* add key types for cmac for transit key creation
* add test for key creation
* fix test logic and add cases
* fix logic for hmac
* add go doc
* fix key size and add check for HMAC key
* Use a less strict URL validation for PKI issuing and crl distribution urls
* comma handling
* limit to ldap
* remove comma hack
* changelog
* Add unit test validating ldap CRL urls
---------
Co-authored-by: Steve Clark <steven.clark@hashicorp.com>
We have many hand-written String() methods (and similar) for enums.
These require more maintenance and are more error-prone than using
automatically generated methods. In addition, the auto-generated
versions can be more efficient.
Here, we switch to using https://github.com/loggerhead/enumer, itself
a fork of https://github.com/diegostamigni/enumer, no longer maintained,
and a fork of the mostly standard tool
https://pkg.go.dev/golang.org/x/tools/cmd/stringer.
We use this fork of enumer for Go 1.20+ compatibility and because
we require the `-transform` flag to be able to generate
constants that match our current code base.
Some enums were not targeted for this change:
When creating database connections, there is a race
condition when multiple goroutines try to create the
connection at the same time. This happens, for
example, on leadership changes in a cluster.
Normally, the extra database connections are cleaned
up when this is detected. However, some database
implementations, notably Postgres, do not seem to
clean up in a timely manner, and can leak in these
scenarios.
To fix this, we create a global lock when creating
database connections to prevent multiple connections
from being created at the same time.
We also clean up the logic at the end so that
if (somehow) we ended up creating an additional
connection, we use the existing one rather than
the new one. This by itself would solve our
problem long-term, however, would still involve
many transient database connections being created
and immediately killed on leadership changes.
It's not ideal to have a single global lock for
database connection creation. Some potential
alternatives:
* a map of locks from the connection name to the lock.
The biggest downside is the we probably will want to
garbage collect this map so that we don't have an
unbounded number of locks.
* a small pool of locks, where we hash the connection
names to pick the lock. Using such a pool generally
is a good way to introduce deadlock, but since we
will only use it in a specific case, and the purpose
is to improve performance for concurrent connection
creation, this is probably acceptable.
Co-authored-by: Jason O'Donnell <2160810+jasonodonnell@users.noreply.github.com>
* Address OCSP client caching issue
- The OCSP cache built into the client that is used by cert-auth
would cache the responses but when pulling out a cached value the
response wasn't validating properly and was then thrown away.
- The issue was around a confusion of the client's internal status
vs the Go SDK OCSP status integer values.
- Add a test that validates the cache is now used
* Add cl
* Fix PKI test failing now due to the OCSP cache working
- Remove the previous lookup before revocation as now the OCSP
cache works so we don't see the new revocation as we are actually
leveraging the cache
* add gosimport to make fmt and run it
* move installation to tools.sh
* correct weird spacing issue
* Update Makefile
Co-authored-by: Nick Cabatoff <ncabatoff@hashicorp.com>
* fix a weird issue
---------
Co-authored-by: Nick Cabatoff <ncabatoff@hashicorp.com>
* Transit: Release locks using defer statements
- Leverage defer statements to Unlock the fetched policy
to avoid issues with forgetting to manually Unlock during
each return statement
* Add cl
* Base Binary Cert and CSR Parse functions.
* Add otherSANS parsing.
* Notate what doesn't exist on a CSR.
* Fix otherSans call err-checking and add basic-constriants to CSR
* Move BasicConstraint parsing to be optionally set.
* Refactored to use existing ParseBasicConstraintsExtension.
* Add handling for the ChangeSubjectName ext on CSR that is needed for EST
* Remove ChangeSubjectName - it's an attribute, not an extension, and there is no clean way to parse it, so pair down for now.
* Make these public methods, so they can be used in vault.
* Add unit tests for certutil.ParseCertificateToCreationParameters.
Also add unit tests for certutil.ParseCertificateToFields.
* Cleanup TestParseCertificate.
* Add unit tests for certutil.ParseCsrToCreationParameters and ParseCsrToFields.
* Fix return values for "add_basic_constraints" in certutil.ParseCsrToFields.
Add a test for parsing CSRs where "add_basic_constraints" is false.
* Clear up some todos.
* Add a test for certutil.ParseCertificateToCreationParameters for non-CA cert.
* Tweak TestParseCertificate/full_non_CA_cert.
* Basics of three remaining fields - keyUsage; extKeyUsage; PolicyIdentifiers
* Fix tests and err handling
* Add unit tests for policy_identifiers; ext_key_usage_oids; key_usage
* Add test on ext_key_usage_oids
* Remove duplicate usages elsewhere.
* Add error handling to csr-checks.
* Remove extranames on returned types.
* Remove useless function.
---------
Co-authored-by: Victor Rodriguez <vrizo@hashicorp.com>
* add new plugin wif fields to AWS Secrets Engine
* add changelog
* go get awsutil v0.3.0
* fix up changelog
* fix test and field parsing helper
* godoc on new test
* require role arn when audience set
* make fmt
---------
Co-authored-by: Austin Gebauer <agebauer@hashicorp.com>
Co-authored-by: Austin Gebauer <34121980+austingebauer@users.noreply.github.com>
This commit introduces two new adaptive concurrency limiters in Vault,
which should handle overloading of the server during periods of
untenable request rate. The limiter adjusts the number of allowable
in-flight requests based on latency measurements performed across the
request duration. This approach allows us to reject entire requests
prior to doing any work and prevents clients from exceeding server
capacity.
The limiters intentionally target two separate vectors that have been
proven to lead to server over-utilization.
- Back pressure from the storage backend, resulting in bufferbloat in
the WAL system. (enterprise)
- Back pressure from CPU over-utilization via PKI issue requests
(specifically for RSA keys), resulting in failed heartbeats.
Storage constraints can be accounted for by limiting logical requests
according to their http.Method. We only limit requests with write-based
methods, since these will result in storage Puts and exhibit the
aforementioned bufferbloat.
CPU constraints are accounted for using the same underlying library and
technique; however, they require special treatment. The maximum number
of concurrent pki/issue requests found in testing (again, specifically
for RSA keys) is far lower than the minimum tolerable write request
rate. Without separate limiting, we would artificially impose limits on
tolerable request rates for non-PKI requests. To specifically target PKI
issue requests, we add a new PathsSpecial field, called limited,
allowing backends to specify a list of paths which should get
special-case request limiting.
For the sake of code cleanliness and future extensibility, we introduce
the concept of a LimiterRegistry. The registry proposed in this PR has
two entries, corresponding with the two vectors above. Each Limiter
entry has its own corresponding maximum and minimum concurrency,
allowing them to react to latency deviation independently and handle
high volumes of requests to targeted bottlenecks (CPU and storage).
In both cases, utilization will be effectively throttled before Vault
reaches any degraded state. The resulting 503 - Service Unavailable is a
retryable HTTP response code, which can be handled to gracefully retry
and eventually succeed. Clients should handle this by retrying with
jitter and exponential backoff. This is done within Vault's API, using
the go-retryablehttp library.
Limiter testing was performed via benchmarks of mixed workloads and
across a deployment of agent pods with great success.
Adds the ability to pin a version for a specific plugin type + name to enable an easier plugin upgrade UX. After pinning and reloading, that version should be the only version in use.
No HTTP API implementation yet for managing pins, so no user-facing effects yet.
* Migration of OtherSANs Parsing Call to SDK helper from pki-issuer
* Based on PR feedback from Steve, remove internal variable, reference certutil directly.
* Support reloading database plugins across multiple mounts
* Add clarifying comment to MountEntry.Path field
* Tests: Replace non-parallelisable t.Setenv with plugin env settings
* PKI refactoring to start breaking apart monolith into sub-packages
- This was broken down by commit within enterprise for ease of review
but would be too difficult to bring back individual commits back
to the CE repository. (they would be squashed anyways)
- This change was created by exporting a patch of the enterprise PR
and applying it to CE repository
* Fix TestBackend_OID_SANs to not be rely on map ordering
* Refactor plugin catalog into its own package
* Fix some unnecessarily slow tests due to accidentally running multiple plugin processes
* Clean up MakeTestPluginDir helper
* Move getBackendVersion tests to plugin catalog package
* Use corehelpers.MakeTestPlugin consistently
* Fix semgrep failure: check for nil value from logical.Storage
- Noticed that our documentation was out of date, we allow 8192
bit RSA keys to be used as an argument to the various PKI
issuer/key creation APIs.
- Augument some unit tests to verify this continues to work
* allow to skip TLS check in acme http-01 challenge
* remove configurable logic, just ignore TLS
* add changelog
* Add test case
---------
Co-authored-by: Steve Clark <steven.clark@hashicorp.com>
* Address a panic export RSA public keys in transit
- When attempting to export the public key for an RSA key that
we only have a private key for, the export panics with a nil
deference.
- Add additional tests around Transit key exporting
* Add cl
- Try to avoid these build failures as our proxy does seem to have
issues around pulling images with the 'latest' tag at times.
```
acme_test.go:206:
Error Trace: /home/runner/actions-runner/_work/vault-enterprise/vault-enterprise/builtin/logical/pkiext/pkiext_binary/acme_test.go:206
/home/runner/actions-runner/_work/vault-enterprise/vault-enterprise/builtin/logical/pkiext/pkiext_binary/acme_test.go:75
Error: Received unexpected error:
container create failed: Error response from daemon: No such image: docker.mirror.hashicorp.services/curlimages/curl:latest
Test: Test_ACME/group/caddy_http_eab
Messages: could not start cURL container
```
* wip
* more pruning
* Integrate OCSP into binary paths PoC
- Simplify some of the changes to the router
- Remove the binary test PKI endpoint
- Switch OCSP to use the new binary paths backend variable
* Fix proto generation and test compilation
* Add unit test for binary request handling
---------
Co-authored-by: Scott G. Miller <smiller@hashicorp.com>
* fix lease revocation when config token exists in one namespace but can create tokens in another
* add test
* Add similar check for admin partition
* Add admin partition test
---------
Co-authored-by: robmonte <17119716+robmonte@users.noreply.github.com>
* Forbid setting auto_rotate_period on transit managed keys
- Prevent and guard against auto-rotating managed keys as we
generate an invalid key version without the uuid field set.
- Hook in the datakey generation api into managed key encryption.
* Add cl
- Do not load existing ACME challenges persisted within storage on non-active nodes. This was the main culprit of the issues, secondary nodes would load existing persisted challenges trying to resolve them but writes would fail leading to the excessive logging.
- We now handle this by not starting the ACME background thread on non-active nodes, while also checking within the scheduling loop and breaking out. That will force a re-reading of the Closing channel that should have been called by the PKI plugin's Cleanup method.
- If a node is stepped down from being the active node while it is actively processing a verification, we could get into an infinite loop due to an ErrReadOnly error attempting to clean up a challenge entry
- Add a maximum number of retries for errors around attempting to decode,fetch challenge/authorization entries from disk. We use double the number of "normal" max attempts for these types of errors, than we would for normal ACME retry attempts to avoid collision issues. Note that these additional retry attempts are not persisted to disk and will restart on every node start
- Add a 1 second backoff to any disk related error to not immediately spin on disk/io errors for challenges.
* secrets/db: add rotation error path test
We add a test to verify that failed rotations can successfully recover
and that they do not occur outside of a rotation window. Additionally,
we remove registering some external plugins in getCluster() that shaves
off about 5 minutes the database package tests.
* remove dead code and add test comment
* revert to original container helper after refactor
This fixes the enterprise failure of the test
```
=== FAIL: builtin/logical/pki TestCRLIssuerRemoval (0.00s)
crl_test.go:1456:
Error Trace: /home/runner/actions-runner/_work/vault-enterprise/vault-enterprise/builtin/logical/pki/crl_test.go:1456
Error: Received unexpected error:
Global, cross-cluster revocation queue cannot be enabled when auto rebuilding is disabled as the local cluster may not have the certificate entry!
Test: TestCRLIssuerRemoval
Messages: failed enabling unified CRLs on enterprise
```
