When running the test suite in CI (where requests are centralized from
relatively few IPs), we'd occasionally hit Dockerhub's rate limits.
Luckily Hashicorp runs a (limited) public mirror of the containers we
need, so we can switch to them here in the tests.
For consistency between developer and CI, we've opted to have the tests
always pull from the Hashicorp mirror, rather than updating the CI
runner to prefer the mirror.
We exclude nomad and influxdb as we don't presently mirror these repos.
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Clarify when -format=raw fails
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Document Vault read's new -format=raw mode
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Add raw format to usage, completion
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Add missing support for raw format field printing
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Prohibit command execution with wrong formatter
This allows us to restrict the raw formatter to only commands that
understand it; otherwise, when running `vault write -format=raw`, we'd
actually hit the Vault server, but hide the output from the user. By
switching this to a flag-parse time check, we avoid running the rest of
the command if a bad formatter was specified.
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Expose raw request from client.Logical()
Not all Vault API endpoints return well-formatted JSON objects.
Sometimes, in the case of the PKI secrets engine, they're not even
printable (/pki/ca returns a binary (DER-encoded) certificate). While
this endpoint isn't authenticated, in general the API caller would
either need to use Client.RawRequestWithContext(...) directly (which
the docs advise against), or setup their own net/http client and
re-create much of Client and/or Client.Logical.
Instead, exposing the raw Request (via the new ReadRawWithData(...))
allows callers to directly consume these non-JSON endpoints like they
would nearly any other endpoint.
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Add raw formatter for direct []byte data
As mentioned in the previous commit, some API endpoints return non-JSON
data. We get as far as fetching this data (via client.Logical().Read),
but parsing it as an api.Secret fails (as in this case, it is non-JSON).
Given that we intend to update `vault read` to support such endpoints,
we'll need a "raw" formatter that accepts []byte-encoded data and simply
writes it to the UI.
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Add support for reading raw API endpoints
Some endpoints, such as `pki/ca` and `pki/ca/pem` return non-JSON
objects. When calling `vault read` on these endpoints, an error
is returned because they cannot be parsed as api.Secret instances:
> Error reading pki/ca/pem: invalid character '-' in numeric literal
Indeed, we go to all the trouble of (successfully) fetching this value,
only to be unable to Unmarshal into a Secrets value. Instead, add
support for a new -format=raw option, allowing these endpoints to be
consumed by callers of `vault read` directly.
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Add changelog entry
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Remove panic
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Store login MFA secret with tokenhelper
* Clean up and refactor tokenhelper paths
* Refactor totp test code for re-use
* Add login MFA command tests
* Use longer sleep times and sha512 for totp test
* Add changelog
Vault's new TLS devvault mode has two nits with certificate
construction:
1. The CA doesn't need to include any SANs, as these aren't checked.
Technically this means the CA could be reused as a leaf certificate
for the one specified IP SAN, which is less desirable.
2. Add hostname to SANs in addition to CNs. This is a best practice, as
(when the CN is a hostname), it is preferable to have everything in
SANs as well.
Neither of these are major changes.
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Add patch support to CLI
This is based off the existing write command, using the
JSONMergePatch(...) API client method rather than Write(...), allowing
us to update specific fields.
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Add documentation on PATCH support
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Add changelog
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
This fixes a couple of references to loop variables in parallel tests
and deferred functions. When running a parallel test (calling
`t.Parallel()`) combined with the table-driven pattern, it's necessary
to copy the test case loop variable, otherwise only the last test case
is exercised. This is documented in the `testing` package:
https://pkg.go.dev/testing#hdr-Subtests_and_Sub_benchmarks
`defer` statements that invoke a closure should also not reference a
loop variable directly as the referenced value will change in each
iteration of the loop.
Issues were automatically found with the `loopvarcapture` linter.
* Added flag and env var which will disable client redirection
* Added changelog
* Docs fix for unsaved file, and test single request made
* Updated test for case when redirect is enabled, updated docs based on suggestions
* Write explicit -help output to stdout
Per the consensus of most programs, and mirroring the GNU Coding
Standards for CLI design, when users request -help explicitly via the
CLI, this should be written to stdout to allow paging of output. stderr
is fine when an invalid usage triggers the help text however.
In our case, mitchellh/cli helpfully adds a HelpWriter that we
previously set to stderr explicitly. This writer is only called to print
user-requested help text; it is not called on error cases (e.g., bad
usage triggering additional help text to the user).
Thus it should safely be settable to stdout, enabling pagers without
additional redirects.
Notably, we do have to set ErrorWriter as it defaults to initializing to
the value of HelpWriter, which we want to set to stdout now, which is
less useful.
See also: https://www.gnu.org/prep/standards/html_node/_002d_002dhelp.htmlResolves: #17004
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Add changelog
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
Signed-off-by: Alexander Scheel <alex.scheel@hashicorp.com>
* Add -plugin-version flag to vault auth/secrets tune
* CLI tests for auth/secrets tune
* CLI test for plugin register
* Plugin catalog listing bug where plugins of different type with the same name could be double counted
* Use constant for -plugin-version flag name
* core: Handle deprecated mounts on enable and unseal
* changelog: Deprecation Status handling
* core: Add Pending Removal override var
* core: Add some documentation for Pending Removal override
* HCP link integration
* update configure-git.yml
* more OSS stuff
* removing internal repos
* adding a nil check
* removing config test to be included in ENT only
* updating hcp-sdk-go to v0.22.0
* remove Hostname and AuthURL link config params
Co-authored-by: Chris Capurso <1036769+ccapurso@users.noreply.github.com>
* auth: Add Deprecation Status to auth list -detailed
* secrets: Add Deprecation Status to secrets list -detailed
* Add changelog entry for deprecation status list
* OSS portion of wrapper-v2
* Prefetch barrier type to avoid encountering an error in the simple BarrierType() getter
* Rename the OveriddenType to WrapperType and use it for the barrier type prefetch
* Fix unit test
Chris Capurso