sjoerd

Hi, I just got home from my summer break and unraid is spamming notifications in the browser. Seems endless. Is it possible to write notifications to a file? I check the settings for notifications but all I can find are settings for agents. Regards Sjoerd

Oke, I created a rhel9 based vm with 8G and 4 cores, installed the docker-engine on it and fired-up a GitLab-CE container with a compose file and it runs butter smooth. The /var/lib/docker is mounted on a btrfs partition but should not be necessary, as docker uses its own overlay these days. I suspect something is not configured entirely correct on the docker-engine on the unraid server itself but there is really not much to configure from the UI itself.. Why do I think that: I fresh install of GitLab-CE on unraids docker crashes over and over (yes, I deleted the entire appdata/gitabce directory) All this started after I upgraded to unraid 7 - not implying unraid 7 is the issue but it might - Other dockers (plex, mariadb, netdata, wordpress, krusader, phpmyadmin (by default disabled)) are still working fine.

Hey Guys, Is is possible to use VirtIO-FS in combination with docker swarm - with swarm you want all your docker-engine hosts(vms)/containers to point to the same storage. I know installing microceph on all the hosts does the trick but I was wondering if you can achieve the same with VirtIO-FS - or is VirtIO-FS not meant for this? Little background: I want to dive much deeper into docker/swam/kubernetes as my team-lead from work wants to start containerizing a dozen of application (which is not a bad idea at all) but knowledge is minimal. Regards, Sjoerd

So i got it working - had the hack the template because somehow the mappings need to be the same so I changed 9080 to 7080 and and had the change the gitlab.rb config for that (nginx['listen_port'] = 7080) but I got no idea what the heck is going on.. I added my own user, added some groups and got tons op redis / sidekiq errors and timeouts.. It used to be very stable stable till last couple of months - say 6 months or something. I'm using the same image as work but on a vm that's configured as a dockerserver, also the latest and we never got issues there - I even got ldap authentication working there... I got the feeling GitLAB-CE really needs a compose file for all the separate services. When I browse and check settings on the admin area the container uses a 8 cores flatout which also makes the entire unraid server somewhat unresponsive. This should not happen. Even stopping the container from the gui results in an error - I got logged of from the unraid server and need to relog to the gui.. That on its own is already a weird behaviour. As much as I love gitlab - this is not funny anymore. Anyone else having issues like this?"

Can anyone help me out here - my current gitab-ce container crashes after using the frontend (browsing, editting etc). I got a bunch of errors in the log I can't figure out so I decided to create a second instance of the image with an other name, volumes and ports obv. The container is starting and according to the logs the container starts but I'm getting this message: My unraid server is obv named "pleiades" with ip 10.34.51.55 2025-04-27_08:52:05.02666 {"time":"2025-04-27T08:52:05.026635033Z","level":"ERROR","msg":"Failed to get receptive agents","mod_name":"kas2agentk_tunnel","error":"Get \"http://pleiades.local:7080/api/v4/internal/kubernetes/receptive_agents\": read tcp 172.17.0.2:43788->10.34.51.55:7080: read: connection reset by peer"} The container starts with these settings (all settings are from the unraid gui) docker run -d --name='GitLab-CE-NEW' --net='bridge' --cpuset-cpus='16,17,18,19,20,21,22,23' --pids-limit 2048 -e TZ="Europe/Berlin" -e HOST_OS="Unraid" -e HOST_HOSTNAME="pleiades" -e HOST_CONTAINERNAME="GitLab-CE-NEW" -l net.unraid.docker.managed=dockerman -l net.unraid.docker.webui='http://pleiades.local:7080' -l net.unraid.docker.icon='https://raw.githubusercontent.com/tynor88/docker-templates/master/images/gitlab_small.png' -p '7080:9080/tcp' -p '7443:9443/tcp' -p '7022:22/tcp' -v '/mnt/user/appdata/gitlab-ce-new/config':'/etc/gitlab':'rw' -v '/mnt/user/appdata/gitlab-ce-new/data':'/var/opt/gitlab':'rw' -v '/mnt/user/appdata/gitlab-ce-new/log':'/var/log/gitlab':'rw' --env GITLAB_OMNIBUS_CONFIG="external_url 'http://pleiades.local:7080'" --shm-size 1G --memory 16G --cpus 8 'gitlab/gitlab-ce' The container should start - the only thing that is different in the setup are the different ports on the HOST side, so why does my browser keep saying "site can't be reached" - unraid is listening on 7080 so ???. When I do "curl http://pleiades:7080" I get back "curl: (56) Recv failure: Connection reset by peer" - I don't get it.. And here is comes: when I change -p '7080:9080/tcp' -p '7443:9443/tcp' -p '7022:22/tcp' back to -p '9080:9080/tcp' -p '9443:9443/tcp' -p '9022:22/tcp' the container start and I can loging (i stop the "main" gitlab container ofc otherwise why would have an issue with ports already in use... Anyhow - I dont understand... I should not matter what the host-ports are as long as they are mapped to the container port. I already tried the changing nginx["listen_port"] = 7080 but that is the container-port.. I should not have to temper with that. Anyone know what's going on.. I really need create a second instance so I can move all my repositories to the new instance

If that is the case maybe check if there is an updated bios/firmware for you motherboard. I got a i9-14900 and had no issue after the upgrade. I know my motherboard (Asus Proart z790 had an update for "The new BIOS includes Intel microcode 0x129 to enhance the stability of Intel Core 13th and 14th Gen processors." somewhere during the summer last year. I'm not sure if my motherboard has this updated firmware are I purchased it in September - I don't know how I can check the current biosversion on a running system so can't check

would be handy if un-get got a silence install - now we always need to tap Y (but I think @ich777 needs to implement a -y or --yes switch. Apart of that - I'm quite happy that you set up a new repo

Yes - that was me being dumb - vim didnt work because I failed to dig a bit deeper. I just had in do "un-get install libsodium" and all was fine. Question was that stupid that I had to remove it

is make a docker for just vim not a tremendous overkill ? Dont get me wrong - im willing to leave nerdtools behind me but for rather simple tooling like vim, btop, nmon, ncdu we should not have to install containers - really glad the plugin python3 for unraid exists

Hi, I was unable to get that board - too much hassle as none of the resellers that I'm allowed to buy from (company policy) wanted to get it. The only brand for motherboards were Asus (or on entire supermicro rack server with a powerhungry xeon with only 8 threads) Instead I got a consumer board (ASUS ProArt Z790) - see my signature.

Hi guys, I got a dire question. My neighbor has a qnap with 4 disks and wont boot anymore. Ofc he did not make backup because "why do we need backups with raid 5 right" - despite the fact I told him many time to get an external 4T drive or something to at least backup personal data once a month. Anyhow - I was already building an unraid server for time (trial key still) and added the 4 drivers to see if one or more failed. Extended smart test shows the drives are just fine but ofc I cannot mount a "linux_raid_member" fs using unassigned drives. I was thinking about using an other qnap (system just might eol as its been running 24/7 for the last 12 years - disk are replaced twice during that time so current set nowhere near end of their lifespan) but all 2nd hand are almost the same cost as a new one - money they cannot spend. So - and maybe I'm not at the right place here - is it possible to somehow bring this linux_raid_member back to life on unraid (would be awesome) - is it possible to somehow bring this linux_raid_member back to life on a "donor system" - I already prepped a bootable disk with ubuntu 24 but I have no experience with mdadm whatsoever - The only thing I know is in which order the disk must are.

in unraid you do not setup the disk from the array in any raid configuration (hence UNraid) but having a raid1 zfs pool for cache ain't such a bad idea.. Google @SpaceInvaderOne cache" on youtube. This unraid user has godlike video's on youtube about setting up and replacing anything that has to do with disks - and more. If I got "stuck" or uncertain how the perform a certain unraid managment procedure. He probably got a vid for it.

Hi, I got the following setup: * asus z790 proart creator wifi (not using wifi) * 128GB mem (2x set of 64G) * i9-14900 (the 65W tdp with igpu version) This board has onboard 4x M.2 and 8x sata. Beware tho coz one op de M.2 slots shares bandwidth with 4 sata port (but lsi 9211 8i are available in wide variety). If you are indeed using a backplane - cables 4x sata -> 8087 do exist. Been using them myself in my old/backup unraid server I got 2x 2T m.2 in raid 1 zfs pool for user shares / appdata, 1 xfs 1T m.2 for vms and on 1 xfs m.2 for whatever (not used but probably going to use it for temp transcoding directory for plex and maybe if i start to reuse sabnzbd to unpack files (to save the use of my 2 expensive 2T nmve drives) Futher 4x 8T ironwolf in a 1 parity configuration. I don't have experience with pass-through but my printer is attached to the network anyways and have no need to hook up the server directly to an AV although I'm going to try to pass-though an installed low power nvidia to a linux desktop and windows desktop although I have no clue how to do that a.t.m. I assume that you can select all you want to pass-through in the properties of the vm.

Eureka! I got it: I still have my backup/secondary unraid server online where gitlab-ce also runs. The automatic backup procedure started to fail 3 days ago. They only thing I changed was the hardware - I did a major, like MAJOR hardware overhaul. Went from an i5-8600k (6cores)/ 16G to a i9--14900 with 64G. I restored the last good backup on my secondary unraid and all was fine, so I concluded (wrongly) that the PostgreSQL database got corrupted somehow. Now that I have a running gitlab-ce docker on the other unraid server I was fine with deleting the image and the content on the volumes (I kept the backups and scripts I made and the ssl certificates), but an entirely new installation of the image/container gave me the same errors. What the heck is going on??? When I googled /dev/shm, docker and postgresql I learned that /dev/shm could be full. Logging into the container console the /dev/shm was 64Mb (default size) and entire used where at my other unraid server it was only user for 20Mb - why? The other results from my search brought me on two gitlab pages: - https://docs.gitlab.com/ee/install/docker/troubleshooting.html - https://gitlab.com/gitlab-org/omnibus-gitlab/-/issues/5476 And there you have it - I added "--shm-size 265m" to the extra parameters. Restarted the container and YEAH. Al i have to do is restore the backup 🙂. When I check the `/dev/shm` I got this now: # df -hP | head -1 && df -hP | grep shm Filesystem Size Used Avail Use% Mounted on shm 256M 91M 166M 36% /dev/shm I also added --memory 16G but that got me a message about kernel not supporting memory limitation although when I check the docker page on advanced I does actually say 14G/16G used for the container.. See this thread

As of yesterday I got an 500 error on the frontend with these lines from the postgresql log: 2024-09-04_05:23:14.34374 FATAL: could not resize shared memory segment "/PostgreSQL.1945094446" to 84976 bytes: No space left on device 2024-09-04_05:23:14.34684 LOG: database system is shut down 2024-09-04_05:23:15.38092 FATAL: could not resize shared memory segment "/PostgreSQL.2018239940" to 84976 bytes: No space left on device 2024-09-04_05:23:15.38360 LOG: database system is shut down I need to fix this quite urgent Additional: A lot of these: {"severity":"WARN","time":"2024-09-04T06:52:35.985Z","message":"writing value to /dev/shm/gitlab/puma/gauge_all_puma_1-0.db failed with unmapped file"}