Hi @atripp,

That’s great news - really glad to hear the fix is on the way! I’m excited to test it out once it’s available. Please do let me know when the update is released.

Thanks so much for the help!

Hi @rhessinger,

Thanks for taking the time to investigate this and for the detailed update.

Given that the fix likely won’t work, I’ll proceed with the workaround on my side.

Appreciate the help.

Hi @rhessinger ,

Just checking in to see if you had a chance to look into this.

And I’d like to know whether a product fix is planned, or if I should rely on the workaround.

Thanks.

Hi @rhessinger,

I currently do not have an overview on how many images without top-level mediaType.

To have an image in Nexus without top-level mediaType, you can use the oras CLI tool to manipulate the manifest.

For my testing, I used skopeo to copy an OCI image (e.g. alpine/git:v2.52.0) to Nexus.

skopeo copy --dest-creds <user>:<password> --dest-tls-verify=false docker://alpine/git:v2.52.0 docker://registry.nexus.example/docker/alpine/git:v2.52.0

Then I used the oras to:

1. fetch the manifest

   oras manifest fetch --insecure registry.nexus.example/docker/alpine/git:v2.52.0 > manifest.json
   

2. Edit the manifest.json, and remove the top-level mediaType

3. Push the manifest back to the registry

   oras manifest push --insecure --media-type application/vnd.oci.image.manifest.v1+json registry.nexus.example/docker/alpine/git:v2.52.0 manifest.json
   

After that, when inspecting the manifest with docker manifest inspect, the top-level mediaType should be missing from the image. And when trying to do migration in ProGet, it will show that the image is not found.

I hope the information shared above is useful. Thank you.

Hi @rhessinger ,

Unfortunately I do not have an example of Dockerfile to build the image.

However, after further investigation, I noticed that this only happens when the top-level mediaType is missing in the image when I used the docker manifest inspect <image> to check the manifest.

After I updated the manifest with top-level mediaType manually, the ProGet is able to identify the image and migrate it from the Nexus.

Hi @rhessinger ,

Thanks for the support. I am now able to import container images from Nexus.

However, I notice that images with OCI media type are not imported. Is that expected?

Output of execution:

INFO: Importing example:0.0.1...
WARN Docker image "example:0.0.1" not found.

The example:0.0.1 has the following value returned from the https://«NEXUS_HOST_AND_PORT»/service/rest/v1/components?repository=«REPOSITORY_NAME» API:

"contentType" : "application/vnd.oci.image.manifest.v1+json",

Hi ProGet team,

I was reading through the backup documentation and saw this statement:

"For smaller, single-server installations, you may find it easiest to simply take a snapshot of the Virtual Machine the product is running on. As long as all components are hosted on a single virtual machine, this will sufficiently backup the product."

Just wanted to clarify how this works in a restore scenario.

If the VM containing the ProGet docker deployment is backed up or snapshotted and later restored into a different cluster (with different underlying hardware), will ProGet work simply by starting the container services? Or are licenses tied to hardware-specific identifiers (CPU, MAC, etc.) that could cause issues after restoring the VM elsewhere?

Basically trying to confirm whether a full VM restore is enough, or if there are licensing steps needed after moving the VM.

Thank you.

Hi @rhessinger ,

Thanks for your help! I’m fine waiting until the release next Friday.

Hi @rhessinger,

I don’t think the reverse proxy strips the component API. In the configuration of the reverse proxy, we map a URL to a port, so when a client accesses the server using a specific URL, the request is routed to the corresponding listening port.

Thanks.

Hi @rhessinger ,

The pseudo-group/subgrouping/groups are just arbitrary name for possible parent groups.

This is a screenshot of the layout when the image are published to Nexus.

Thanks.

Hi @rhessinger,

Thanks for the reply. I am running Nexus 3.76.1.

Here's the response from the API:

{
  "items" : [ {
    "id" : "<obfuscated>",
    "repository" : "docker",
    "format" : "docker",
    "group" : "",
    "name" : "pseudo-group/subgrouping/groups/individual/project/image",
    "version" : "stable",
    "assets" : [ {
      "downloadUrl" : "https://my-nexus-server.foo.bar.com/repository/docker/v2/pseudo-group/subgrouping/groups/individual/project/image/manifests/stable",
      "path" : "/v2/pseudo-group/subgrouping/groups/individual/project/image/manifests/stable",
      "id" : "<obfuscated>",
      "repository" : "docker",
      "format" : "docker",
      "checksum" : {
        "sha256" : "<obfuscated>",
        "sha1" : "<obfuscated>"
      },
      "contentType" : "application/vnd.docker.distribution.manifest.v2+json",
      "lastModified" : "2025-12-02T12:15:33.452+00:00",
      "lastDownloaded" : null,
      "uploader" : "<obfuscated>",
      "uploaderIp" : "<obfuscated>",
      "fileSize" : <obfuscated>,
      "blobCreated" : "2025-12-02T12:15:33.454+00:00",
      "blobStoreName" : null,
      "docker" : { }
    } ]
  } ],
  "continuationToken" : null
}

The Docker API Override URL is set to https://nexus-registry.foo.bar.baz.com/v2/ which is the URL returned by the docker client.

Here's the outputs when running the migration:

info: System.Net.Http.HttpClient.Default.LogicalHandler[100]
      Start processing HTTP request GET https://nexus-registry.foo.bar.baz.com/v2/ouping/groups/individual/project/image/manifests/stable
info: System.Net.Http.HttpClient.Default.ClientHandler[100]
      Sending HTTP request GET https://nexus-registry.foo.bar.baz.com/v2/ouping/groups/individual/project/image/manifests/stable

Hope the info above is helpful. Thanks!

Hi ProGet team,

Thanks for adding support to override the Docker API URL. However, during my testing, it doesn’t seem to work as expected.

For Docker images with long paths, based on the logs, it appears that the requested URL path is truncated. I’m not sure if this is just a visual effect in the logs or if the actual request is being truncated. For example, the URL should be:

https://nexus-registry.example/v2/my-example/subgroup/container/project/subname/image/manifests/stable

but the log shows:

https://nexus-registry.example/v2/group/container/project/subname/image/manifests/stable

and it returns a 404, even though I can access the original URL directly in the browser.

Furthermore, with shorter image paths, e.g.:

https://nexus-registry.example/v2/python/manifests/3.11.2-bullseye

I don’t see the request in the log, and it fails with the following error message:

Error: Unhandled exception: System.ArgumentOutOfRangeException: startIndex cannot be larger than length of string. (Parameter 'startIndex')
   at System.String.ThrowSubstringArgumentOutOfRange(Int32 startIndex, Int32 length)
   at System.String.Substring(Int32 startIndex, Int32 length)

I would appreciate any guidance on this matter.

Thank you for your support!

Hi @rhessinger,

Our Nexus docker repository is configured to use specific port (e.g. port 8080), but this port is not exposed externally. Here's an example response returned by the Nexus endpoint:

{
  "name" : "example",
  "url" : "https://nexus-example/repository/example",
  "online" : true,
...
  "docker" : {
    "v1Enabled" : false,
    "forceBasicAuth" : true,
    "httpPort" : 8080,
    "httpsPort" : null,
    "subdomain" : null
  },
...
  "format" : "docker",
  "type" : "hosted"
}

Instead, we have a proxy server that maps a custom URL (e.g., https://nexus-docker/) to the internal Nexus port.

All Docker operations (push/pull) work correctly through this custom URL.

Could this proxy-based URL mapping affect the Docker migration feature, or is there a way to specify the external (proxy) URL for migration?

Thanks.

Hi ProGet team,

I’m trying to use the Docker migration feature in ProGet to import images from a Sonatype Nexus repository, but it doesn’t seem to be working as expected with the following error.

ERROR: Unhandled exception: System.Net.Http.HttpRequestException: Response status code does not indicate success: 404 (Not Found).

The Nexus docker repository can be found in the selection of the import dialog.

Before I proceed further, I’d like to confirm whether there are any specific requirements or limitations for this migration feature.

In our setup:

• The Nexus Docker repository is configured using a port connector.
• We have a proxy server that maps this port to a different external URL.

Would this configuration affect or break the migration process?
If so, is there a recommended way to handle such setups?

Any clarification or documentation on the expected configuration for Docker migration from Nexus would be greatly appreciated.

Thanks.

Hi @atripp,

Thanks for the reply.

Could you elaborate more on the data deduplication you mentioned?

Additionally, is there any documentation or guide on how to configure or enable this?

Thanks in advance for your help.

Hi ProGet team,

I came across the blog post mentioning that high traffic from around 50 or more users on a single-server ProGet instance can lead to issues such as HTTP request timeouts, unhandled exceptions, and temporary service unavailability.

I’d like to understand more about the performance capacity of a single ProGet instance, particularly when running the official Docker image with the embedded PostgreSQL server.

The documentation lists the recommended container sizes (up to 4 vCPU and 8 GB RAM for “Large”), but I’m wondering:

• What kind of workload or number of concurrent requests the “Large” configuration is expected to handle reliably
• Whether scaling resources beyond that point would improve performance
• Or if it’s generally more effective to move to an external PostgreSQL setup or a load-balanced configuration instead

Any guidance, documentation, or internal benchmarks on this would be greatly appreciated.

Thanks!

Hi @stevedennis,

Thanks for the update! I’m fine with waiting for the official release.

Hi @stevedennis,

I see, thanks for the explanation. However, based on the logs from running the retention rules in dry-run mode, it shows that foo.zip and bar.zip — which were newly uploaded to the assets directory and haven’t been downloaded yet — are marked for deletion.

Will this actually happen during the real run, or is it just a misleading result from the dry-run output?

Retention Rules Configuration:

Output logs:

Running in dry run mode...
Checking rule 1...
Only delete packages that have not been requested in the last 90 days (since 08/07/2025 14:50:16)
Finding packages that match retention rule 1...
Getting count of matching packages...
2 packages qualify for deletion under this rule.
Deleting matching packages...
Dry run mode is set; nothing will actually be deleted.
Deleting example/foo.zip ...
Deleting example/bar.zip ...
Deleted 2 packages (154347 KB total).
Done checking rule 1.

Hi @stevedennis,

Thanks for the input. I would like to understand more on what the command do. Does the command only reset the Administrator password (as the name suggests), or does it also recreate the Administrator user and reassign the Administer role/privileges if that privileges were removed? I want to be sure whether running it is safe when the privileges were deleted versus only when the password is lost.

Also — a small documentation/case-sensitivity note I ran into: the command shown in the docs appears as proget on Linux (image version 25.0.13), not ProGet.Service. Since Linux is case-sensitive, that difference matters and could confuse people following the instructions. Could the docs be updated or clarified?

Appreciate any details on the exact effects of the reset command. Thanks!

Hi ProGet team,

While configuring the retention rules to manage when files should be deleted, I noticed something that I’d like to confirm.

When using the setting Recent downloads: Never delete recently downloaded files, it appears that the rules are applied to newly created files that haven’t been downloaded yet.

If my understanding is correct, this means a local artifact that was recently created — but has not yet been downloaded — could still be deleted when the retention rules run.

Is this the expected behavior?
And is there a way to ensure that newly created files (e.g., files younger than a certain number of days) are not deleted by the retention rules?

Thanks in advance for any clarification or best practices!