This implements permissive mTLS , which allows toggling services into "permissive" mTLS mode.
Permissive mTLS mode allows incoming "non Consul-mTLS" traffic to be forward unmodified to the application.
* Update service-defaults and proxy-defaults config entries with a MutualTLSMode field
* Update the mesh config entry with an AllowEnablingPermissiveMutualTLS field and implement the necessary validation. AllowEnablingPermissiveMutualTLS must be true to allow changing to MutualTLSMode=permissive, but this does not require that all proxy-defaults and service-defaults are currently in strict mode.
* Update xDS listener config to add a "permissive filter chain" when MutualTLSMode=permissive for a particular service. The permissive filter chain matches incoming traffic by the destination port. If the destination port matches the service port from the catalog, then no mTLS is required and the traffic sent is forwarded unmodified to the application.
* add test-integrations workflow
* add test-integrations success job
* update vault integration testing versions (#16949)
* change parallelism to 4 forgotestsum. use env.CONSUL_VERSION so we can see the version.
* use env for repeated values
* match test to circleci
* fix envvar
* fix envvar 2
* fix envvar 3
* fix envvar 4
* fix envvar 5
* make upgrade and compatibility tests match circleci
* run go env to check environment
* debug docker
Signed-off-by: Dan Bond <danbond@protonmail.com>
* debug docker
Signed-off-by: Dan Bond <danbond@protonmail.com>
* revert debug docker
Signed-off-by: Dan Bond <danbond@protonmail.com>
* going back to command that worked 5 days ago for compatibility tests
* Update Envoy versions to reflect changes in #16889
* cd to test dir
* try running ubuntu latest
* update PR with latest changes that work in enterprise
* yaml still sucks
* test GH fix (localhost resolution)
* change for testing
* test splitting and ipv6 lookup for compatibility and upgrade tests
* fix indention
* consul as image name
* remove the on push
* add gotestsum back in
* removing the use of the gotestsum download action
* yaml sucks today just like yesterday
* fixing nomad tests
* worked out the kinks on enterprise
---------
Signed-off-by: Dan Bond <danbond@protonmail.com>
Co-authored-by: John Eikenberry <jae@zhar.net>
Co-authored-by: Dan Bond <danbond@protonmail.com>
Co-authored-by: Nathan Coleman <nathan.coleman@hashicorp.com>
Co-authored-by: Sarah <sthompson@hashicorp.com>
* add ability to start container tests in debug mode and attach a debugger to consul while running it.
* add a debug message with the debug port
* use pod to get the right port
* fix image used in basic test
* add more data to identify which container to debug.
* fix comment
Co-authored-by: Evan Culver <eculver@users.noreply.github.com>
* rename debugUri to debugURI
---------
Co-authored-by: Evan Culver <eculver@users.noreply.github.com>
This commit adds the PrioritizeByLocality field to both proxy-config
and service-resolver config entries for locality-aware routing. The
field is currently intended for enterprise only, and will be used to
enable prioritization of service-mesh connections to services based
on geographical region / zone.
- added Sameness Group to config entries
- added Sameness Group to subscriptions
* generated proto files
* added Sameness Group events to the state store
- added test cases
* Refactored health RPC Client
- moved code that is common to rpcclient under rpcclient common.go. This will help set us up to support future RPC clients
* Refactored proxycfg glue views
- Moved views to rpcclient config entry. This will allow us to reuse this code for a config entry client
* added config entry RPC Client
- Copied most of the testing code from rpcclient/health
* hooked up new rpcclient in agent
* fixed documentation and comments for clarity
* circleci: remove fronted jobs
Signed-off-by: Dan Bond <danbond@protonmail.com>
* remove frontend-cache
Signed-off-by: Dan Bond <danbond@protonmail.com>
---------
Signed-off-by: Dan Bond <danbond@protonmail.com>
* Add a test to reproduce the race condition
* Fix race condition by publishing the event after the commit and adding a lock to prevent out of order events.
* split publish to generate the list of events before committing the transaction.
* add changelog
* remove extra func
* Apply suggestions from code review
Co-authored-by: Dan Upton <daniel@floppy.co>
* add comment to explain test
---------
Co-authored-by: Dan Upton <daniel@floppy.co>
Prior to this change, peer services would be targeted by service-default
overrides as long as the new `peer` field was not found in the config entry.
This commit removes that deprecated backwards-compatibility behavior. Now
it is necessary to specify the `peer` field in order for upstream overrides
to apply to a peer upstream.
The old setting of 24 hours was not enough time to deal with an expiring certificates. This change ups it to 28 days OR 40% of the full cert duration, whichever is shorter. It also adds details to the log message to indicate which certificate it is logging about and a suggested action.
Paul Glass