* fix: Improve `coder server` shutdown procedure
This commit improves the `coder server` shutdown procedure so that all
triggers for shutdown do so in a graceful way without skipping any
steps.
We also improve cancellation and shutdown of services by ensuring
resources are cleaned up at the end.
Notable changes:
- We wrap `cmd.Context()` to allow us to control cancellation better
- We attempt graceful shutdown of the http server (`server.Shutdown`)
because it's less abrupt (compared to `shutdownConns`)
- All exit paths share the same shutdown procedure (except for early
exit)
- `provisionerd`s are now shutdown concurrently instead of one at a
time, the also now get a new context for shutdown because
`cmd.Context()` may be cancelled
- Resources created by `newProvisionerDaemon` are cleaned up
- Lifecycle `Executor` exits its goroutine on context cancellation
Fixes#3245
The device keeps logging to the logger even after `dev.Close()` but
doing that with `t.Log` is unsafe (test has ended). This is why
`slogtest` was used.
`dev.Close()` has a wait on encryption and decryption routines, however,
these are left running even after the wait. The implementation uses the
WaitGroups in a weird way.
* fix: Change uses of t.Cleanup -> defer in test bodies
Mixing t.Cleanup and defer can lead to unexpected order of execution.
* fix: Ensure t.Cleanup is not aborted by require
* chore: Add helper annotations
This was using the incorrect GitHub endpoint prior, which fetched a team
by slug. Any user in a GitHub organization can view all teams, so this
didn't block signups like intended.
I've verified this API returns an error when the calling user is not a
member of the team requested.
Fixes#3105.
* fix: Remove use of `require` in `require.Eventually` in tests
Because require uses `t.FailNow()` and `require.Eventually` runs the
function in a goroutine, which is not allowed.
* feat: Add ruleguard for require.Eventually
Co-authored-by: Cian Johnston <cian@coder.com>
* Dev tunnel test uses local fake server; fixed port
Signed-off-by: Spike Curtis <spike@coder.com>
* Remove parallel for test
Signed-off-by: Spike Curtis <spike@coder.com>
* Fix segfault
Some contexts were moved closer to use so that test setup doesn't affect
timeout. And timeout was increased for some others to avoid flakyness
due to slow test runners.
* set a failed canceled job status correctly
resolves#1374
* added unit test for convertProvisionerJob
* Update coderd/provisionerjobs_internal_test.go
Co-authored-by: Cian Johnston <cian@coder.com>
* PR feedback
Co-authored-by: Cian Johnston <cian@coder.com>
Two coderd unit tests (TestPatchCancelTemplateVersion/Success and TestPatchCancelWorkspaceBuild) implied erroneously that the job was canceled successfully.
This is not the case, as these unit tests do not include a Provision_Complete response in the input to the
echo provisioner. Now explicitly checking the job error and bumping the force cancel interval to be longer.
Fixes#3083.
* Pass workspace owner email address to provisioner
* Remove owner_email and owner_username fields from agent metadata
* Add Git environment variables to example templates
* Remove "owner_name" field from provisioner metadata, use username instead
* Remove Git configuration from most templates, add documentation
* Proofreading/typo fixes from @mafredri
* Update example templates to latest version of terraform-provider-coder
* return parameters from Terraform provisioner in sorted order
* persist parameter indices in database and return them in correct order from API
* don't re-sort parameters by name when creating templates
- make default template max TTL 24 hours (still less than 168)
- make default workspace autostop 2 hours unless specified otherwise
- add instance type selector to aws templates
* test: Use a template to prevent migrations from running for every test
* Create a single makefile target
* Fix built-in race
* Extend timeout of built-in PostgreSQL fetch
