The E2E provisioning used local-exec to call ssh in a for loop in a hacky
workaround https://github.com/hashicorp/terraform/issues/25634, which
prevented remote-exec from working on Windows. Move to a newer version of
Terraform that fixes the remote-exec bug to make provisioning more reliable
and observable.
Note that Windows remote-exec needs to include the `powershell` call itself,
unlike Unix-alike remote-exec.
The AWS EBS CSI plugin was observed to return a EOF when we get to the end of
the paging for `ListSnapshots`, counter to specification. Handle this case
gracefully, including for `ListVolumes` (which EBS doesn't support but has
similar semantics).
Also fixes a timestamp formatting bug on `ListSnapshots`
Split the EBS and EFS tests out into their own test cases:
* EBS exercises the Controller RPCs, including the create/snapshot workflow.
* EFS exercises only the Node RPCs, and assumes we have an existing volume
that gets registered, rather than created.
The goal is to always find an interface with an address, preferring
sandbox interfaces, but falling back to the first address found.
A test was added against a known CNI plugin output that was not handled
correctly before.
With the updated undefined variable code, we attempt to pick the text of
`${....}` verbatim from the hcl body. Previously, we'd attempt to
regenerate the string from the AST and pray it matches input; the
generation is lossy, as the same AST can represent multiple variations
(e.g. `${v.0}` and `${v[0]}` have the same HCLv2 AST). In this change,
we attempt to go back to the hcl2 source and find the string snippet
corresponding to the variable reference.
Our API client `delete` method doesn't include a request body, but accepts an
interface for the response. We were accidentally putting the request body into
the response, which doesn't get picked up in unit tests because we're not
reading the (always empty) response body anyways.
The plugin stub object does not include fine-grained capability checks, which
means `nomad volume status -verbose` will return ugly and verbose error
"Unimplemented" messages from the plugin if it does not support the CSI
`ListVolumes` RPC. Return a nicer error message from our RPC handler instead.
The HTTP router did not correctly route `/v1/volumes/external` without being
explicitly added to the top-level router. Break this out into its own request
handler.
If a plugin returns an error, we should continue at the outer scope to query
the next plugin, otherwise we just retry the plugin we got an error
on (potentially infinitely if it's an invalid request like an unsupported
plugin).
Add new `access_mode`/`attachment_mode` fields. Make it more clear which set
of fields belong to create vs register. Update the example spec that's
generated by `volume init`.
Registration of Nomad volumes previously allowed for a single volume
capability (access mode + attachment mode pair). The recent `volume create`
command requires that we pass a list of requested capabilities, but the
existing workflow for claiming volumes and attaching them on the client
assumed that the volume's single capability was correct and unchanging.
Add `AccessMode` and `AttachmentMode` to `CSIVolumeClaim`, use these fields to
set the initial claim value, and add backwards compatibility logic to handle
the existing volumes that already have claims without these fields.
Previously installing buf was left out of `make bootstrap` because it
had conflicts with the `tools/go.mod` file and dependencies used by
other tools. With Go 1.16 we eliminated that `go.mod` file, and can
now just install `buf` with `go install` like everything else.
This PR adds the common OSS changes for adding support for Consul Namespaces,
which is going to be a Nomad Enterprise feature. There is no new functionality
provided by this changeset and hopefully no new bugs.
