-
Two days ago I started getting an error message on one of my containers: time="2023-01-21T13:33:59Z" level=warning msg="Could not do a head request for "meisnate12/plex-meta-manager:develop", falling back to regular pull." container=/plex-meta-manager image="meisnate12/plex-meta-manager:develop" And if this helps, here is the email subject line: This container has been running fine (for months) and then this just started happening. Any ideas on what I should do? Ignore it? Is there somehting I should fix? |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 11 replies
-
This is due to a new format of the manifest that is produced with the new version of If you are responsible for creating said image, you can look at the upstream bug for ways to opt-out of the unsupported manifest until it's supported: docker/buildx#1509 (comment) Known images that uses the unsupported manifest:
|
Beta Was this translation helpful? Give feedback.
-
@piksel What do you mean by "[...]as long as you don't use a really short interval for watchtower it should be fine[...]" ? I have it set up to run once per night and I get this error every day. Or is there another setting I can tweak which I'm unaware of? That's from last night:
|
Beta Was this translation helpful? Give feedback.
This is due to a new format of the manifest that is produced with the new version of
docker buildx
. The only consequence is that the docker API rate limiting will be applied to every check for an image update, so as long as you don't use a really short interval for watchtower it should be fine. In that case you can set theWATCHTOWER_WARN_ON_HEAD_FAILURE=never
environment variable for watchtower, or add the equivalent argument--warn-on-head-failure never
(see https://containrrr.dev/watchtower/arguments/#head_failure_warnings).If you are responsible for creating said image, you can look at the upstream bug for ways to opt-out of the unsupported manifest until it's supported: docker/build…