* update TestHTTPHandlers_AgentMetrics_LeaderShipMetrics to use 3 servers instead of 2 to allow quorom when leadership flails.
* properly sequence defers
* dns v2 - both empty string and default should be allowed for namespace and partition in Ce
* add changelog
* use default partition constant
* use constants in validation.
---------
Co-authored-by: Michael Zalimeni <michael.zalimeni@hashicorp.com>
* NET-5879 - move the filter for non-passing to occur in the health RPC layer rather than the callers of the RPC
* fix import of slices
* NET-5879 - expose sameness group param on service health endpoint and move sameness group health fallback logic into HealthService RPC layer
* fixing deepcopy
* fix license headers
This refactors and relocates the following packages to live under internal/gossip instead of either in the toplevel lib or agent/consul:
- librtt : related to serf coordinates
- libserf : random serf stuff
* Define file-system-certificate config entry
* Collect file-system-certificate(s) referenced by api-gateway onto snapshot
* Add file-system-certificate to config entry kind allow lists
* Remove inapplicable validation
This validation makes sense for inline certificates since Consul server is holding the certificate; however, for file system certificates, Consul server never actually sees the certificate.
* Support file-system-certificate as source for listener TLS certificate
* Add more required mappings for the new config entry type
* Construct proper TLS context based on certificate kind
* Add support or SDS in xdscommon
* Remove unused param
* Adds back verification of certs for inline-certificates
* Undo tangential changes to TLS config consumption
* Remove stray curly braces
* Undo some more tangential changes
* Improve function name for generating API gateway secrets
* Add changelog entry
* Update .changelog/20873.txt
Co-authored-by: Jared Kirschner <85913323+jkirschner-hashicorp@users.noreply.github.com>
* Add some nil-checking, remove outdated TODO
* Update test assertions to include file-system-certificate
* Add documentation for file-system-certificate config entry
Add new doc to nav
* Fix grammar mistake
* Rename watchmaps, remove outdated TODO
---------
Co-authored-by: Melisa Griffin <melisa.griffin@hashicorp.com>
Co-authored-by: Jared Kirschner <85913323+jkirschner-hashicorp@users.noreply.github.com>
This operation would previously fail due to unconsumed bytes in the
decoder buffer when reading the Ent snapshot (the first byte of the
record would be misinterpreted as a type indicator, and the remaining
bytes would fail to be deserialized or read as invalid data).
Ensure restore succeeds by decoding the ignored record as an
interface{}, which will consume the record bytes without requiring a
concrete target struct, then moving on to the next record.
* update go-control-plane envoy dependency to 0.12.0
* add changelog
* go mod tidy
* fix linting issues
* add agent/grpc-internal to the list of SA1019 ignores
* Include SNI + root PEMs from peered cluster on terminating gw filter chain
This allows an external service registered on a terminating gateway to be exported to and reachable from a peered cluster
* Abstract existing logic into re-usable function
* Regenerate golden files w/ new listener logic
* Add changelog entry
* Use peering bundles that are stable across test runs
* put conditionals are hcp initialization for consul server
* put more things behind configuration flags
* add changelog
* TestServer_hcpManager
* fix TestAgent_scadaProvider
Currently, when a client starts a blocking query and an ACL token expires within
that time, Consul will return ACL not found error with a 403 status code. However,
sometimes if an ACL token is invalidated at the same time as the query's deadline is reached,
Consul will instead return an empty response with a 200 status code.
This is because of the events being executed.
1. Client issues a blocking query request with timeout `t`.
2. ACL is deleted.
3. Server detects a change in ACLs and force closes the gRPC stream.
4. Client resubscribes with the same token and resets its state (view).
5. Client sees "ACL not found" error.
If ACL is deleted before step 4, the client is unaware that the stream was closed due to
an ACL error and will return an empty view (from the reset state) with the 200 status code.
To fix this problem, we introduce another state to the subsciption to indicate when a change
to ACLs has occured. If the server sees that there was an error due to ACL change, it will
re-authenticate the request and return an error if the token is no longer valid.
Fixes#20790
* Fix streaming RPCs for agentless.
This PR fixes an issue where cross-dc RPCs were unable to utilize
the streaming backend due to having the node name set. The result
of this was the agent-cache being utilized, which would cause high
cpu utilization and memory consumption due to the fact that it
keeps queries alive for 72 hours before purging inactive entries.
This resource consumption is compounded by the fact that each pod
in consul-k8s gets a unique token. Since the agent-cache uses the
token as a component of the key, the same query is duplicated for
each pod that is deployed.
* Add changelog.
* Fix xDS deadlock due to syncLoop termination.
This fixes an issue where agentless xDS streams can deadlock permanently until
a server is restarted. When this issue occurs, no new proxies are able to
successfully connect to the server.
Effectively, the trigger for this deadlock stems from the following return
statement:
https://github.com/hashicorp/consul/blob/v1.18.0/agent/proxycfg-sources/catalog/config_source.go#L199-L202
When this happens, the entire `syncLoop()` terminates and stops consuming from
the following channel:
https://github.com/hashicorp/consul/blob/v1.18.0/agent/proxycfg-sources/catalog/config_source.go#L182-L192
Which results in the `ConfigSource.cleanup()` function never receiving a
response and holding a mutex indefinitely:
https://github.com/hashicorp/consul/blob/v1.18.0/agent/proxycfg-sources/catalog/config_source.go#L241-L247
Because this mutex is shared, it effectively deadlocks the server's ability to
process new xDS streams.
----
The fix to this issue involves removing the `chan chan struct{}` used like an
RPC-over-channels pattern and replacing it with two distinct channels:
+ `stopSyncLoopCh` - indicates that the `syncLoop()` should terminate soon. +
`syncLoopDoneCh` - indicates that the `syncLoop()` has terminated.
Splitting these two concepts out and deferring a `close(syncLoopDoneCh)` in the
`syncLoop()` function ensures that the deadlock above should no longer occur.
We also now evict xDS connections of all proxies for the corresponding
`syncLoop()` whenever it encounters an irrecoverable error. This is done by
hoisting the new `syncLoopDoneCh` upwards so that it's visible to the xDS delta
processing. Prior to this fix, the behavior was to simply orphan them so they
would never receive catalog-registration or service-defaults updates.
* Add changelog.
* Shuffle the list of servers returned by `pbserverdiscovery.WatchServers`.
This randomizes the list of servers to help reduce the chance of clients
all connecting to the same server simultaneously. Consul-dataplane is one
such client that does not randomize its own list of servers.
* Fix potential goroutine leak in xDS recv loop.
This commit ensures that the goroutine which receives xDS messages from
proxies will not block forever if the stream's context is cancelled but
the `processDelta()` function never consumes the message (due to being
terminated).
* Add changelog.
* disable terminating gateway auto host rewrite
* add changelog
* clean up unneeded additional snapshot fields
* add new field to docs
* squash
* fix test
* Update agent/dns.go
Co-authored-by: Michael Zalimeni <michael.zalimeni@hashicorp.com>
* PR feedback
* split tests out into multiple files.
* Extract responsibilities from router into discoveryResultsFetcher, messageSerializer, responseGenerator.
* adding recordmaker tests
* add response generator test coverage.
* changing tests case name based on PR feedback
---------
Co-authored-by: Michael Zalimeni <michael.zalimeni@hashicorp.com>
* NET-7813 - DNS : SERVFAIL when resolving PTR records
* Update agent/dns.go
Co-authored-by: Michael Zalimeni <michael.zalimeni@hashicorp.com>
* PR feedback
---------
Co-authored-by: Michael Zalimeni <michael.zalimeni@hashicorp.com>
Ensure all topics are refreshed on FSM restore and add supervisor loop to v1 controller subscriptions
This PR fixes two issues:
1. Not all streams were force closed whenever a snapshot restore happened. This means that anything consuming data from the stream (controllers, queries, etc) were unaware that the data they have is potentially stale / invalid. This first part ensures that all topics are purged.
2. The v1 controllers did not properly handle stream errors (which are likely to appear much more often due to 1 above) and so it introduces a supervisor thread to restart the watches when these errors occur.
John Maguire
John Murret
Dan Stough
Kiran Naidoo
sarahalsmiller
Deniz Onur Duzgun
Nathan Coleman
Dhia Ayachi
R.B. Boyer
Michael Zalimeni
Eric Haberkorn
George Ma
Iryna Shustava
Chris S. Kim
Derek Menteer
Matt Keeler
Semir Patel