search - Docker registry 2.0 API v2

Question

I pulled and setup the local docker registry:2.0

I have tried pushing an image successfully, but when ever I try searching for an image I get 404:

root@ip-10-232-0-153:~# curl -v -X GET http://localhost:5000/v2/search
* Hostname was NOT found in DNS cache
*   Trying 127.0.0.1...
* Connected to localhost (127.0.0.1) port 5000 (#0)
> GET /v2/search HTTP/1.1
> User-Agent: curl/7.35.0
> Host: localhost:5000
> Accept: */*
>
< HTTP/1.1 404 Not Found
< Content-Type: text/plain; charset=utf-8
< Docker-Distribution-Api-Version: registry/2.0
< Date: Fri, 08 May 2015 00:00:45 GMT
< Content-Length: 19
<
404 page not found
* Connection #0 to host localhost left intact

Also when I try to curl localhost:5000, I just get a 404:

404 page not found

Answer 1

UPDATE (14 April 2016): Still not here in the distribution roadmap, but here is a particular issue about search.

UPDATE (12 November 2015): The API endpoints still do not yet exist and are not yet in the Docker Registry roadmap.

The problem here is that the new v2 Docker registry doesn't support that particular endpoint yet, as of this question and answer. You can check the source itself for the route endpoints, and you'll see that most of the API endpoints involve simple operations like uploading and tagging, but no implementation yet of the search endpoint. It's important here to note that the v2 registry is a completely different project than the v1 registry. It's even written in a completely different language (v1 was a Python project, whereas v2 uses Go, which is more in line with the rest of the Docker projects). It took me some time and serious reading to understand the dichotomy here between the registries. It is worth looking at this particular Github issue about the v2 registry for a deep-dive into a recent discussion on the state of the v2 registry, as well as some discussion about where they've been taking it.

So there's no search endpoint in the v2 registry yet. You can list your image by tag or by the image name itself as mentioned in task number 8 in this documentation.