MowMdown

Yes it is supposed to start all stopped containers

You can also just type the IP address into the URL bar of your web browser and it will take you to the webUI instead of using the terminal screen. For example

Like I said. You would want to setup a docker container file sharing service (ie Nextcloud/Filerun) which will do what you want. Unraid itself (nor really any other NAS type OS) by itself is suitable. Im not sure why that option is off the table. Unraid is more than capable of running these services, I myself use them and they work.

You cant, you either would have to assign them a user/pass and they would have to input those into their windows crediential manager or they would require you to set up the SMB user to match their windows credentials. You would be better off hosting a service in which does what you need like Nextcloud/Seafile of any number of the self-hosted document services if that's what youre doing. SMB to random users is a horrible idea and a security nightmare.

You can also try this if youre running a linuxserver image of plex: https://forums.unraid.net/topic/198283-arrow-lake-video-transcoding-not-working/#findComment-1618506

This was actually a plex issue and they very recently pushed a fix that may not yet be available to download as of this message. If you are running a linuxserver plex container you can add the following variables to pull the specific version. VERSION: 1.43.2.10623-07eb705bc PLEX_DOWNLOAD: https://downloads.plex.tv/plex-media-server-builds Source: https://forums.plex.tv/t/hardware-gpu-quicksync-transcoding-broken-on-beta-1-43-1-10540/937145/118

I am not aware if that is possible.

Yeah I suspect the ollama container is causing the issue with the NVRM out of memory, I dont see any logged now

Your GPU is what is running out of memory. It looks like (to my untrained eye) you are crashing the kernel driver for the Nvidia GPU right after all your containers are started up and I see you are running the ollama llm container. Set this container to not auto-start and reboot, run without this docker container and post new diags please

Powering on and off a remote server is going to require some kind of device like a KVM that can wire into the hardware or send a magic packet for Wake-On-Lan for out-of-band management. I myself am in the process of setting up a self-hosted S3 API compliant container called Garage which will let me use backrest to sync my data in an S3 compatible format which im already using to send backups to backblaze b2 storage. I don't see why unraid cant work. I dont see why something else cant work either. There are 100 different ways to skin the cat.

Just be aware that mover will not move files from this storage device you will need to use other means/methods to move files to the array, this was only to have SAB report accurate free space.

make sure you un-check the secondary and that the primary is set to the SSD youre using for downloads.

Only way to fix that is to assign the download share exclusively to the specific pool you created in the share settings and setting secondary to none Otherwise sab sees the /mnt/user combined free space for that share which would be SSD + Array You can also see on the unraid "Share" Tab what it will appear as, note the column on the far right under FREE

you might want to scrub the docker.img for corruption, I dont see anything logged. Otherwise if the issue persists, I would delete it and re-create it.

Well the risk is losing a single drive means total loss of the pool and most people, including myself, prefer our systems remain running and functioning in the event of a drive loss. 256GB is what I consider on the low end but I dont see why it wont be usable. I ran unraid for years with only a 250GB single drive pool before switching to a RAID10 setup where I now run mirrored pairs for my pool. As long as you have a good backup strategy, which you should have regardless of your pool configuration, and are prepared to deal with it shall a failure occur, it's your system to do as you wish with.

Have you ruled out the plugins youre using? I see corefreq and other system performance altering plugins, I would run in safe mode and only plex, and see if you can replicate the issue. You need to try and limit variables.

See if you can connect a keyboard and monitor to the system and if you can log into the shell, you should be able to type in diagnostics and it will run it and save to flash drive. I suspect the flash drive may be dying. If you cant access the shell via the local monitor/keyboard method, I would pull the flash drive and see if it can be read by your PC and maybe even run a fsck from windows on it.

please post full diagnostics in a reply to this comment. also if you have not yet, please enable remote syslog server, however you will want to use the IP of unraid or just 127.0.0.1 where it asks for a remote IP. this will cause it to save to the system itself.

I filed a bug report a while ago due to the calculation often setting a default min free space larger than the pool device when users have large array sizes. Say you have a 60TB array w/ a 1TB pool. Unraid will set this to 6.1TB which exceeds the pool size by 6x. I believe it's a 10% default calculation of the combined pool+array.

You will want to be on the 7.3.0 beta at a minimum as the 6.12 kernel 7.2.X uses doesnt have full Xe support and wont work with Intel Arc GPUs Open terminal and type in lspci -nn | grep VGA you should get an output similar to 2a:00.0 VGA compatible controller [0300]: Intel Corporation DG2 [Arc A310] [8086:56a6] (rev 05) Note down the [8086:56a6] the important piece is the right side of the : the 56a6 in my case your will be different Then go to Tools > System Drivers and scroll down to the i915 driver, click the little pencil and type in options i915 force_probe=!56a6 then srcoll down to Xe and type in options xe force_probe=56a6 remember to replace 56a6 from my example with your own value from the lspci command and then click the save icon.

@cjlmediasolutions Yes I was running the Xe drivers and switched back to i915 to test this specifically.

I ran 3 simultaneous transcodes using my A310 and the i915 drivers with plex 1.43.1.10611 for about 20 minutes and zero crashes or unraid going unresponsive, whatever your issue is I do not suspect it is caused by plex. Ive also been using Xe drivers as well, no issues.

Im not sure that applies to a zfs raidz2 pool does it? When replacing a single disk in a multi-disk pool like that. They would want to offline the bad disk, remove it, and import a new disk and re-silver it. I dont know the technical steps to advise.

Yeah sorry, mixed you up with someone else regarding the CPU, 8300 still has a decent iGPU for media transcoding and still more efficient than a P400. Regardless. Ok so yes the 580 driver is the correct version, lots of people have issues due to not having the correct version selected, anything newer than 580 wont work on legacy cards. You'd need to post your docker run on plex so I can see how you configured it for the nvidia gpu since it requires extra stuff.

4 dead drives all at once? I would say highly unlikely. Did you try a different power cable? or just connecting a single disk with a known working power and SATA cable? Make sure you're not using any power splitters either.