This PR makes the following changes:
- enforces lower and upper limits on template `max_ttl_ms`
- adds a migration to enforce 7-day cap on `max_ttl`
- allows setting template `max_ttl` to 0
- updates template edit CLI help to be clearer
* fix: Add `SIGHUP` and `SIGTERM` handling to `coder server`
To prevent additional signals from aborting program execution, signal
handling was moved to the beginning of the main function, this ensures
that signals stays registered for the entire shutdown procedure.
Fixes#1529
* fix: Reduce variables needed for Docker template
This should make initial setup a bit simpler!
* Fix for M2 Macbooks
PostgreSQL 13 doesn't support the M series architecture.
* Fix name <-> id swap
* Update Docker provider to remove host requirement
Co-authored-by: Kyle Carberry <kyle@air.local>
* chore: Add nix shell for simple development setup
This enables contributors using Nix to set up their environment with ease.
* improve nix style, flake output schema
* fix error message
* Update scripts/build_go_slim.sh
Co-authored-by: Mathias Fredriksson <mafredri@gmail.com>
* Update scripts/build_go_slim.sh
Co-authored-by: Mathias Fredriksson <mafredri@gmail.com>
* Add UTC default for timezone and remove unnecessary goreleaser dependency
* Skip TZ test if localtime does not exist
Co-authored-by: Charlie Moog <moogcharlie@gmail.com>
Co-authored-by: Mathias Fredriksson <mafredri@gmail.com>
This allows deployments using our Prometheus export t determine
the number of active users in the past hour.
The interval is an hour to align with API key last used refresh times.
SSH connections poll to check shutdown time, so this will be accurate
even on long-running connections without dashboard requests.
This seems to have caused flakes on Windows, the reason could be that
the input is lost due to writing to stdin before the shell is ready, or
simply that the command wasn't echoed (for the same reason).
We no longer need to consume the output since #2122 has been fixed, so
this might remove the flake in the latter case.
Ideally we would wait for the prompt to be present, but since we are
spawning the users shell, we have no control of what the prompt looks
like. In CI we can make assumption but even then it could change in the
future.
* fix: Improve shutdown procedure of ssh, portforward, wgtunnel cmds
We could turn it into a practice to wrap `cmd.Context()` so that we have
more fine-grained control of cancellation. Sometimes in tests we may be
running commands with a context that is never canceled.
Related to #3221
* fix: Set ssh session stderr to stderr
- Problems can arise spawning workspaces if a schemeless URL is passed
as the access URL.
If an access url is detected to not have an "http" or "https" scheme
then it is prepended with "https". If the hostname is detected
to be a loopback device then "http" is preferred.
Before, there was a `template edit` AND a `template update`. The
distinction between both commands was easy to forget. `push` more
clearly indicates that the template's source code is being updated.
It is also complimentary to existing `template pull`.
* 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
I also noticed we don't have `goleak` enabled for CLI tests, this commit
adds it, but commented out. The reason being that we're nowhere near
being able to enable it yet.
Co-authored-by: Cian Johnston <cian@coder.com>
* 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
Kyle Carberry