Geck0

I went AMD for the first time and I'm happy with the chip. Although, I think unRaid plays better with Intel.

Simple, in safe mode you're on a reduced system. No plugins, etc. This results in lower wattage. Go to your normal system, that's pushing all the plugins, graphics cards vm's, parity checks, etc. Its under more load, more processing, more fan usage / water cooling pumps. The smart swtich will likely have a maximum load before it trips, ook up the model.Especially if you have other devices on the same plug (don't know what kind of smart switch).

How much money abd time did you spend on this issue? 🤣

That's okay, we can agree to disagree. Thanks for agreeing with me. Nope, you can change it back to :latest or whichever the latest tag available, I do it all the time. You're referring to people that pin the tag and leave it for months / years and get left behind. This is not what I said. , but you must look by yourself which those are you have to go through.

A bit late to reply on this, but hopefully it helps others as well. If you add a file to a Nextcloud folder, the database isn't updated, so Nextcloud doesn't know the file is there. You can add files this way, I do it all the time, if I'm moving a lot of files from areas of my server into Nextcloud. There is no point in downloading them to my client and Syncing, that would be daft. So, after you've done this, you need to run: docker exec -it nextcloud-container-name occ files:scan --all or, if you've added it to one person's folder only, instead of --all just use the person's name (case is important). Hope that helps somebody / everybody.

How are you connected to the database? Can you show us your config.php (cleanse it), also your docker templates for mariadb or mysql. Does the database docker stop or does the connection just drop? Have you checked the mysql logs??

linuxserver.io disabled the gui updater, but its still available on the CLI. docker exec -it "nextcloud-container-name" occ upgrade If you can't get back into the gui, you need to disable maintenance mode: docker exec -it "nextcloud-container-name" occ maintenance-mode:disable (I might have the maintenance-mode:disable incorrect, but you can type up to and including occ and it will show you all the options). Once an upgrade has been done, you can't switch to a lower version. You're better off restoring a database and nextcloud backup and doing it again. Edit: I forgot to add, if you want a smooth upgrade. Its worth doing an appdata backup, including mysql / mariadb. Secondly, I would disable the apps in Nextcloud. If there's going to be a failure, its normally a compatibility with apps between versions. I've had issues, mainly with the Elasticsearch plugin not working afterwards, despite being the correct version and I've had to reinstall it. Worst case, your upgrade fails. I agree with @jademonkee. I also don't use the :latest tag. You're asking for issues. Nextcloud is great, but latest versions are buggy as hell. You're better waiting for a couple of updates. I'm on the latest now, but I've had to fix a few things and there is still one repetitive error that they reckon will be fixed in 29.04, which they're just releasing an RC for now.

I would load a trial version and don't mount the disks. If that lasts for 24 hours, then mount the disks and give it another 24 hours. Then load a basic docker, like Plex and the disks and another 24 hours. If its conking out at any point along this basic line, then you have a problem that has to be physical. There is no way a clean unRaid build, running Plex and some disks can fall over because of the dockers or issues in the base config. Anything is possible, but highly unlikely. If it doesn't conk out, I would use the same usb and re-build the image (don't restore the image) and re-license. My two cents for what its worth. 🤪

That's what I've got for plugins and dockers, fairly unremarkable. I don't have any nVidia plugins, etc. I run a couple of basic Windows 11 VMs. I still don't think its the plugins though. The only other thing I can think of, is your motherboard's firmware. Sometimes the latest firmware doesn't play nice. I'm guessing if you've tried everything, you've upgraded, etc. and people normally do that when they have a new motherboard. Have you Googled to check if anybody is having trouble with their firmware version that you're on? (Sorry, if that's obvious). Also, are you using an LSI SAS board? If you are, have you tried moving the drives to the motherboard's SATA? Have you Memtested? Short of the above, have you got an empty SSD / nvme that you can use with a Windows or Linux build. Boot with it and leave it for a few days to see if this still occurs (leaving the physical part of the server exactly the same). If nothing, trash the unRaid image and start clean. Backups If you're machine is going down as often as this, I would really look at backuping up the important stuff and I mean daily. You're going to end up with a corruption somewhere along the line. When I was dealing with the similar issue (which was PSU), I eventually trashed the whole lot and started again, but it was an old rig that I'd had in storage for a while.

List your plugins and dockers. I'll compare it with mine.

Seriously? How annoying. It's difficult t9 diagnose when it happens, since you haven't got logs. I k ow how you feel. I had the same issue for more than a couple of weeks. Maybe unclean power? Are you connected to mains or UPS?

I agree with Kilrah. I've actually done it before, using existing cables from a previous psu, that was a different manufacturer to the new psu. Fortunately, not the motherboard, but there was a puff of smoke. You also don't know if the old cables aren't the issue. Its best to clean install. All of that said, I've installed my current PSU, keeping the existing cables but the system wasn't that old and I knew it was the PSU that was done. I still find it very unlikely that a plugin or app would cause this issue. Not impossible though.

Only other thing I just thought of, have you checked your usb key?

That was quick. I have to wait over a week to get a new one, where I live it takes ages. Hence why I have spares.

I forgot to mention. I use a mirrored pool x2 Ironwolf m.2 for data caching. Single SSD Ironwolf for downloads, Jellyfin media caching and transcoding. I use a separate 2tb Sabrent Rocket m.2 for VMs. The m.2 are on the motherboard's heating. To reach it is a full strip, because of the watercooling and graphics card.

Do you need that many drives starting off? I would go 20tb drives, price difference isn't much. I would go 4 array drives. 1 drive parity and 3 data. Keep a 5th drive precleared and ready to go, as a replacement or an array expansion. Keep the rest of the money for Black Friday. I keep a spare PSU and a spare hdd (already installed and orecleared). I don't keep spare nvme or ssd, owing to being able to keep operating without cache drives. Also, the price gigabyte tends to drop over time for these drive. 64gb memory is the sweet spot. I've got 128gb and I've never cracked 50% usage. I've got.the Fractal Design Mesh XL. I've water-cooled it but kept the HDD cage (custom setup). I've got room for approx 3 more drives. It's plenty of space. By the time I run out, it'll be time to swap out 14tb drives for what will likely be 25tb.

Yeah, I would plug for a 1500W Corsair. The fan doesn't even turn, so its pretty quiet and its been rock solid. If plugins were causing this much issue for you, there would be other people in the same boat and the unRaid guys would be all over it by now. Sometimes there are bugs, but not to the point of a failing system. That's my thoughts anyway.

I think you may find XFS will be more suitable. Bitrot can happen but, in my opinion, its unlikely. You haven't provided any info to work with in what your build will look like. However, you're usage is not too dissimilar to mine. Here's my setup: Have a look at the images above. Personally, the system may be overkill. You definately don't need 128GB memory. I would go for 64mb max. Its best if you digest the above and then ask questions. I may move to ZFS for data healing, however I keep good backups of my data. The movies, etc. if I lose them I will be disappointed but I won't sweat it. I've got terabyte external drives to write backups to. Eventually, when I'm not busy, I will build another unRaid for remote backing up to. So, XFS is pretty robust. I've had drives fail and when I've slipped a new drive in, its rebuild the data fine.

Well, if you figure it out, post it. the outcome will be interesting.

Well, look at it this way. If it turns out to not be the PSU, then you have a spare. It does sound awfully familiar though. Mine was going down randomly. It started off as reboots and then eventually it became crashes. It could be the case that you're running something like the dynamix file check and the load is tipping the PSU over the edge. Especially if the capacitors are giving out. Any coil whine or ticking noises?

I also checked for hard links, by comparing inode numbers for /mnt/user/appdata/mnt and /mnt. They're not the same numbers, if they're hardlinked the numbers would be the same. Also, I'm not sure that a hardlink would survive an Unraid reboot anyway (never tried). If they're not the same inode, can I simply delete the /mnt/user/appdata/mnt?? If I'm wrong, I'm guessing it may wreck my system.

Hi @Kilrah, No, its not that. I ran To list all symbolic links. Its got to be something else. I'm guessing its a mapping in docker, however I'm not sure how to diagnose it further.

PSU issue. I had the same problem a couple of years ago. I replaced it with a Corsair 1500. I purchased a spare one that is sitting in my cupboard, just in case. The issue you're describing is almost identical to what I was experiencing. Its either a load issue or your PSU is on its way out.