Ensure that enabling AutoConfig sets the tls configurator properly
This also refactors the TLS configurator a bit so the naming doesn’t imply only AutoEncrypt as the source of the automatically setup TLS cert info.
Most of the groundwork was laid in previous PRs between adding the cert-monitor package to extracting the logic of signing certificates out of the connect_ca_endpoint.go code and into a method on the server.
This also refactors the auto-config package a bit to split things out into multiple files.
This implements a solution for #7863
It does:
Add a new config cache.entry_fetch_rate to limit the number of calls/s for a given cache entry, default value = rate.Inf
Add cache.entry_fetch_max_burst size of rate limit (default value = 2)
The new configuration now supports the following syntax for instance to allow 1 query every 3s:
command line HCL: -hcl 'cache = { entry_fetch_rate = 0.333}'
in JSON
{
"cache": {
"entry_fetch_rate": 0.333
}
}
On the servers they must have a certificate.
On the clients they just have to set verify_outgoing to true to attempt TLS connections for RPCs.
Eventually we may relax these restrictions but right now all of the settings we push down (acl tokens, acl related settings, certificates, gossip key) are sensitive and shouldn’t be transmitted over an unencrypted connection. Our guides and docs should recoommend verify_server_hostname on the clients as well.
Another reason to do this is weird things happen when making an insecure RPC when TLS is not enabled. Basically it tries TLS anyways. We should probably fix that to make it clearer what is going on.
The envisioned changes would allow extra settings to enable dynamically defined auth methods to be used instead of or in addition to the statically defined one in the configuration.
There are a couple of things in here.
First, just like auto encrypt, any Cluster.AutoConfig RPC will implicitly use the less secure RPC mechanism.
This drastically modifies how the Consul Agent starts up and moves most of the responsibilities (other than signal handling) from the cli command and into the Agent.
All commands which read config (agent, services, and validate) will now
print warnings when one of the config files is skipped because it did
not match an expected format.
Also ensures that config validate prints all warnings.
Previously the logic for reading ConfigFiles and produces Sources was split
between NewBuilder and Build. This commit moves all of the logic into NewBuilder
so that Build() can operate entirely on Sources.
This change is in preparation for logging warnings when files have an
unsupported extension.
It also reduces the scope of BuilderOpts, and gets us very close to removing
Builder.options.
The nil value was never used. We can avoid a bunch of complications by
making the field a string value instead of a pointer.
This change is in preparation for fixing a silent config failure.
Flags is an overloaded term in this context. It generally is used to
refer to command line flags. This struct, however, is a data object
used as input to the construction.
It happens to be partially populated by command line flags, but
otherwise has very little to do with them.
Renaming this struct should make the actual responsibility of this struct
more obvious, and remove the possibility that it is confused with
command line flags.
This change is in preparation for adding additional fields to
BuilderOpts.
This field was populated for one reason, to test that it was empty.
Of all the callers, only a single one used this functionality. The rest
constructed a `Flags{}` struct which did not set Args.
I think this shows that the logic was in the wrong place. Only the agent
command needs to care about validating the args.
This commit removes the field, and moves the logic to the one caller
that cares.
Also fix some comments.
Currently opaque config blocks (config entries, and CA provider config) are
modified by PatchSliceOfMaps, making it impossible for these opaque
config sections to contain slices of maps.
In order to fix this problem, any lazy-decoding of these blocks needs to support
weak decoding of []map[string]interface{} to a struct type before
PatchSliceOfMaps is replaces. This is necessary because these config
blobs are persisted, and during an upgrade an older version of Consul
could read one of the new configuration values, which would cause an error.
To support the upgrade path, this commit first introduces the new hooks
for weak decoding of []map[string]interface{} and uses them only in the
lazy-decode paths. That way, in a future release, new style
configuration will be supported by the older version of Consul.
This decode hook has a number of advantages:
1. It no longer panics. It allows mapstructure to report the error
2. It no longer requires the user to declare which fields are slices of
structs. It can deduce that information from the 'to' value.
3. It will make it possible to preserve opaque configuration, allowing
for structured opaque config.
This allows the operator to disable agent caching for the http endpoint.
It is on by default for backwards compatibility and if disabled will
ignore the url parameter `cached`.
* testing: replace most goe/verify.Values with require.Equal
One difference between these two comparisons is that go/verify considers
nil slices/maps to be equal to empty slices/maps, where as testify/require
does not, and does not appear to provide any way to enable that behaviour.
Because of this difference some expected values were changed from empty
slices to nil slices, and some calls to verify.Values were left.
* Remove github.com/pascaldekloe/goe/verify
Reduce the number of assertion packages we use from 2 to 1
Three of the checks are temporarily disabled to limit the size of the
diff, and allow us to enable all the other checks in CI.
In a follow up we can fix the issues reported by the other checks one
at a time, and enable them.
Based on work done in https://github.com/hashicorp/memberlist/pull/196
this allows to restrict the IP ranges that can join a given Serf cluster
and be a member of the cluster.
Restrictions on IPs can be done separatly using 2 new differents flags
and config options to restrict IPs for LAN and WAN Serf.
This will emit warnings about the configs not doing anything but still allow them to be parsed.
This also added the warnings for enterprise fields that we already had in OSS but didn’t change their enforcement behavior. For example, attempting to use a network segment will cause a hard error in OSS.
* Implements a simple, tcp ingress gateway workflow
This adds a new type of gateway for allowing Ingress traffic into Connect from external services.
Co-authored-by: Chris Piraino <cpiraino@hashicorp.com>
On recent Mac OS versions, the ulimit defaults to 256 by default, but many
systems (eg: some Linux distributions) often limit this value to 1024.
On validation of configuration, Consul now validates that the number of
allowed files descriptors is bigger than http_max_conns_per_client.
This make some unit tests failing on Mac OS.
Use a less important value in unit test, so tests runs well by default
on Mac OS without need for tuning the OS.
I spent some time today on my local Mac to figure out why Consul 1.6.3+
was not accepting limits.http_max_conns_per_client.
This adds an explicit check on number of file descriptors to be sure
it might work (this is no guarantee as if many clients are reaching
the agent, it might consume even more file descriptors)
Anyway, many users are fighting with RLIMIT_NOFILE, having a clear
message would allow them to figure out what to fix.
Example of message (reload or start):
```
2020-03-11T16:38:37.062+0100 [ERROR] agent: Error starting agent: error="system allows a max of 512 file descriptors, but limits.http_max_conns_per_client: 8192 needs at least 8212"
```
When run in with `-dev` in DevMode, it is not possible to replace
the embeded UI with another one because `-dev` implies `-ui`.
This commit allows this an slightly change the error message
about Consul 0.7.0 which is very old and does not apply to
current version anyway.
This is like a Möbius strip of code due to the fact that low-level components (serf/memberlist) are connected to high-level components (the catalog and mesh-gateways) in a twisty maze of references which make it hard to dive into. With that in mind here's a high level summary of what you'll find in the patch:
There are several distinct chunks of code that are affected:
* new flags and config options for the server
* retry join WAN is slightly different
* retry join code is shared to discover primary mesh gateways from secondary datacenters
* because retry join logic runs in the *agent* and the results of that
operation for primary mesh gateways are needed in the *server* there are
some methods like `RefreshPrimaryGatewayFallbackAddresses` that must occur
at multiple layers of abstraction just to pass the data down to the right
layer.
* new cache type `FederationStateListMeshGatewaysName` for use in `proxycfg/xds` layers
* the function signature for RPC dialing picked up a new required field (the
node name of the destination)
* several new RPCs for manipulating a FederationState object:
`FederationState:{Apply,Get,List,ListMeshGateways}`
* 3 read-only internal APIs for debugging use to invoke those RPCs from curl
* raft and fsm changes to persist these FederationStates
* replication for FederationStates as they are canonically stored in the
Primary and replicated to the Secondaries.
* a special derivative of anti-entropy that runs in secondaries to snapshot
their local mesh gateway `CheckServiceNodes` and sync them into their upstream
FederationState in the primary (this works in conjunction with the
replication to distribute addresses for all mesh gateways in all DCs to all
other DCs)
* a "gateway locator" convenience object to make use of this data to choose
the addresses of gateways to use for any given RPC or gossip operation to a
remote DC. This gets data from the "retry join" logic in the agent and also
directly calls into the FSM.
* RPC (`:8300`) on the server sniffs the first byte of a new connection to
determine if it's actually doing native TLS. If so it checks the ALPN header
for protocol determination (just like how the existing system uses the
type-byte marker).
* 2 new kinds of protocols are exclusively decoded via this native TLS
mechanism: one for ferrying "packet" operations (udp-like) from the gossip
layer and one for "stream" operations (tcp-like). The packet operations
re-use sockets (using length-prefixing) to cut down on TLS re-negotiation
overhead.
* the server instances specially wrap the `memberlist.NetTransport` when running
with gateway federation enabled (in a `wanfed.Transport`). The general gist is
that if it tries to dial a node in the SAME datacenter (deduced by looking
at the suffix of the node name) there is no change. If dialing a DIFFERENT
datacenter it is wrapped up in a TLS+ALPN blob and sent through some mesh
gateways to eventually end up in a server's :8300 port.
* a new flag when launching a mesh gateway via `consul connect envoy` to
indicate that the servers are to be exposed. This sets a special service
meta when registering the gateway into the catalog.
* `proxycfg/xds` notice this metadata blob to activate additional watches for
the FederationState objects as well as the location of all of the consul
servers in that datacenter.
* `xds:` if the extra metadata is in place additional clusters are defined in a
DC to bulk sink all traffic to another DC's gateways. For the current
datacenter we listen on a wildcard name (`server.<dc>.consul`) that load
balances all servers as well as one mini-cluster per node
(`<node>.server.<dc>.consul`)
* the `consul tls cert create` command got a new flag (`-node`) to help create
an additional SAN in certs that can be used with this flavor of federation.
Fixes#7231. Before an agent would always emit a warning when there is
an encrypt key in the configuration and an existing keyring stored,
which is happening on restart.
Now it only emits that warning when the encrypt key from the
configuration is not part of the keyring.
Matt Keeler