| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
...to work for specific edge cases with a simpler solution.
Re-reads hooks directories after any changes are detected by the watchers.
Added monitoring test for adding a different invalid hook to primary directory.
Some issues with prior code:
- ReadDir would stop when it encounters an invalid hook, rather than registering an error but continuing to read the valid hook.
- Wouldn’t account for Rename and Chmod events.
- After doing a mv of the hooks file instead of rm, it would still think the hooks file is in the directory, but it has been moved to another location.
- If a hook file was renamed, it would register the renamed file as a separate hook and not delete the original, so it would then execute the hook twice - once for the renamed file, and once for the original name which it did not delete.
Signed-off-by: samc24 <sam.chaturvedi24@gmail.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This isn't an issue with podman, which will only ever use one
directory. But CRI-O generally uses two directories, and we want to
make sure that changes to the fallback directory are not clobbering
hooks configured in the override directory. More background in [1].
I've split the handling into a single-directory block and a
multiple-directory block so we don't waste time polling the filesystem
for single-directory removals.
I'm using the single-directory block for the the zero-directory case
as well. Managers with zero directories should not be receiving
fsnotify events, so I don't think it really matters which block
handles them. If we want to handle this case robustly (because we're
concerned about something in the hook package adjusted the private
.directories property on the fly?), then we'll probably want to add an
explicit zero-directory block in future work.
[1]: https://github.com/kubernetes-incubator/cri-o/pull/1470
Signed-off-by: W. Trevor King <wking@tremily.us>
Closes: #757
Approved by: rhatdan
|
|
|
This shifts the matching logic out of libpod/container_internal and
into the hook package, where we can reuse it after vendoring into
CRI-O. It also adds unit tests with almost-complete coverage. Now
libpod is even more isolated from the hook internals, which makes it
fairly straightforward to bump the hook config file to 1.0.0. I've
dubbed the old format 0.1.0, although it doesn't specify an explicit
version. Motivation for some of my changes with 1.0.0:
* Add an explicit version field. This will make any future JSON
structure migrations more straightforward by avoiding the need for
version-guessing heuristics.
* Collect the matching properties in a new When sub-structure. This
makes the root Hook structure easier to understand, because you
don't have to read over all the matching properties when wrapping
your head around Hook.
* Replace the old 'hook' and 'arguments' with a direct embedding of
the runtime-spec's hook structure. This provides access to
additional upstream properties (args[0], env, and timeout) and
avoids the complication of a CRI-O-specific analog structure.
* Add a 'when.always' property. You can usually accomplish this
effect in another way (e.g. when.commands = [".*"]), but having a
boolean explicitly for this use-case makes for easier reading and
writing.
* Replace the previous annotations array with an annotations map. The
0.1.0 approach matched only the values regardless of key, and that
seems unreliable.
* Replace 'cmds' with 'when.commands', because while there are a few
ways to abbreviate "commands", there's only one way to write it out
in full ;). This gives folks one less thing to remember when
writing hook JSON.
* Replace the old "inject if any specified condition matches" with
"inject if all specified conditions match". This allows for more
precise targeting. Users that need more generous targeting can
recover the previous behavior by creating a separate 1.0.0 hook file
for each specified 0.1.0 condition.
I've added doc-compat support for the various pluralizations of the
0.1.0 properties. Previously, the docs and code were not in
agreement. More on this particular facet in [1].
I've updated the docs to point out that the annotations being matched
are the OCI config annotations. This differs from CRI-O, where the
annotations used are the Kubernetes-supplied annotations [2,3]. For
example, io.kubernetes.cri-o.Volumes [4] is part of CRI-O's runtime
config annotations [5], but not part of the Kubernetes-supplied
annotations CRI-O uses for matching hooks.
The Monitor method supports the CRI-O use-case [6]. podman doesn't
need it directly, but CRI-O will need it when we vendor this package
there.
I've used nvidia-container-runtime-hook for the annotation examples
because Dan mentioned the Nvidia folks as the motivation behind
annotation matching. The environment variables are documented in [7].
The 0.1.0 hook config, which does not allow for environment variables,
only works because runc currently leaks the host environment into the
hooks [8]. I haven't been able to find documentation for their usual
annotation trigger or hook-install path, so I'm just guessing there.
[1]: https://github.com/kubernetes-incubator/cri-o/pull/1235
[2]: https://github.com/kubernetes-incubator/cri-o/blob/v1.10.0/server/container_create.go#L760
[3]: https://github.com/kubernetes-incubator/cri-o/blob/v1.10.0/server/container_create.go#L772
[4]: https://github.com/kubernetes-incubator/cri-o/blob/v1.10.0/pkg/annotations/annotations.go#L97-L98
[5]: https://github.com/kubernetes-incubator/cri-o/blob/v1.10.0/server/container_create.go#L830-L834
[6]: https://github.com/kubernetes-incubator/cri-o/pull/1345/
[7]: https://github.com/NVIDIA/nvidia-container-runtime/tree/v1.3.0-1#environment-variables-oci-spec
[8]: https://github.com/opencontainers/runc/pull/1738
Signed-off-by: W. Trevor King <wking@tremily.us>
Closes: #686
Approved by: mheon
|
