either way, I finally got it working. For anyone in the future looking for help, I found this post on Reddit, with a modified compose script, and the enviroment variables to accompany it. I applied the enviroment variables in the Services->Compose->Files->Edit Global Env Variables, changing the required variables where my details were needed. Saved, pressed the UP button, and it installed. Can now access the GUI.
I dont know, im just saying what I read and what I tried, because the Immich instuctions page are genearl instructions, and not specific to OMV. I came here looking for answers, and found that suggestion in another thread from someone also having the 500 Server error. I tried it both ways and neither worked, and I dont have enough knowledge about this process to say why.
I also read another thread on here which said remove all references in the compose script to the .env files and delete the .env files from disc (or somethign to that effect. i dont have it open anymore). That didnt seem right either, and I didnt try that one.
So if OMV Compose is creating the .env file on its own (which seems to be the case since I found it had done so when I went looking for it while following the Immich instructions), and that file tells me NOT to modifty it because changes will be lost, then Im not sure why OMV Compose cant find the file it created itself, or how to set the information it might be missing. Or perhaps the .env file isnt the problem at all. I see alot of 'variable is not set' errors in the red box, but nowhere to set them if I cant edit the .env, and the Compose script contains references (e.g. POSTGRES_USER: ${DB_USERNAME}) to information which I believe comes from the .env file. So a circular problem?
Im trying to install Immich, and being very new to Linux, docker, etc, Im having some trouble. I've tried looking for up to date guides to follow along, but havent come across anything thats working so far. I typed out a very long explanation of my steps, only to answer 1 of my own questions, which I thought might solve the problem, but it did not and only raised another problem and question. So, edited version to follow...
I've tried the official steps on the Immich App site, but no luck. My understanding of OMV Compose is that i can take an existing compose file supplied online, e.g. the one provided on the Immich github, and paste it into the OMV Compose interface, save, and run (UP arrow button), and it will do all the work to set things up correctly, in the proper directory structure.
I made a couple attemps unsuccesfully to get this to work, and then I made my 1st realization about why it might not be...
The 1st steps on the Immage site are to create an Immich directory, an .env file, and a .yml compose file, in the User Home directory. As i said above, since Im expecting OMV to create all the directories for me, if I need to create something to get the process started, I need to put it in the right place. After some digging, i remembered that I created a folder thru OMV when setting up Compose, and that is where all the compose files are stored. So instead of following the Immich guide, i navigated to THAT directory instead, and created the .env there.
In the CLI, did and ls to find "immich.evn" already existed from my 2 previous attemps to launch the Compose file. Id read in another thread that someone had renamed the .env file to 'immich.env' and that had solved their problem, so i decided to go with this file. i did a nano to edit it and found there was some default text from OMV about not editing the file, and it would be overwritten. Since my original attempt didnt work, i figured id try anyway. I went back to the github and copy/pasted the code from the example.env into mine. I changed my DB_LOCATION by copy/pasting my share folders absolute path from OMV. I then changed the DB_PASSWORD, wrote out the file, and went back into OMV.
From the github, I copy/pasted the compose.yml it into a new Compose. I was unsure about whether or not i needed to change anything here. I did not enter my Share folder here, as i did above in the env, because there is a # comment about leaving that line as is. I then did 2 attempts with tweaks to the .env lines. I tried once leaving it at default...
env_file:
- .envand when that didnt work, id read elsewhere a suggestion to specify explicitly 'immich.env'. But, specifiying that also didnt work.
Here is the error i keep getting...
So at this point Im stuck. The reason im trying to do everything though OMV with this build is because my understanding is that all the Compose containers and the plugins are started automatically when OMV boots up. If i install things outside of OMV, they might not start automatically with all the other services. Its also probably easier to manage all the services directly from inside OMV with a GUI, rather than all through the CLI.
The only other potential issue i can come up with (and its probably irrelevant) is that, as seen in the attachment, the command being run is apparenlty the "docker-compose" command, where as the Immage instructions page it say the correct command should be just "docker compose" (no hyphen). Perhaps this only applies when doing this through the command line with a non-OMV version of docker (i dont know how else to word this?)
Any help is appreciated, as I dont know what else to try. thanks.
ok, thanks for the replies. I'll press on from here and see what bottle neck i run into next.
Thanks for the reply. The time I encountered the error, it was with windows. It was something completely unrelated to this project. 35yrs of windows use, and it was the 1st time i'd encountered the issue. It will probably never happen again, but it just brought it to mind when seeing the long mount path.
But thanks for clearing up that Linux doesnt suffer from the same issue.
If i need to shorten the path, ill use the Symlink option mentioned above.
But on a related point, i was reading a through about someone building a nas, building the raid from the terminal, and they were having an issue having the raid auto-mount when the system was powered up. Is this something OMV handles automatically? I've only powered on/off once (its only been assembled a few days), and the raid was still present, so Im guessing OMV is handling this, whereas had I dont the raid configuration on my own, it would not automatically be mounted and useable after reboots?
ok, understood. So, to backtrack, if i started again from scratch (there's nothing on the drives yet, so Im wouldnt be losing anything), and built the raid from the command line, rather than letting OMV do it, would there be a way to create a different path, or would the result be the same? I was quickly reading thru this guide, and it never seems to cover anything about the mounting path for the raid, unless I overlooked that part.
The only reason I care about this is because I recently ran into a situation where I was getting errors from an excessively long file path/name conbination. It was a pretty extreme situation, but i was reminded of it when seeing this mount path earlier today and thinking it was long, and might possibly cause the same issue in the future, however unlikely.
thanks! however, im guessing this is not actually changing the path, right? The only reason i found this very long path to begin with was because i had connected my Android phone to the Pi, and was flipping thru all the folders trying to find where OMV had created the actual "share" folder, and i found it in the above mentioned path. So again, will symlink actually change the path, or just create a shortcut so to speak? And if its just a shortcut, Im guessing thats only useful inside OMV? Or does it apply this change even outside OMV? ... Like can this shortcut be used in the command line as well? And further to that point, is symlink a linux tool that OMV is just making use of? Because if thats the case, then i can see the path shortcut being applicable even outside OMV.
apologies for the noob-level questions.
Hi. Im just installing OMV for the 1st time, building a little Pi Nas. Everything has worked fine so far. However, I've found 1 potential quality of life issue that may become a problem in the future.
I've noticed the absolute path that the Raid5 was mounted was a very long name/location...
/srv/dev-disk-by-uuid-26d2e52e-6c26-48c5-956e-1fe0d87c8666/
Im wondering if there is a way to change this to a more user-friendly path...
/srv/raid-disk/ .... or something like that?
Im familiar with the very very basics of working with a Pi/linux, and can follow along with guides, but dont know anywhere near enough as to where to start looking for the answer to this, or how to implement a solution on my own. I just know that i didnt seem to see any way in OMV to change this, so Im guessing its a command line fix, and if so, Im not sure if OMV would accept the new location.
Any help would be greatly appreciated. Thanks!