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
Although we document that this is allowed, and the ACL code handles it,
some of the validation code did not. This would cause a semi
intermittent problem because that validation code is used only in some
code paths; specifically we skip it (for now) to avoid erroring out on
some deprecated grant string formats when reading grants from the DB,
but we use it to error out when validating incoming grants.
* sdk: add host plugin protobufs and interface
This adds the protobufs and interface for the upcoming plugin-based host
feature for boundary, providing a common interface for building targets
and host catalogs off of dynamic cloud data.
* Fix signature formatting and separate out/flesh out descriptions
* sdk: re-gen with protobuf v3.15.8
* sdk: update protobuf according to @talanknight's comments
* Update internal/proto/local/plugin/host_catalog.proto
Co-authored-by: Todd Knight <T.Alan.Knight@gmail.com>
* Add TODO for host catalog persisted data for size limits
* sdk: move id to external_id and make protobuild
* Move plugin interface to GRPC service
This re-uses existing protobufs where possible, a few more messages have
been created for persisted data and for the host response where it's
important that we make it clear that the external ID is *not* the same
ID as the calculated stable ID for the host within Boundary.
* Update internal/proto/local/plugin/host_plugin_service.proto
Co-authored-by: Todd Knight <T.Alan.Knight@gmail.com>
* Move proto target to plugin/proto, fix formatting
Co-authored-by: Todd Knight <T.Alan.Knight@gmail.com>
* feat (events): Add new options needed for AuditConfig
WithAuditWrapper and WithFilterOperations all for optional paramters
when creating AuditConfig types.
* refactor (event): Change EventerConfig.Sinks to a ptr
We need to modify the SinkConfig.AuditConfig during the NewEventer factory, so we have to use a ptr not a ref.
* feat (events): Add audit configuration types
* feat (events): add audit config to sink configuration.
* feat (events): add redact/encrypt node for audit events to Eventer
Integrate the encrypt.Filter node from the eventlogger into the
event.Eventer, so audit events can be redacted/encrypted based
on data classifications.
* test (events) Add test protobufs for events
Add test protobufs along with a test service for testing
the audit event encrypt node
These fields are being refactored so they conform with the new
wh_credential_dimension. This will inform Business Intelligence tools
that these attributes conform across dimensions.
The tables in the operational part of the Boundary database (meaning
non-warehouse tables) suffix primary key columns and foreign key columns
with `_id`. Many of these columns appear in the warehouse dimension and
fact tables as normal attributes, not primary keys or foreign keys. To
disambiguate ID columns from operational tables, the primary key column
for dimension tables will be renamed from `id` to `key` and the foreign
key columns in the fact tables will be renamed to change the suffix from
`_id` to `_key`.
Now when initializing the test database for sql tests,
it will load the colors persona data into the database. This will
provide a data set that tests can use without needing to call
`wtt_load`.
The main motivation is to allow for a separation of setup and test
execution that requires crossing a transaction boundary. The pgTap tests
are required to exectue as a single transaction, so that the rollback
can reset the state, and each test can be isolated. However, can cause
issues with some test senarios, particuarly with the use of some times
releated functions like `current_timestamp` and `now` which return the
time that the current transaction started. This creates an issue if a
test relies on some setup that sets a time via `now`, then an update
that expects a new time via `now`.
See: https://www.postgresql.org/docs/13/functions-datetime.html#FUNCTIONS-DATETIME-CURRENT
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.
* website: bump packages for global style upgrades
* website: backfill missing type styles
* website: upgrade security page to use markdown-page
* website: upgrade g-container to g-grid-container
* website: upgrade community page
* website: hide alert-banner on mobile
* website: fix mismatched border colour on downloads page
* website: fix incorrect className
* website: bump to fix issue with typography on OpenAPI pages
Four new columns were added to the `wh_user_dimension` table to support
OIDC. The data type for the columns was originally `wh_dim_text` which
is a domain type with a `not null` restriction. Adding the new columns
caused the migration to fail if the `wh_user_dimension` table had any
rows in it because the existing rows would violate the `not null`
restriction.
Closes#1469
* 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>
This adds new columns to the wh_user_dimension to provide additional
data from OIDC that does not apply to Password Auth.
Note: Given the specifics of the data, it is possible these columns may
be used by other auth methods in the future, hence the lack of `oidc` in
the column names.
The specific columns are:
- auth_method_external_id
For OIDC this is the Issuer
- auth_account_external_id
For OIDC this is the Subject
- auth_account_full_name
For OIDC this is the account's Full Name, and can be `None`.
- auth_account_email
For OIDC this is the account's Email, and can be `None`.
For all of these new columns, they will be `Not Applicable` if the auth
method is Password.
This sets up the configuration for running a new suite of tests that are
written in sql, to test the sql logic implemented in plpgsql functions
and triggers. It is primarily designed to provides tests for the data
warehouse functionality that is implemented purely in sql.
The infrastructure to run these tests leverages docker. It creates one
docker container using the `library/postgres` image and provisions the
database with the boundary sql migrations and test helper functions. A
separate docker container is run that has the pgtap extension and
pg_prove cli to execute the tests.
As noted, pgtap is used for the tests. This postgresql extension
provides a number of helpful assertion functions to make writing sql
tests easier. It also, along with the pg_prove cli, provides readable
test output.
Along with the new testing infrastructure this commit establishes a
baseline of tests for the wh_user_dimension:
- Unit tests for individual views and functions:
- whx_user_dimension_source
- whx_user_dimension_target
- wh_upsert_user
- Integration tests for that verify that the wh_user_dimension gets
populated correctly when sessions are inserted or updated. This
ensures that:
- the trigger runs on session insert
- the trigger results in the correct changes to the
wh_user_dimension.
This commit provides a `Makefile` to run the tests locally that provides
support for:
- Setting the postgres docker image to easily test multiple versions.
- Passing through options to pg_prove
- Passing the test list to run individual tests
- Additional targets to separately startup the database image to quickly
re-run tests while developing new tests.
Lastly configures circleci to run this test suite as a matrix build
in parallel with the other tests. The matrix build tests multiple
versions of postgres:
- latest
- 13-alpine
- 12-alpine
- 11-alpine
See:
https://pgtap.org/
* 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>
Chris Marchesi
Jeff Mitchell
Louis Ruch
Jim
Randall Morey
Todd Knight
Rolf Sommerhalder
Timothy Messier
Pamela Bortnick
Mike Wickett
Zachary Shilton
Michael Gaffney
Taylor Dolezal
Susmitha Girumala