This change updates the pinniped CLI entrypoint to prevent browser
processes that we spawn from polluting our std out stream.
For example, chrome will print the following message to std out:
Opening in existing browser session.
Which leads to the following incomprehensible error message from
kubectl:
Unable to connect to the server: getting credentials:
decoding stdout: couldn't get version/kind; json parse error:
json: cannot unmarshal string into Go value of type struct
{ APIVersion string "json:\"apiVersion,omitempty\"";
Kind string "json:\"kind,omitempty\"" }
This would only occur on the initial login when we opened the
browser. Since credentials would be cached afterwards, kubectl
would work as expected for future invocations as no browser was
opened.
I could not think of a good way to actually test this change. There
is a clear gap in our integration tests - we never actually launch a
browser in the exact same way a user does - we instead open a chrome
driver at the login URL as a subprocess of the integration test
binary and not the pinniped CLI. Thus even if the chrome driver was
writing to std out, we would not notice any issues.
It is also unclear if there is a good way to prevent future related
bugs since std out is global to the process.
Signed-off-by: Monis Khan <mok@vmware.com>
At a high level, it switches us to a distroless base container image, but that also includes several related bits:
- Add a writable /tmp but make the rest of our filesystems read-only at runtime.
- Condense our main server binaries into a single pinniped-server binary. This saves a bunch of space in
the image due to duplicated library code. The correct behavior is dispatched based on `os.Args[0]`, and
the `pinniped-server` binary is symlinked to `pinniped-concierge` and `pinniped-supervisor`.
- Strip debug symbols from our binaries. These aren't really useful in a distroless image anyway and all the
normal stuff you'd expect to work, such as stack traces, still does.
- Add a separate `pinniped-concierge-kube-cert-agent` binary with "sleep" and "print" functionality instead of
using builtin /bin/sleep and /bin/cat for the kube-cert-agent. This is split from the main server binary
because the loading/init time of the main server binary was too large for the tiny resource footprint we
established in our kube-cert-agent PodSpec. Using a separate binary eliminates this issue and the extra
binary adds only around 1.5MiB of image size.
- Switch the kube-cert-agent code to use a JSON `{"tls.crt": "<b64 cert>", "tls.key": "<b64 key>"}` format.
This is more robust to unexpected input formatting than the old code, which simply concatenated the files
with some extra newlines and split on whitespace.
- Update integration tests that made now-invalid assumptions about the `pinniped-server` image.
Signed-off-by: Matt Moyer <moyerm@vmware.com>
This flag is (for now) meant only to facilitate end-to-end testing, allowing us to force the "manual" login flow. If it ends up being useful we can un-hide it, but this seemed like the safest option to start with.
There is also a corresponding `--oidc-skip-listen` on the `pinniped get kubeconfig` command.
Signed-off-by: Matt Moyer <moyerm@vmware.com>
This is a more restrictive library interface that more closely matches the use cases of our new form_post login flow.
Signed-off-by: Matt Moyer <moyerm@vmware.com>
- Note that this adds an extra check of the response, which is that
the issuer string in the response must match issuer of the requested
URL.
- Some of the error messages also changed to match the errors provided
by oidc.NewProvider
- And perform auto-discovery when the flags are not set
- Several TODOs remain which will be addressed in the next commit
Signed-off-by: Margo Crawford <margaretc@vmware.com>
- Also some light prefactoring in login.go to make room for LDAP-style
login, which is not implemented yet in this commit. TODOs are added.
- And fix a test pollution problem in login_oidc_test.go where it was
using a real on-disk CLI cache file, so the tests were polluted by
the contents of that file and would sometimes cause each other to
fail.
This isn't strictly necessary because we currently always have the concierge endpoint and CA as CLI flags, but it doesn't hurt and it's better to err on the side of _not_ reusing a cache entry.
Signed-off-by: Matt Moyer <moyerm@vmware.com>
Before this change, the "context", "cluster", and "user" fields in generated kubeconfig YAML were always hardcoded to "pinniped". This could be confusing if you generated many kubeconfigs for different clusters.
After this change, the fields will be copied from their names in the original kubeconfig, suffixed with "-pinniped". This suffix can be overridden by setting the new `--generated-name-suffix` CLI flag.
The goal of this change is that you can distinguish between kubeconfigs generated for different clusters, as well as being able to distinguish between the Pinniped and original (admin) kubeconfigs for a cluster.
Signed-off-by: Matt Moyer <moyerm@vmware.com>
We were previously issuing both client certs and server certs with
both extended key usages included. Split the Issue*() methods into
separate methods for issuing server certs versus client certs so
they can have different extended key usages tailored for each use
case.
Also took the opportunity to clean up the parameters of the Issue*()
methods and New() methods to more closely match how we prefer to call
them. We were always only passing the common name part of the
pkix.Name to New(), so now the New() method just takes the common name
as a string. When making a server cert, we don't need to set the
deprecated common name field, so remove that param. When making a client
cert, we're always making it in the format expected by the Kube API
server, so just accept the username and group as parameters directly.
