I hypothesize that the flakiness in rolling update is due to shutting
down s3 server before s4 is properly added as a voter.
The chain of the flakiness is as follows:
1. Bootstrap with s1, s2, s3
2. Add s4
3. Wait for servers to register with 3 voting peers
* But we already have 3 voters (s1, s2, and s3)
* s4 is added as a non-voter in Raft v3 and must wait until autopilot promots it
4. Test proceeds without s4 being a voter
5. s3 shutdown
6. cluster changes stall due to leader election and too many pending configuration
changes (e.g. removing s3 from raft, promoting s4).
Here, I have the test wait until s4 is marked as a voter before shutting
down s3, so we don't have too many configuration changes at once.
In https://circleci.com/gh/hashicorp/nomad/57092, I noticed the
following events:
```
TestAutopilot_RollingUpdate: autopilot_test.go:204: adding server s4
TestAutopilot_RollingUpdate: testlog.go:34: 2020-04-03T20:08:19.789Z [INFO] nomad/serf.go:60: nomad: adding server: server="nomad-137.global (Addr: 127.0.0.1:9177) (DC: dc1)"
TestAutopilot_RollingUpdate: testlog.go:34: 2020-04-03T20:08:19.789Z [INFO] raft/raft.go:1018: nomad.raft: updating configuration: command=AddNonvoter server-id=c54b5bf4-1159-34f6-032d-56aefeb08425 server-addr=127.0.0.1:9177 servers="[{Suffrage:Voter ID:df01ba65-d1b2-17a9-f792-a4459b3a7c09 Address:127.0.0.1:9171} {Suffrage:Voter ID:c3337778-811e-2675-87f5-006309888387 Address:127.0.0.1:9173} {Suffrage:Voter ID:186d5e15-c473-e2b3-b5a4-3259a84e10ef Address:127.0.0.1:9169} {Suffrage:Nonvoter ID:c54b5bf4-1159-34f6-032d-56aefeb08425 Address:127.0.0.1:9177}]"
TestAutopilot_RollingUpdate: autopilot_test.go:218: shutting down server s3
TestAutopilot_RollingUpdate: testlog.go:34: 2020-04-03T20:08:19.797Z [INFO] raft/replication.go:456: nomad.raft: aborting pipeline replication: peer="{Nonvoter c54b5bf4-1159-34f6-032d-56aefeb08425 127.0.0.1:9177}"
TestAutopilot_RollingUpdate: autopilot_test.go:235: waiting for s4 to stabalize and be promoted
TestAutopilot_RollingUpdate: testlog.go:34: 2020-04-03T20:08:19.975Z [ERROR] raft/raft.go:1656: nomad.raft: failed to make requestVote RPC: target="{Voter c3337778-811e-2675-87f5-006309888387 127.0.0.1:9173}" error="dial tcp 127.0.0.1:9173: connect: connection refused"
TestAutopilot_RollingUpdate: retry.go:121: autopilot_test.go:241: don't want "c3337778-811e-2675-87f5-006309888387"
autopilot_test.go:241: didn't find map[c54b5bf4-1159-34f6-032d-56aefeb08425:true] in []raft.ServerID{"df01ba65-d1b2-17a9-f792-a4459b3a7c09", "186d5e15-c473-e2b3-b5a4-3259a84e10ef"}
```
Note how s3, c3337778, is present in the peers list in the final
failure, but s4, c54b5bf4, is added as a Nonvoter and isn't present in
the final peers list.
This change updates tests to honor `BootstrapExpect` exclusively when
forming test clusters and removes test only knobs, e.g.
`config.DevDisableBootstrap`.
Background:
Test cluster creation is fragile. Test servers don't follow the
BootstapExpected route like production clusters. Instead they start as
single node clusters and then get rejoin and may risk causing brain
split or other test flakiness.
The test framework expose few knobs to control those (e.g.
`config.DevDisableBootstrap` and `config.Bootstrap`) that control
whether a server should bootstrap the cluster. These flags are
confusing and it's unclear when to use: their usage in multi-node
cluster isn't properly documented. Furthermore, they have some bad
side-effects as they don't control Raft library: If
`config.DevDisableBootstrap` is true, the test server may not
immediately attempt to bootstrap a cluster, but after an election
timeout (~50ms), Raft may force a leadership election and win it (with
only one vote) and cause a split brain.
The knobs are also confusing as Bootstrap is an overloaded term. In
BootstrapExpect, we refer to bootstrapping the cluster only after N
servers are connected. But in tests and the knobs above, it refers to
whether the server is a single node cluster and shouldn't wait for any
other server.
Changes:
This commit makes two changes:
First, it relies on `BootstrapExpected` instead of `Bootstrap` and/or
`DevMode` flags. This change is relatively trivial.
Introduce a `Bootstrapped` flag to track if the cluster is bootstrapped.
This allows us to keep `BootstrapExpected` immutable. Previously, the
flag was a config value but it gets set to 0 after cluster bootstrap
completes.
Copy the updated version of freeport (sdk/freeport), and tweak it for use
in Nomad tests. This means staying below port 10000 to avoid conflicts with
the lib/freeport that is still transitively used by the old version of
consul that we vendor. Also provide implementations to find ephemeral ports
of macOS and Windows environments.
Ports acquired through freeport are supposed to be returned to freeport,
which this change now also introduces. Many tests are modified to include
calls to a cleanup function for Server objects.
This should help quite a bit with some flakey tests, but not all of them.
Our port problems will not go away completely until we upgrade our vendor
version of consul. With Go modules, we'll probably do a 'replace' to swap
out other copies of freeport with the one now in 'nomad/helper/freeport'.
Mahmood Ali