But that clarified that the basic auth credentials are somehow not being used.

But since posting, the newest release versions of Docker Engine, Swarm (and possibly Distribution) seem to have eliminated the need for me to specify the X-Registry-Auth header in the.dockerconfig.json file.

On a whim I took it out.) Now pulls across the swarm work with both images from my private registry server and public images from Docker Hub.

When I check the swarm worker logs its saying the image was not found.

But if I run the same on swarm worker directly its working fine.

Another thing is, if I pull the image manually on all swarm workers and keep it available, then the docker service create is successfully creating the containers across all swarm workers.

I am behind the firewall and proxy and not able to use public docker hub for testing.

The daemon starts normally, images can be pulled and pushed, but upon starting a container with docker run.

There are a couple of issues with the data-only containers approach.

Basically the need is for a persistent named set of data volumes which are initialised with the contents of respective dirs in the base image.

However data-only containers add noise to both the list of containers and the list of images, and prevent me from assigning meaningful names to data volumes.rnrnConsider these two scenarios:rnrn1.

I need a persistent volume for storing uploaded images or I want to deploy my app codejars somewhere but still be able to re-tweak the container using it.rn2.

Create a named data volume.rn2.

I create this volume, I want it initialised from varlibmysql of a particular image.

Create a data only container.rn2.

Use the --volumes-from option when creating the app container.rnrnIn this case:rn Applications Desktop Documents Downloads Google Drive Library Movies Music Pictures Public src docker ps shows x2 number of containers, and unless I choose a good naming scheme could end up very messy.rn Applications Desktop Documents Downloads Google Drive Library Movies Music Pictures Public src docker volume ls shows a bunch of random ids that tell me nothing about whats inside those volumes or where Ive mounted them.rn Applications Desktop Documents Downloads Google Drive Library Movies Music Pictures Public src docker inspect again shows the random ids which are less helpful.rn Applications Desktop Documents Downloads Google Drive Library Movies Music Pictures Public src docker images will show unnamed images, since if I make changes to the base image, I cant easily re-create the data-only container to use the new image with the existing volumes.

This container sticks around pointing to the old image.rnrn----rnrnThere are a couple of ways I can think of that can help here:rnrn1.

This isnt the ideal solution, but I can easily script the rest.

However even if using a different solution, I think this operation should be available regardless.rn2.

Add some option to initialise a new volume from a declared volume of some image.

The problem here is that an image can declare several volumes and they have no names, so will have to be referred to by path.

Not as convenient as the --volumes-from switch.rn3.

Add or tweak existing options to run or create to give names to certain volumes when they are auto-created and initialised just as they are now.

This will replace the need for --volumes-from and data-only containers altogether.

For example:rnrn docker run -d --name us.mysql -v us.data:varlibmysql mysql rn docker run -d --name eu.mysql -v eu.data:varlibmysql mysql rnrn Will result in the creation of two data volumes named us.data and eu.data and their init from the contents of varlibmysql of the mysql image if they dont exist, or mount the existing ones if they already exist.rn4.

Another option I thought of is declaring data volume sets or groups.

In this case the commands might look something like:rnrn docker volume create-set --name us.mysql.data mysqlrn docker run -d --name us.mysql --volumes-from us.mysql.data mysqlrnrn In this case the --volumes-from refers to the set of data volumes which are collectively named rather than individually named.

Its a temporary workaround for TP4 until the right fix can be made to the Windows code.

This fixes the workflow issue mentioned in rnrnTibor, could we get this marked for TP4rnrnSigned-off-by: Stefan J.

Command returns: unauthorized: access to the requested resource is not authorized.rnrnActually, checking the code, I could not detect an actual problem with the dockerdocker but maybe its an registry issue.rnrnIm proceed into double checking the functionality with other Registry.rnrnMy examples used the basic Getting Started Guide for the sake of understanding: rnrnActually Im a docker newbie but in love with the project and looking forward into contributing.

If the argument changes then the cache is busted at the first usage of that argument (not at the argument declaration)rn2.

The cache remembers previous invocations (not just the last one).

This appears to be because of selinux support, running the container with --security-opt label:disable fixes this.rnrnLooking deeper into this, I believe this is related to syscalllinux:138( ) where the daemon tries to set an selinux context on the volume path.

Docker For Mac Disable-Legacy-Registry