Mainfrezzer

if you want the servers attached to your custom network, put in network_mode your custom network

yeah wing connects to the panel and then modifies the config.yml in its folder. the displayed config in the panel is just for the initial setup

if its that "empty", you didnt connect the node to the panel yet. the full file would look something along the lines of this debug: false app_name: pelican uuid: token_id: token: api: host: 0.0.0.0 port: 8080 ssl: enabled: false cert: /etc/letsencrypt/live/10.0.0.2/fullchain.pem key: /etc/letsencrypt/live/10.0.0.2/privkey.pem disable_remote_download: false remote_download: max_redirects: 10 upload_limit: 256 trusted_proxies: [] system: root_directory: /mnt/user/appdata/pelican-wing log_directory: /mnt/user/appdata/pelican-wing/log data: /mnt/user/appdata/pelican-wing/data archive_directory: /mnt/user/appdata/pelican-wing/data/archives backup_directory: /mnt/user/appdata/pelican-wing/data/backups tmp_directory: /tmp/pelican username: nobody timezone: Europe/Berlin user: rootless: enabled: false container_uid: 0 container_gid: 0 uid: 99 gid: 100 passwd: enable: true directory: /tmp/pelican machine_id: enable: true directory: /tmp/pelican/machine-id disk_check_interval: 150 activity_send_interval: 60 activity_send_count: 100 check_permissions_on_boot: true enable_log_rotate: true websocket_log_count: 150 sftp: bind_address: 0.0.0.0 bind_port: 2022 read_only: false key_only: false crash_detection: enabled: true detect_clean_exit_as_crash: true timeout: 60 crash_detection_activity_lines: 2 backups: write_limit: 0 compression_level: best_speed remove_backups_on_server_delete: true transfers: download_limit: 0 openat_mode: auto docker: network: interface: 10.16.5.1 dns: - 10.0.0.10 - 10.0.0.9 - 1.1.1.1 - 1.0.0.1 name: pelican_net ispn: false IPv6: false driver: bridge network_mode: pelican_net is_internal: false enable_icc: true network_mtu: 1500 interfaces: v4: subnet: 10.16.5.0/24 gateway: 10.16.5.1 v6: subnet: fd00:10:16:5::/64 gateway: fd00:10:16:5::1011 domainname: "" registries: {} tmpfs_size: 100 container_pid_limit: 512 installer_limits: memory: 1024 cpu: 100 overhead: override: false default_multiplier: 1.05 multipliers: {} use_performant_inspect: true userns_mode: "" system_ips: [] log_config: type: local config: compress: "false" max-file: "1" max-size: 5m mode: non-blocking throttles: enabled: true lines: 2000 line_reset_interval: 100 remote: http://10.0.0.2:8088 remote_query: timeout: 30 boot_servers_per_page: 50 custom_headers: {} allowed_mounts: [] Search: blacklisted_dirs: - node_modules - .git - .wine - appcache - depotcache - vendor max_recursion_depth: 8 BlockBaseDirMount: true allowed_origins: [] allow_cors_private_network: false ignore_panel_config_updates: false

in the wings config.yml is a section called docker. you can change the network_mode to whatever other docker network you have

This has caused severe emotional distress in me. I didn't know where to lurk, so I went outside and got a sunburn. Unacceptable. Please make it right. Thank you!

For anyone updating to the newest release of the pelican panel (as of the 22nd of june, beta35). you need to add a new mapping to your template to include plugins now https://github.com/Mainfrezzer/UnRaid-Templates/commit/310b40843ba29f1dbe4b179aefa6b7c8a54676ee

It does port forward just fine. It uses whatever mapping you use for container. So if you try to host a website in the container on port 80, but map it it 2345. It's port 2345 for the wireguard forwarding. Granted, gonna be a pain in the ass for anything dynamic but when you know what you need, it works fine. My main page and repo is hosted that way, including their proxy, as there I don't need ipv6.

Theres some posts in the discord about it, been borked for a while and seems to be somewhat browser dependent but you do get different results for each person as it appears. Dont worry about it

im not sure where you looked but its certainly 2023.11.22 https://raw.githubusercontent.com/giganode/unraid-i915-sriov/refs/heads/master/i915-sriov.plg

You actually never have to touch any of the command line, everything one could do with basic storage needs can be done completely via gui. Fuse is just the layer the array, the "unraid special", uses to combine all single array disks to one "big disk". Fuse does have its quirks but theyre well documented here. From your earlier comments, i dont know why you were playing around with the array anyway. Since you mentioned windows storage spaces, you should be aware of the performance of the "parity protected" option, it also suffers from the CMR/SMR mix (well SMR as parity device), so you must be using the raid profiles there. The pools would be the options to go there, since youre not limited by single disks speeds or the fuse overhead there. But that contradicts with what you actually want, since you seems too look for mixing disks. So im confused. And yeah its says 30 days, but you can extend 15 days for 2 times iirc, which should come up to 60 days total. Edit: You might have a look, if youre still interested in something linux, into mergerfs with snapraid. Maybe thats something for you. Works kinda the same but not really. I do use it for my backup server.

Just in case somebody happens to upgrade to 7.3.1 and experiences some funkyness with the ark container. Im aware of it looping. Im not absolutely certain as to why but i think i narrowed it down. I currently have a testbuild under the tag :test available. (includes a bump to wine11) Edit; Should be sorted via regular update

Just the way it goes for almost every package. Even a variation of the badge, that makes it clear that it just uses an image that happens to be curated by dockerhub would certainly make things more clear than a flatout "OFFICIAL" one.

who gets peddled by Dockerhub? Certainly helpful. Adguard/Adguardhome doesnt have it. And also ignores all that publish mainly to literally any other registry than Dockerhub. Also, thats flatout wrong, for CA. Binhex isnt a Maintainer of Pihole, and if his template is borked, Pihole cant do anything about it. Edit: I do wanna elaborate a bit further. The official tag makes maybe sense just on docker hub seen byitself. For CA it shouldn't be pulled at all. Official should be a manual tag that Lime/Squid attaches to container if there actually happens to be the same entity behind the app, the image and template. It just causes confusion and at the worst case annoys the Devs, when something is tagged as official but the issue is something they have no control over.

I do wanna mention that this "official" icon is a rather meaningless and dockerhub only feature.

The windows machine had in this case 2 Nics, didnt change anything May 6 10:26:25 Magnon-Box smbd[2168712]: [2026/05/06 10:26:25.696302, 0] ../../source3/smbd/smb2_read.c:265(smb2_sendfile_send_data) May 6 10:26:25 Magnon-Box smbd[2168712]: smb2_sendfile_send_data: sendfile failed for file Archive/Archive.part20.rar (Connection reset by peer) for client PID=2168712,CLIENT=REDACTED,channel=2,remote=ipv6:fd0> May 6 10:26:25 Magnon-Box smbd[965126]: [2026/05/06 10:26:25.804240, 0] ../../source3/smbd/durable.c:951(vfs_default_durable_reconnect) May 6 10:26:25 Magnon-Box smbd[965126]: vfs_default_durable_reconnect: default_durable_reconnect_fn [Archive/Archive.part20.rar] failed: NT_STATUS_OBJECT_NAME_NOT_FOUND May 6 23:59:26 Magnon-Box smbd[1542941]: [2026/05/06 23:59:26.039488, 0] ../../source3/smbd/smb2_read.c:265(smb2_sendfile_send_data) May 6 23:59:26 Magnon-Box smbd[1542941]: smb2_sendfile_send_data: sendfile failed for file Archive/Archive.part20.rar (Connection reset by peer) for client PID=1542941,CLIENT=REDACTED,channel=2,remote=ipv6:fd0> May 6 23:59:26 Magnon-Box smbd[144767]: [2026/05/06 23:59:26.101105, 0] ../../source3/smbd/durable.c:951(vfs_default_durable_reconnect) May 6 23:59:26 Magnon-Box smbd[144767]: vfs_default_durable_reconnect: default_durable_reconnect_fn [Archive/Archive.part20.rar] failed: NT_STATUS_OBJECT_NAME_NOT_FOUND This error came from another a whole different platform as the 2 above Edit: i did test around. It seems like the issue appears only when you write to the array and my best guess is extrem IO wait that screws it over. Now, why it happens to specific file types, thats the question

Funny, i did run into this issue as well. With a encrypted fully compressed multi part (50gb chunks) rar archive May 24 14:15:44 Magnon-Box smbd[1894656]: [2026/05/24 14:15:44.407591, 0] ../../source3/smbd/durable.c:951(vfs_default_durable_reconnect) May 24 14:15:44 Magnon-Box smbd[1894656]: vfs_default_durable_reconnect: default_durable_reconnect_fn [Archive/rar/Archive.part16.rar] failed: NT_STATUS_OBJECT_NAME_NOT_FOUND The file was absolutely present in the time of that error. Now, i think the original file in that particular archive, also gave me that error. although that was a massive 300gb single zip. Ive no clue what upsets it, i just loaded the files onto a usb external drive and dumped it that way unto unraid. Hella annoying tho.

nothing changes until the next array stop and start.

You can either click this checkbox in the unraid webgui or rename the EFI- folder to EFI Im gonna assume now that the motherboard is set to boot uefi/bios at the same time

you need to boot in UEFI mode, youre currently using legacy. Also check that Resizable BAR Support and Above 4G Decoding is enabled

Will update it in a couple hours. Makes absolutely more sense to send a notifications on failure instead of logging hey it worked.

ah yeah, good that you remind me. I did handle that in my image but didnt update it in the "official" image. although im not sure yet if i update that part yet as they wanted to publish their image themselves

it should be availabe at your unraid-ip:5000 then. Although there is a docker bug that has sneaked in, if youre using the beta version, that lets you start container whichs ports are already unavailable. I just had a look at it and it work as one would expect. Although one thing to keep in mind, since the container runs as host, if you change the port its running on, the webui button will still send you to port 5000

In the grand scheme, it isnt horrible. So there is that. But i absolutely agree that 6 months is a bit eh, to keep shipping something that has known issues, how trivial they might be. At this point, theres also the go tls issue which pilled untop of issues with that docker version, which is remedied in the 7.3 beta. That one is also hard to exploit, as a sidenote. But as time goes on, im sure more stuff will pop up and pile unto it and we all know that people set up their system like today and dont touch it and after a while, there will be more/easy ways that can be exploited. Edit: forgot to note, the versions prior to 28 are all nat unprotected. You can do funny things with docker on unraid while just being on a system within the same network.

this has nothing to do with it being slackware, its easy to update, you can do it at any point in time yourself. The issue comes with it being immutable and nobody caring enough to just update the shipped version at any point, while having ample opportunities. Edit: Since i did talk about it before in November, i want to put it here as well Unraid 7.1 beta started with Docker 27.5.1 around March 18 2025 or so, at that point, Docker version 28 was already out, it released mid February 2025. Beta2 also didnt bump it. Then comes 7.2 and nobody updated Docker to v28 either. And then came November, where i did state the cve fixes and got told, nah 7.3. Which could have been a quick update, if, at any point, we moved to 28 instead of sticking to 27 for what is over a year now.

good that you remind me, your delugevpn image is lacking the healthcheck^^