* Remove unused backend wrapper from Cache
* Remove double printShutdownStatus
* Fix readyz race condition
* Test coverage for the readyz.monitor fix
* Close listeners immediately in proxy.shutdown
* Use and handle net.ErrClosed correctly
This adapts utils.IsUseOfClosedNetworkError to check for net.ErrClosed
even inside trace.Aggregate errors, makes it so that we always return
something that would pass errors.Is(err, net.ErrClosed) when returning
from a (net.Listener).Accept(), and handles closed listeners within our
various Serve() loops so that we don't hit spurious backoff waits while
shutting down.
* Close listeners early and emitters late
* Test coverage for the proxy listener changes
* Revert some errors back to trace.ConnectionProblem
* Reduce PR scope to just the proxy, add comments
* Improve error logging.
Prior to this patch, the `LoadBalancer` method `nextBackend()` returned
a pointer to an element of the `LoadBalancer`'s `backends` slice.
It is possible for the element referenced by that pointer to be overwritten
during backend removal (see loadbalancer.go:120). Even though the backend
removal itself is carried out under a lock, it is possible for code
dereferencing a pointer returned by `nextBackend() to get an unexpectedly
modified value.
This patch changes `nextBackend()` to return by value, severing the link
between the returned value and the backing store.
* Use cmp.Equal instead of manual Equals methods
Equals methods can get out-of-sync with the fields added in structs they
compare. Using `cmp.Equal` handles that, removes a ton of code and makes
it more explicit when specific fields are excluded from comparison.
* Use gogoproto equal plugin for comparing proto values
This will be faster than reflect-based go-cmp.
This code is not caught by linters because it's exported and they assume
there's some external users.
Since teleport is relatively self-contained, we can tell for sure
whether something is called or not.
This commit refactor discovery protocol
to make it less dependent on the database and
scale better on large numbers of tunnels.
Reverse tunnel is now always sending
back the list of all proxies registered in the
cluster in the form of discovery requests.
Before this commit, reverse tunnel server was comparing
existing TunnelConnection with the Proxies
and sending back the list of proxies that were not
discovered.
This required nodes to register tunnel connections
in the database and servers poll the connections.
On 10K clusters this is not scalable. Instead,
the change assumes that there is not a lot of
proxies so it's OK to send the information about
them back to all connected agents.
Agent pools can make up their own mind about what to
do with the information - they can ignore
the request as long as they observe all agents
connected to the requested proxies.
At the same time, to avoid using too much traffic,
reverse tunnel server only sends the discovery requests
after the first agent heartbeat and in case if
proxy list changes. To make it possible reverse tunnel
sets up a watch on the proxies.
David Boslee