bpie

my docker folders disappeared except one. I don't see them in configuration files. Does anyone has a clue what could happen?

I'm not an expert but it looks like you have Gen 11 intel GPU and this image has intel-compute-runtime (openCL runtime for intel) installed which supports Gen 12 and up. openCL is used for some operations which are not supported by eg. QSV in this example. You need to install intel-compute-runtime-legacy from AUR. Dunno if there is binary version, when i installed it i had to install some additional packages to speed up (I installed `paru` to install from AUR because this is what I use on my main system) and than I could install the legacy version of runtime, but its not binary and compilation took a lot of time and space (needed to increase docker disk size twice xD). I did this mainly to understand what is going on. Doing this every update is not the way to go. But I have no better advice so there is that. I had no more time to play with this. I'll post if I find a better way Side note: with intel-compute-runtime-legacy now jellyfin works ok for me. I have troubles on chrome, transcoding is 2x slower than on firefox (dunno why). Side note 2: Intel is dropping support for MediaSDK on linux and new sdk (don't remember the name) will support only newer GPUs so probably there will be even more troubles ahead sadly. Especially since intel thinks 4yo GPU is old... Edit: there is a intel-compute-runtime-legacy-bin package in AUR, dunno how i missed it. So I can share how to do that. It is not script ready, there will be a need for entering password (pass) or accepting (y<enter> or just enter if there is 'yes/y' starts from big letter which means it is default) commands to do (one by one) in binhex-jellyfin container's shell, DO AT YOUR OWN RISK # root pacman -S --noconfirm git pacman -S --noconfirm --needed base-devel useradd -m -s /bin/bash newuser usermod -aG wheel newuser echo '%wheel ALL=(ALL) ALL' >> /etc/sudoers echo 'pass' | passwd newuser --stdin su - newuser # newuser git config --global user.name "Your Name" git config --global user.email [email protected] cd /tmp git clone https://aur.archlinux.org/paru.git cd paru makepkg -si # MAKEFLAGS="-j 4" paru -S intel-compute-runtime-legacy # for compiled version paru -S intel-compute-runtime-legacy-bin notes: - legacy version of the package is not available in arch repositories (binhex containers are arch based) but we can install it from AUR. AUR CAN BE DANGEROUS, especially when package is not inspected properly by you. Especially bin versions. So do it at your own risk. - I use paru (to install from AUR) because I know it and it's fast. But it is not suited for scripts. - there is also (commented out) installation of compiled version. But it took many hours on my 4-core cpu. - those steps need to be run AFTER EVERY binhex-jellyfin UPDATE! One could try to make it work in cron on unraid (copy script to container and than run the script in container) but I do not know how to make paru work in a script. - I do not update repositories on purpose. If some URLs do not work while installing something (git or runtime package) update could be needed (`pacman -Syy` or `paru -Syy` probably will fix this,)

I’m troubleshooting why Jellyfin stopped working and came across similar issues with Quick Sync Video (QSV). Hoping this helps someone or maybe someone can help me. Jellyfin was working fine but started showing errors. The logs include: [AVHWDeviceContext @ 0x555eb74ecbc0] Failed to get number of OpenCL platforms: -1001. Device creation failed: -19. Failed to set value 'opencl=ocl@va' for option 'init_hw_device': No such device Error parsing global options: No such device This seems related to the OpenCL runtime. After some digging, I found intel-compute-runtime supports Gen12, but my GPU (UHD Graphics P630, Gen9.5) isn’t compatible. Installing intel-compute-runtime-legacy from the AUR improved things slightly, but performance is worse than before. It seems some API/SDK/driver doesn’t fully support my hardware, especially for tone mapping (that's why it used openCL). Compiling intel-compute-runtime-legacy (and dependencies) took ages on my Unraid server (4 cores, -j 4) and used a lot of Docker volume space. I’ll check the Jellyfin forums but wanted to share here in case it’s related to Arch's cutting-edge packages.

it should be 2 as in attached image. cache.cfg states: diskFsType="zfs" diskFsProfile="mirror" diskId="Lexar_SSD_NM790_2TB_NJF5744104557P2202" diskId.1="Lexar_SSD_NM710_2TB_NES132R002654P2200" diskType="Cache" diskType.1="Cache" and diskFsWidth is back to 3 in /boot/config/pools/cache.cfg

~# zpool status pool: cache state: ONLINE status: Some supported and requested features are not enabled on the pool. The pool can still be used, but some features are unavailable. action: Enable all features using 'zpool upgrade'. Once this is done, the pool may no longer be accessible by software that does not support the features. See zpool-features(7) for details. config: NAME STATE READ WRITE CKSUM cache ONLINE 0 0 0 nvme1n1p1 ONLINE 0 0 0 errors: No known data errors

I edited the pool config file (diskFsWidth="2" from 4), rebooted and it started up. But something is not right. Is it safe to run this while a array with cache is running? Need to educate myself about zpool / zfs.

I did nothing new, just stopped the array and now it wont start again with or without docker or VM on ;( now it says 1 group of 4 devices, previously it was 3, should be 2 dell-server-diagnostics-20240831-1131.zip

Didn't know that, thanks

Yes, I forgot to disable docker daemon (only containers were down). But something strange did happen to the cache config I think. Can I recover from this? Should I start a new config or try to modify existing one by hand or some other tool? Edit: Oh, I disabled Docker and VM and the array started (after a few refreshes of UI, probably some warnings "cached"). Should I clean it now somehow- repeat the actions with unbalanced plugin to make sure all files are on one disk? I do not really know how this system works under the hood so I have no idea how to clean this up. Edit2: docker daemon started, docker containers also... ok I thought it will be something more serious thanks for the help

Setup: - array: 3 disks, 2 data and 1 parity (xfs) - cache 2 ssd zfs mirrored When I saw my disks on array were constantly spinned up, i checked that there are some files on system and domains shares. I tried using mover (Cache <- Array) but it did not fix it so I searched the Internet and found It can be done easily by rsync (or sth. similar) or "unbalanced" plugin, so I did the latter: - stopped the array - no VMs running so nothing to do there - run unbalanced from disk1 (both folders) to Cache (mirrored zfs) - started dockers and webUI stopped responding - restarted server - array did not start ("Wrong Pool State cache-invalid config" pop up) I see something strange in cache settings- it says "File system type: zfs mirror 1 group of 3 devices" but there should be 2 devices. I see also in cache.cfg diskFsWidth="3" the only 3 i found there but don't really know what it means for sure. do anyone knows how to fix the issue? Can I just edit the diskFsWidth (probably it's not that easy)?