[GH-ISSUE #37] Plugin error after a reboot #14

Closed
opened 2026-10-15 12:54:35 +00:00 by bsakaguchi · 20 comments
Owner

Originally created by @bpivk on GitHub (May 22, 2026).
Original GitHub issue: https://github.com/games-on-whales/unraid-plugin/issues/37

Is there a reason why the plugin goes into plugins with errors after a reboot?

I can install it fine and the docker service persists but each time I reboot my server the plugin vanishes and If I go under plugins I find it errored out and was disabled.

Image
Originally created by @bpivk on GitHub (May 22, 2026). Original GitHub issue: https://github.com/games-on-whales/unraid-plugin/issues/37 Is there a reason why the plugin goes into plugins with errors after a reboot? I can install it fine and the docker service persists but each time I reboot my server the plugin vanishes and If I go under plugins I find it errored out and was disabled. <img width="1743" height="124" alt="Image" src="https://github.com/user-attachments/assets/b7da1846-a4d7-4f7d-a4d5-20870b031a92" />
Author
Owner

@JBailes commented on GitHub (May 22, 2026):

Could you provide more detail? Perhaps a log?

I would be very happy to fix this, but a log would be extremely helpful, as would hardware information about your server.

<!-- gh-comment-id:4518561497 --> @JBailes commented on GitHub (May 22, 2026): Could you provide more detail? Perhaps a log? I would be very happy to fix this, but a log would be extremely helpful, as would hardware information about your server.
Author
Owner

@bpivk commented on GitHub (May 22, 2026):

Where can I find the logs? I'm searching for them but I can't see any errors anywhere.

My server info:

<html> Model: |   -- | -- M/B: | Jingsha X99 Dual Version BQ36 V1.1 s/n MJ202011204 Chassis: | s/n MJ202011204 BIOS: | American Megatrends Inc. Version BQ36RD24 Dated 11/20/2020 CPU: | Intel® Xeon® CPU E5-2666 v3 @ 2.90GHz GPU: | Nvidia GTX 1070 HVM: | Enabled IOMMU: | Enabled Cache: | CPU Internal L1: 640 KiB, CPU Internal L2: 2560 KiB, CPU Internal L3: 25 MiB, CPU Internal L1: 640 KiB, CPU Internal L2: 2560 KiB, CPU Internal L3: 25 MiB Memory: | 64 GiB DDR4 Multi-bit ECC (max. installable capacity 1536 GiB) Network: | bond0: IEEE 802.3ad Dynamic link aggregation, mtu 1500 Kernel: | Linux 6.18.29-Unraid x86_64 OpenSSL: | 3.5.6 </html>
<!-- gh-comment-id:4518579955 --> @bpivk commented on GitHub (May 22, 2026): Where can I find the logs? I'm searching for them but I can't see any errors anywhere. My server info: <html> <body> <!--StartFragment--> Model: |   -- | -- M/B: | Jingsha X99 Dual Version BQ36 V1.1 s/n MJ202011204 Chassis: | s/n MJ202011204 BIOS: | American Megatrends Inc. Version BQ36RD24 Dated 11/20/2020 CPU: | Intel® Xeon® CPU E5-2666 v3 @ 2.90GHz GPU: | Nvidia GTX 1070 HVM: | Enabled IOMMU: | Enabled Cache: | CPU Internal L1: 640 KiB, CPU Internal L2: 2560 KiB, CPU Internal L3: 25 MiB, CPU Internal L1: 640 KiB, CPU Internal L2: 2560 KiB, CPU Internal L3: 25 MiB Memory: | 64 GiB DDR4 Multi-bit ECC (max. installable capacity 1536 GiB) Network: | bond0: IEEE 802.3ad Dynamic link aggregation, mtu 1500 Kernel: | Linux 6.18.29-Unraid x86_64 OpenSSL: | 3.5.6 <!--EndFragment--> </body> </html>
Author
Owner

@JBailes commented on GitHub (May 22, 2026):

Ah hah, I think it's an older nvidia card issue.

I think I've got what I need, I'll have a PR out to try to fix this in a couple of hours!

<!-- gh-comment-id:4518588036 --> @JBailes commented on GitHub (May 22, 2026): Ah hah, I think it's an older nvidia card issue. I think I've got what I need, I'll have a PR out to try to fix this in a couple of hours!
Author
Owner

@bpivk commented on GitHub (May 22, 2026):

Also this as another error:

May 22 12:16:40 Tower avahi-daemon[85005]: *** WARNING: Detected another IPv4 mDNS stack running on this host. This makes mDNS unreliable and is thus not recommended. ***

root@Tower:~# ss -ulpn | grep 5353
UNCONN 0 0 0.0.0.0:5353 0.0.0.0:* users:(("avahi-daemon",pid=85005,fd=14))
UNCONN 0 0 0.0.0.0:5353 0.0.0.0:* users:(("wolf",pid=15598,fd=43))
UNCONN 0 0 *:5353 : users:(("wolf",pid=15598,fd=51))

Could this also cause problems?

<!-- gh-comment-id:4518602295 --> @bpivk commented on GitHub (May 22, 2026): Also this as another error: May 22 12:16:40 Tower avahi-daemon[85005]: *** WARNING: Detected another IPv4 mDNS stack running on this host. This makes mDNS unreliable and is thus not recommended. *** root@Tower:~# ss -ulpn | grep 5353 UNCONN 0 0 0.0.0.0:5353 0.0.0.0:* users:(("avahi-daemon",pid=85005,fd=14)) UNCONN 0 0 0.0.0.0:5353 0.0.0.0:* users:(("wolf",pid=15598,fd=43)) UNCONN 0 0 *:5353 *:* users:(("wolf",pid=15598,fd=51)) Could this also cause problems?
Author
Owner

@JBailes commented on GitHub (May 22, 2026):

Oh, thank you! That is possible, I'm starting to look into this!

<!-- gh-comment-id:4518604716 --> @JBailes commented on GitHub (May 22, 2026): Oh, thank you! That is possible, I'm starting to look into this!
Author
Owner

@JBailes commented on GitHub (May 22, 2026):

Thanks for the extra details. PR #38 has been merged: https://github.com/games-on-whales/unraid-plugin/pull/38

What changed:

  • Plugin boot replay now treats Docker/network/NVIDIA driver availability as warnings instead of disabling the plugin while Unraid is still starting services.
  • The /boot/config/go autostart hook now waits for Docker and GPU device nodes before running docker compose.
  • NVIDIA compose devices are generated from the /dev/nvidia* nodes that actually exist, which should help older or variant NVIDIA layouts.
  • The settings page now warns when Avahi and Wolf are both bound to UDP 5353, since that can make Moonlight auto-discovery unreliable.

Leaving this issue open for reporter confirmation after testing.

<!-- gh-comment-id:4518737022 --> @JBailes commented on GitHub (May 22, 2026): Thanks for the extra details. PR #38 has been merged: https://github.com/games-on-whales/unraid-plugin/pull/38 What changed: - Plugin boot replay now treats Docker/network/NVIDIA driver availability as warnings instead of disabling the plugin while Unraid is still starting services. - The /boot/config/go autostart hook now waits for Docker and GPU device nodes before running docker compose. - NVIDIA compose devices are generated from the /dev/nvidia* nodes that actually exist, which should help older or variant NVIDIA layouts. - The settings page now warns when Avahi and Wolf are both bound to UDP 5353, since that can make Moonlight auto-discovery unreliable. Leaving this issue open for reporter confirmation after testing.
Author
Owner

@bpivk commented on GitHub (May 22, 2026):

One ides that might help a lot with port 5353 issue...
Is there a way to reconfigure the plugin to add a br0 (custom IP) option there? If docker creates ports on a custom IP there won't be any clashes with unraid services and ports used.

<!-- gh-comment-id:4518827984 --> @bpivk commented on GitHub (May 22, 2026): One ides that might help a lot with port 5353 issue... Is there a way to reconfigure the plugin to add a br0 (custom IP) option there? If docker creates ports on a custom IP there won't be any clashes with unraid services and ports used.
Author
Owner

@JBailes commented on GitHub (May 22, 2026):

The problem is there that the moonlight protocol itself uses 5353, that's not our protocol.

One option that I might consider is giving Wolf itself an IP, and that way there would be no port conflict. But that's something I have in mind.

<!-- gh-comment-id:4518911441 --> @JBailes commented on GitHub (May 22, 2026): The problem is there that the moonlight protocol itself uses 5353, that's not our protocol. One option that I might consider is giving Wolf itself an IP, and that way there would be no port conflict. But that's something I have in mind.
Author
Owner

@bpivk commented on GitHub (May 22, 2026):

The problem is there that the moonlight protocol itself uses 5353, that's not our protocol.

One option that I might consider is giving Wolf itself an IP, and that way there would be no port conflict. But that's something I have in mind.

That's exactly what I was referring to with the custom IP. If a user can enter their own IP interface then the whole docker stack can run without conflicts.

<!-- gh-comment-id:4518927361 --> @bpivk commented on GitHub (May 22, 2026): > The problem is there that the moonlight protocol itself uses 5353, that's not our protocol. > > One option that I might consider is giving Wolf itself an IP, and that way there would be no port conflict. But that's something I have in mind. That's exactly what I was referring to with the custom IP. If a user can enter their own IP interface then the whole docker stack can run without conflicts.
Author
Owner

@JBailes commented on GitHub (May 23, 2026):

Follow-up update: PR #39 has been merged: https://github.com/games-on-whales/unraid-plugin/pull/39

This builds on PR #38 by adding the custom-IP path discussed above:

  • The settings GUI can now put Wolf on a custom Docker network such as br0.
  • A static Wolf IPv4 can be configured so Moonlight can connect to Wolf on that IP instead of the Unraid host IP.
  • WOLF_INTERNAL_IP is set from that static IP so Wolf advertises the selected address back to Moonlight.
  • Bridge mode is also available and publishes Wolf's default Moonlight ports, though custom br0/static IP is the intended option for avoiding host port and mDNS conflicts.

Leaving this issue open for reporter confirmation after testing.

<!-- gh-comment-id:4524917428 --> @JBailes commented on GitHub (May 23, 2026): Follow-up update: PR #39 has been merged: https://github.com/games-on-whales/unraid-plugin/pull/39 This builds on PR #38 by adding the custom-IP path discussed above: - The settings GUI can now put Wolf on a custom Docker network such as br0. - A static Wolf IPv4 can be configured so Moonlight can connect to Wolf on that IP instead of the Unraid host IP. - WOLF_INTERNAL_IP is set from that static IP so Wolf advertises the selected address back to Moonlight. - Bridge mode is also available and publishes Wolf's default Moonlight ports, though custom br0/static IP is the intended option for avoiding host port and mDNS conflicts. Leaving this issue open for reporter confirmation after testing.
Author
Owner

@JBailes commented on GitHub (May 23, 2026):

@bpivk It would be fantastic if I could get verification from your end that the two PRs fix the issues!

<!-- gh-comment-id:4525685875 --> @JBailes commented on GitHub (May 23, 2026): @bpivk It would be fantastic if I could get verification from your end that the two PRs fix the issues!
Author
Owner

@bpivk commented on GitHub (May 23, 2026):

I'm sorry. I'm not at my server today. Wife dragged me to an event.

Is the update on the main branch or where can I download it?

<!-- gh-comment-id:4525743549 --> @bpivk commented on GitHub (May 23, 2026): I'm sorry. I'm not at my server today. Wife dragged me to an event. Is the update on the main branch or where can I download it?
Author
Owner

@JBailes commented on GitHub (May 23, 2026):

I'm sorry. I'm not at my server today. Wife dragged me to an event.

Is the update on the main branch or where can I download it?

Yep, update is on main. Both PRs were merged to main, they work on my test environment. Just waiting on you to verify!

<!-- gh-comment-id:4525773967 --> @JBailes commented on GitHub (May 23, 2026): > I'm sorry. I'm not at my server today. Wife dragged me to an event. > > Is the update on the main branch or where can I download it? Yep, update is on main. Both PRs were merged to main, they work on my test environment. Just waiting on you to verify!
Author
Owner

@bpivk commented on GitHub (May 23, 2026):

I may be blind but I can't find the latest version. The last posted one is from 2 days ago.

I am browsing through a phone though on a VPN in a shoddy network location. Not an experience I'd recommend.

<!-- gh-comment-id:4525804048 --> @bpivk commented on GitHub (May 23, 2026): I may be blind but I can't find the latest version. The last posted one is from 2 days ago. I am browsing through a phone though on a VPN in a shoddy network location. Not an experience I'd recommend.
Author
Owner

@bpivk commented on GitHub (May 23, 2026):

I did reinstall to check if the script changes anything but I can't see any changes in the gui itself.

Image
<!-- gh-comment-id:4525807299 --> @bpivk commented on GitHub (May 23, 2026): I did reinstall to check if the script changes anything but I can't see any changes in the gui itself. <img width="1115" height="741" alt="Image" src="https://github.com/user-attachments/assets/9cc9bcd7-ae9a-4776-b38f-9ee1ee9b117a" />
Author
Owner

@JBailes commented on GitHub (May 23, 2026):

Hmm. I'm taking a look, it sounds like Unraid isn't picking up the latest release...or it may take a bit of time for it to do that.

<!-- gh-comment-id:4525822567 --> @JBailes commented on GitHub (May 23, 2026): Hmm. I'm taking a look, it sounds like Unraid isn't picking up the latest release...or it may take a bit of time for it to do that.
Author
Owner

@bpivk commented on GitHub (May 23, 2026):

OK I'm home now. The problem lies here:

Downloading the latest version like this: https://raw.githubusercontent.com/games-on-whales/unraid-plugin/refs/heads/main/gow.plg

Gives this error:
Image

Because the file is pointing to here:

&gitPkgURL;/scripts/vars.sh

And it even states so in your script:
Scripts: downloaded from the tagged release so hashes match exactly.
During development, point gitPkgURL at your local HTTP server (see DEVELOPING.md).

So I can't test it unless I either run a http server (which is a hassle if you want to test it with multiple people or we will all need to make servers) or alternatively unless you create a branch or something where we can install the files the old fashioned way.

<!-- gh-comment-id:4526439590 --> @bpivk commented on GitHub (May 23, 2026): OK I'm home now. The problem lies here: Downloading the latest version like this: https://raw.githubusercontent.com/games-on-whales/unraid-plugin/refs/heads/main/gow.plg Gives this error: <img width="633" height="267" alt="Image" src="https://github.com/user-attachments/assets/428b5519-cbee-4304-90fc-98d6f070ddd7" /> Because the file is pointing to here: <FILE Name="&script_dir;/vars.sh"> <URL>&gitPkgURL;/scripts/vars.sh</URL> </FILE> And it even states so in your script: Scripts: downloaded from the tagged release so hashes match exactly. During development, point gitPkgURL at your local HTTP server (see DEVELOPING.md). So I can't test it unless I either run a http server (which is a hassle if you want to test it with multiple people or we will all need to make servers) or alternatively unless you create a branch or something where we can install the files the old fashioned way.
Author
Owner

@JBailes commented on GitHub (May 24, 2026):

Woah,

Thanks for the information! I am fixing that immediately!

<!-- gh-comment-id:4527363853 --> @JBailes commented on GitHub (May 24, 2026): Woah, Thanks for the information! I am fixing that immediately!
Author
Owner

@JBailes commented on GitHub (May 24, 2026):

Fix pushed and merged! Thank you @bpivk, I should have fully automated the release process now!

<!-- gh-comment-id:4527379556 --> @JBailes commented on GitHub (May 24, 2026): Fix pushed and merged! Thank you @bpivk, I should have fully automated the release process now!
Author
Owner

@bpivk commented on GitHub (May 24, 2026):

I can confirm that the plugin does not error out any more. Also the br0 setup works.

Image
<!-- gh-comment-id:4529908944 --> @bpivk commented on GitHub (May 24, 2026): I can confirm that the plugin does not error out any more. Also the br0 setup works. <img width="893" height="267" alt="Image" src="https://github.com/user-attachments/assets/be2037b8-bf5e-486d-9f68-84b18bc91946" />
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
games-on-whales/unraid-plugin#14
No description provided.