* Migrated all of the old leader task tests and got them passing
* Refactor and consolidate task killing code in AR to always kill leader
tasks first
* Fixed lots of issues with state restoring
* Fixed deadlock in AR.Destroy if AR.Run had never been called
* Added a new in memory statedb for testing
Saves a tiny bit of cpu and some IO. Sadly doesn't prevent all IO on
duplicate writes as the transactions are still created and committed.
$ go test -bench=. -benchmem
goos: linux
goarch: amd64
pkg: github.com/hashicorp/nomad/helper/boltdd
BenchmarkWriteDeduplication_On-4 500 4059591 ns/op 23736 B/op 56 allocs/op
BenchmarkWriteDeduplication_Off-4 300 4115319 ns/op 25942 B/op 55 allocs/op
Previously we did a validation pass over CA PEM files before calling
Go's CertPool.AppendCertsFromPEM to provide more detailed error messages
than the stdlib provides.
Unfortunately our validation was overly strict and rejected valid CA
files. This is actually the reason the stdlib PEM parser doesn't return
meaningful errors: PEM files are extremely permissive and it's difficult
to tell the difference between invalid data and valid metadata.
This PR removes our custom validation as it would reject valid data and
the extra error messages were not useful in diagnosing the error
encountered.
Not setting the host name led the Go HTTP client to expect a certificate
with a DNS-resolvable name. Since Nomad uses `${role}.${region}.nomad`
names ephemeral dir migrations were broken when TLS was enabled.
Added an e2e test to ensure this doesn't break again as it's very
difficult to test and the TLS configuration is very easy to get wrong.
The default timeout is too short for some overburdened or resource
constrained machines to complete the WMI query before the context
deadline expires. This causes them to be unable to fingerprint the CPU
properly.
The new version of gopsutil introduces a 3 second timeout that could come up as an error here; however, we are outputting the wrong variable and eating the error.
Mahmood Ali