mirror of
https://github.com/Comfy-Org/desktop.git
synced 2026-11-04 23:47:43 +00:00
[GH-ISSUE #1522] [Bug]: The electron Desktop App is updating every start now #446
Labels
No labels
area: desktop-app
Bounty
bug
bug-cop:answered
bug-cop:response-received
component: installer
component: preload
custom node
dependencies
Design
documentation
duplicate
enhancement
expected
frontend
good first issue
installation
Low-Priority
performance
Planned
platform: cross-platform
platform: windows
Potential Bug
Potential Bug
pull-request
question
refactor
Release
stale
User Support
V1
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
Comfy-Org/desktop#446
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Originally created by @quentar82 on GitHub (Jan 9, 2026).
Original GitHub issue: https://github.com/Comfy-Org/desktop/issues/1522
App Version
[Bug]: The electron Desktop App is updating every start now
working around nvidia conv3d memory bug.
WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations.
Expected Behavior
comfy updating once and its fine
Actual Behavior
comfy updating everytime you start it
Steps to Reproduce
start comfy
Debug Logs
Browser Logs
No response
Settings JSON
No response
Other
No response
┆Issue is synchronized with this Notion page by Unito
@dosubot[bot] commented on GitHub (Jan 9, 2026):
The app updating every start is likely caused by an issue with how the update setting is saved. The ComfyUI Desktop app stores its settings (including whether auto-update is enabled) in a comfy.settings.json file inside the user/default directory of your ComfyUI base path. If this file is missing, corrupted, or cannot be written (for example, due to permissions or being on a restricted path), the app will revert to default settings and re-enable auto-update every time, causing repeated updates docs.
Check that comfy.settings.json exists and is writable. Make sure your user/config directories are not in restricted locations like the app install directory, updater cache, or cloud-synced folders (like OneDrive), as these can cause permission errors issue. Review main.log and comfyui.log in AppData/Roaming/ComfyUI/logs for errors about settings or updates.
Try disabling automatic updates in the app settings and restarting to see if the behavior changes. If issues persist, back up and delete your settings/config files to let the app regenerate them, then reconfigure as needed. If you continue to see the problem, check for errors in the logs that might point to permission or config issues.
To reply, just mention @dosu.
Docs are dead. Just use Dosu.
@benceruleanlu commented on GitHub (Jan 9, 2026):
Hey @quentar82, I'm looking into this, in the meantime, do you happen to have any logs? They should be under
C:\Users\ben\AppData\Roaming\ComfyUI\logsas main.log and comfyui.log@quentar82 commented on GitHub (Jan 9, 2026):
Main
Comfy UI
@quentar82 commented on GitHub (Jan 13, 2026):
2026-01-09 18:57:56.107] [info] x No solution found when resolving dependencies:
[2026-01-09 18:57:56.107] [info]
`-> Because there is no version of torch==2.9.1+cu130 and you require
torch==2.9.1+cu130, we can conclude that your requirements are
unsatisfiable.
the executed uv command is:
[2026-01-09 18:57:55.448] [info] > & "C:\Users\info\AppData\Local\Programs@comfyorgcomfyui-electron\resources\uv\win\uv.exe" "pip" "install" "torch==2.9.1+cu130" "torchaudio==2.9.1+cu130" "torchvision==0.24.1+cu130" "--index-url" "https://download.pytorch.org/whl/cu128"
maybe you have to change index-url to cu130? as the pytorch command for this is:
pip3 install torch torchvision --index-url https://download.pytorch.org/whl/cu130
@benceruleanlu commented on GitHub (Jan 13, 2026):
As a temporary solution, can you replace the index URL https://download.pytorch.org/whl/cu128 in your user settings with https://download.pytorch.org/whl/cu130?
I think a long time ago, users were bumped to cu129, but existing users were not force updated, so if you had your installation this whole time, the user settings would still point to cu128 because there was no migration.
I think the solution here is to just mark index/mirror URLs as either manually overridden or kept default values, and allow users to choose whether or not they want to update to newer pytorch/cu versions, becauseit also depends on the user's GPU driver version.
@quentar82 commented on GitHub (Jan 14, 2026):
its working now when changing to 130 in user settings. so comfy updating once and starting normally afterwards, thank you.
still got those strange message at start:
working around nvidia conv3d memory bug.
Found comfy_kitchen backend cuda: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
Found comfy_kitchen backend triton: {'available': False, 'disabled': True, 'unavailable_reason': "ImportError: No module named 'triton'", 'capabilities': []}