* Reference PreferredEndpoints by top level struct name instead of embedded proto GetPreferredEndpoint method.
Reorganize the sql files a little.
* Make PreferredEndpoints replayable. Test refactoring to remove variables that persist across test calls. DeleteItems instead of trying to use WithWhere.
This adds a parameter to host sets allowing the sync interval to be
user-specified on a per-host-set basis. A value of -1 disables syncing;
a value of null uses the current default.
* internal/host/plugin: add UpdateSet
This adds the UpdateSet repository function.
UpdateSet runs the OnUpdateSet plugin call within the transaction as per
the changes in #1683.
The plugin call returns the updated set, the hosts in the set, the rows
updated, and any error.
* Ensure that attributes can be cleared by supplying a full null
* move version increment to the end and track perferred endpoint update separately
* move plugin call to end of transaction
* only clone returned set once, and move host fetch call out of transaction
* Request a sync when attributes are updated
* Add test to assert that old preferred endpoints are being removed correctly
* Only call the Plugin at create/update time after the db has been partially updated. This allows more constraints to be checked prior to calling the plugin.
* Create catalog service handler for updating plugin. Add tests for handler and run make fmt. Silently drop attributes where the value is nil when creating the catalog or set.
* chore: Update eventlogger dependency.
We need the latest version, so we can limit the type of events
that are signed.
* fix (events): Only "sign" audit events
This should help prevent issues like the one fixed in
f2ad4de554.
A version for a particular edition should only exist once, previously
which ever file was read last for a version would be included, which
would lead to some migrations not running.
* internal/host/plugin: Add UpdateCatalog
This adds support for updating plugin-based host catalogs.
The workflow is:
* The existing catalog is fetched.
* Any attribute changes are patched against it, along with other changes
such as name and description.
* Both the existing catalog, and the state of the catalog as it will
exist post-update, along with any persisted secrets, are sent to the
plugin via an OnUpdateCatalog request.
Upon success, the following is done:
* The record is updated if non-secret updates need to be made.
* Secrets are updated if they still exist.
* If an empty secret set was sent back, secrets are instead deleted.
* If the secrets field was instead null, secret update is skipped
altogether.
There has also been a lot of helper code added to the tests to help make
the update tests more readable and reduce boilerplate. While the create
tests have not been updated to use these new helpers, it might be a good
idea to do so for consistency.
* move secret/persisted state closer to plugin call
* Update internal/host/plugin/repository_host_catalog.go
Co-authored-by: Todd <T.Alan.Knight@gmail.com>
* Update internal/host/plugin/repository_host_catalog_test.go
Co-authored-by: Todd <T.Alan.Knight@gmail.com>
* move all test helpers for UpdateCatalog inline
* WIP - transparent attributes struct/wire conversion
* Revert "WIP - transparent attributes struct/wire conversion"
This reverts commit 31af1b1b7e535119aa23bd0aa936da1a80eb5aa2.
* deal with multiple attribute masks
* Note TODO for upsertQuery for HostCatalogSecret
* add version check guard
* Update internal/host/plugin/host_catalog_secret.go
Co-authored-by: Todd <T.Alan.Knight@gmail.com>
* remove redundant JIRA reference
* don't separate number of updated catalogs and secrets
* increment version when secrets is the only thing being updated
* add another test for updating both catalog and secrets at once
* correct comment for mock plugin manager section
* add first-part path matching for secrets
* update plugin lookup error to return errors.Internal
* only add name/description to plugin request if they are non-empty
* update duplicate tests to use two project scopes
* remove unused test checks
* return plugin record along with host catalog in update
* move away from using passed in catalog after patching newCatalog
* use t.Cleanup() instead of defer
Co-authored-by: Todd <T.Alan.Knight@gmail.com>
* Update the mask manager to always allow fields containing a prefix which matchs a passedThroughPrefix.
This will be useful when updating plugin based host catalog and sets since attributes and secrets cannot be contained in the mask manager mapping.
* feature (db): Add the ability to enable debug output for a connection
* refactor: Use scope.Global instead of "global"
* feature (db/migrations): Add kms audit keys
Add schema changes to support a new DEK for audit crypto operations.
* feature (kms/audit): Add audit key type
All the bits required to define a new DEK concrete type for
operating on audit entries.
* refactor (db): Remove deprecated migration unit tests.
We added a new audit DEK to kms.CreateKeysTx(...). Unfortunately,
unit tests like TestMigrations_UserDimension,
Test_PrimaryAuthMethodChanges and TestMigrations_CredentialDimension
have an indirect dependency on a specific historical set of DEKs via
an indirect call kms.CreateKeysTx. This of course fails because the
new DEK's tables is not a part of pinned migration versions that these
tests depend on.
* Add ability to read and list hosts. Tests to follow.
* Add tests for reading and listing as well as validation checks to not modify hosts.
* Add IpAddress and DnsNames to the api resource. Populate the plugin info.
* Adding go sdk tests for reading and listing hosts and ensuring update and deleting is disallowed.
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.
Target subtypes must now implement another repository hook to validate
any CredentialLibraries that are being added/set. This allows for
subtypes to determine the valid set of CredentialLibraries.
For tcp.Targets only CredentialLibaries with
credential.ApplicationPurpose are allowed. If any other purpose is
provided an error is returned.
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.)
* refactor (oidc): Stop emitting error for not found token id
Stop emitting errors when authtoken repo.IssueAuthToken is called and
there's if no pending token. It's not an "normal" state and not
an error condition.
* feature (audit tagging): Add tags for auth method requests/responses
Including testing functions for asserting audit events created
when a service is called
Todd
Jim
Jeff Mitchell
Chris Marchesi
Timothy Messier
Louis Ruch