This test started failing when the order of the list response changed.
Since the order does not matter, this test was updated to not depend on
the order.
This change removes the hard-coded credential purpose for the Credential
Libraries. It also moves the logic of creating CredentialLibraries from
CredentialSources out of the repository and into the targets.Service.
This refactor will allow for additional credential purposes to be
supported in both the service and the repository.
This stores worker connection nonces within the database to ensure
uniqueness of connections. Prior the connection required any attacker to
be able to decrypt the incoming encrypted authentication message in
order to replay (and if the attacker has access to the KMS you have
bigger problems), but with this change even if an attacker does have
access to the KMS key, they cannot reuse the same message. (If they're
able to encrypt against the KMS key, they can still act as a worker with
a new message, so this only really helps if they only have decryption
capabilities.)
For the target Repository's Create method, the options of
WithHostSources and WithCredentialSources where never used by the
targets.Service. As such the repository does not need to support these
options.
See: 51812ff779
Now that subtypes are registered via sub-packages, this TargetType is
not needed and should not be used. This removes it and updates the
target.Repository and targets.Service to use the updated code.
This was missed in eac433f7ff.
* refactor (controller): Use a grpc server for the grpc-gateway
This starts a grpc server for the grpc-gateway. The new gateway
server uses an in-memory listener which also supports "dialing", so
the http proxy can connect to the in-memory listener.
This change allows us to use all grpc server features and was in part
motivated by a need for an interceptor of inbound requests. With that
said, having a "proper" grpc server for the gateway will provide a lot
more flexibility and capabilities moving forward.
The request context will still be verified via middleware in the
http proxy (see wrapHandlerWithCommonFuncs). This verified context
will be serialized into a protobuf and sent along to the grpc-gateway
via metadata where an interceptor will add the verified request
info to the request's context (see requestCtxInterceptor). The
interceptor will verify the serialized request info via a shared ticket
with it's companion in-memory http proxy. FYI: the ticket is simply
random base62 string (see db.NewPrivateId).
This change also necessitated a minor change to services. We need to
return an empty response instead of nil, nil when there are no errors
and the response is empty.
* refactor (handlers): Rename gRPC Gateway filter
Originally, this was called an OutgoingInterceptor, which is an
unfortunate name since it's not an interceptor. Now that we're
refactoring the gateway stuff and have proper interceptors, renaming
this should remove ambiguity
* refactor: Rework how 204 with empty msg is handled
Recent changes to grpc gateway has necessitated a change to how
204s are handled by the gateway
* tests (bats): Check status codes for delete operations
This moves all of the tcp target functionality into a tcp package. It
also establishes a Register function in the target package to allow
specific target types to be registered at initilazation time. When
registering, a target type needs to provide its Subtype, as well as some
RepositoryHooks that are used by the target Repository.
This puts it more in line with other upcoming options that use "dir"
instead of "directory" (plus it's easier to type), and makes it clear
that this directory and the passthrough behavior is related to the UI as
opposed to other upcoming items.
* feat (events): Add status code to audit events
* feat (events): Add the resp proto msg to every audit event
* chore (events): Update eventlogger dep
This latest version will redact any map and/or map field which isn't
explicitly classified via a Taggable interface
* feat (events): Begin to add tagging to response proto msgs
* refact: Changes required for gorm v2
* refact: Add db.DB wrapper and refact all test fixtures to use it
Stop leaking the gorm abstraction within our unit tests.
* refact: Migrate direct calls to pq to pgx driver
The refactor will remove the remaining direct dependencies on
the pq driver which has been deprecated in favor of the pgx
driver.
It includes the addition of common.SqlOpen which is a wrapper around
sql.Open, which is needed to deal with the driver name changing from
"postgres" to "pgx" but in our code base the dialect and driver are
synonyms.
Instead of requiring the each proxy handler to handle the webSocket
connection closure. On error the worker handler will appropriately
close the connection and event any errors.
This also affected credential libraries and managed groups. It seems
that any end-to-end tests of these grants currently had them included in
wildcards instead of individually, but these types were never added to
the valid list for resources. Additionally, credential store wasn't
added as a valid top level type for pinning purposes.
This swaps the test initialization to use the dbtest package so each
test will get a fresh database created from a template database, instead
of a completely new docker container.
* `recovery` is only used by the API (and the main module, for decryption)
* `wrapper` is only used by the CLI right now
Moving these brings them closer to the code that uses them, and makes
the SDK package a bit more clear in terms of functionality.
When a session is terminated by the user, the cancel session path is
triggered, in this path the websocket connection is closed. If the
connection is already closed, an EOF is returned. In this situation an
EOF is expected and should not be logged to the user.
Example log line:
{
"id": "ezqk6w9dOQ",
"source": "https://hashicorp.com/boundary/dev-controller/boundary-dev",
"specversion": "1.0",
"type": "error",
"data": {
"error": "failed to close WebSocket: failed to read frame header: EOF",
"error_fields": {},
"id": "e_n60YPsDkaj",
"version": "v0.1",
"op": "worker.(Worker).handleProxy",
"info": {
"msg": "error closing client connection"
}
},
"datacontentype": "text/plain",
"time": "2021-09-02T10:05:46.587513-07:00"
}
The existing logic was to ignore Forbidden errors when recursively
listing in favor of looking for downstream scopes with appropriate
permissions. The problem is that when there is actually no
authentication token at all, what comes back is an Unauthenticated
error. Therefore if you didn't have permissions on the ingress scope
(e.g. the default of `global`) but did downstream for the anonymous
user, you'd get back an error.
I don't think there's any security issue with doing this; the user could
always have simply started the search at a downstream scope to begin
with so nothing more is being revealed, this just fixes behavior to
match expectations in this case.
A test was added for one of the services; the code is exactly the same
across them. The reason that the existing tests that were added with the
output fields testing didn't catch this is that even though they were
specifying `u_anon` as the user, the disabled test context skipped the
logic that would return a 401 in this case. It turns out that fixing
this is very complicated because we don't have a great way in most tests
of defining grants to put onto the generated scopes, so even if a true
anonymous context was added each test would require the anonymous user
to have list permissions on whatever resource was under test, requiring
a lot of role and grant and principal finagling. This is probably a good
longer-term TODO but not one that seemed worth it for this fix.
Fixes ICU-2301
This moves the generated pb files to the `sdk` module. This allows third parties to make use of them if desired (e.g. in upcoming plugins) and moves away from the `gen` nomenclature in the directory, since we are going to be adding some extra interface functions to the messages.
* refactor(worker): Move tcp_proxy into its own package
This refactors:
- the tcp_proxy.go into worker/proxy/tcp
- add a missing unit test for the HandleTcpProxyV1
- refactors session.go into worker/session, needed to avoid import cycle
- adds a baseline for a mock controller session client
- go lint fixes
* Add sleep between spinning listener goroutine and dialing
Co-authored-by: Jeff Mitchell <jeffrey.mitchell@gmail.com>
* Normalize event configuration
This updates event configuration stanzas to more closely match behavior
for listeners and kmses:
* Singular "sink" instead of "sinks" (probably not the most correct
option, but a consistent option, and one that makes sense for HCL)
* Allows sinks to have the type defined at the start of the block
* Moves type-specific configuration to a specific named block
Co-authored-by: Jim <jlambert@hashicorp.com>
* add new error funcs which take a ctx and deprecate existing funcs
* use context where available when writing error events
* remove circular internal/errors pkg dep
* suppress some chatty errors using errors.WithoutEvent()
* convert auth oidc funcs to take context where needed to write events
Co-authored-by: Jim <jlambert@hashicorp.com>
This adds functionality to ensure that connections are closed on a
worker when the server is shut down or reloaded.
Connections are closed, and then the shutdown process waits for the next
report to the controller to go through, waiting the grace period that
was added in #1330. The rest of shutdown is then proceeded with.
Listener shutdown has been moved to the start of the process to ensure
that new connections can't be made while we are shutting down.
Additionally, functionality has been added to force shut down servers
after a second interrupt has been received after the first one. This
takes the form of code in both the server and dev mode code paths that
just performs an os.Exit if another shutdown is received during the
normal shutdown process.
Additionally, the commit also includes some refactoring to tests so that
we could use some of the new E2E testing code to test session
connections during shutdown and reload.
Additionally, this fixes a panic in worker startup when there are no
controllers defined. Looks like the shutdown call was attempting to call
into the controller, not the worker.
Finally, this also removes the extra shutdown within the outer worker
startup.
Shutdown is actually handled within the inner StartWorker(), attempting
shutdown on errors was actually leading to potential use of an
uninitialized worker, which is leading to panics in testing.
Timothy Messier
Jim
Jeff Mitchell
Louis Ruch
Oleg Butuzov
s-christoff
Todd Knight
Chris Marchesi