Skip to content

Tags: cloudfoundry/bosh-agent

Tags

v2.862.0

Toggle v2.862.0's commit message
Update vendored dependencies

v2.861.0

Toggle v2.861.0's commit message
Bump actions/checkout from 6 to 7

Bumps [actions/checkout](https://github.com/actions/checkout) from 6 to 7.
- [Release notes](https://github.com/actions/checkout/releases)
- [Changelog](https://github.com/actions/checkout/blob/main/CHANGELOG.md)
- [Commits](actions/checkout@v6...v7)

---
updated-dependencies:
- dependency-name: actions/checkout
  dependency-version: '7'
  dependency-type: direct:production
  update-type: version-update:semver-major
...

Signed-off-by: dependabot[bot] <support@github.com>

v2.860.0

Toggle v2.860.0's commit message

Verified

This commit was created on GitHub.com and signed with GitHub’s verified signature.
Set command to spawn with lower priority in GenericScript Run method (#…

…438)

v2.859.0

Toggle v2.859.0's commit message
move from docker-image to registry-image

Pushing images with the deprecated docker-image resource
is broken in recent concourse versions.

v2.858.0

Toggle v2.858.0's commit message
move from docker-image to registry-image

Pushing images with the deprecated docker-image resource
is broken in recent concourse versions.

v2.857.0

Toggle v2.857.0's commit message
move from docker-image to registry-image

Pushing images with the deprecated docker-image resource
is broken in recent concourse versions.

v2.856.0

Toggle v2.856.0's commit message
Update vendored dependencies

v2.855.0

Toggle v2.855.0's commit message

Verified

This commit was created on GitHub.com and signed with GitHub’s verified signature.
Implement Runtime NVMe Instance Storage Discovery Using AWS EBS Symli…

…nks (#396)

* Implement runtime NVMe instance storage discovery using EBS symlinks

* Fix nvme instance storage discovery (#400)

* Refactor instance storage discovery into configurable component

Implement auto-detection for instance storage disk type

* Fix windows tests

* Fix windows tests (but for real this time)

* Remove leftover path normalization (#401)

* Fix nvme instance storage discovery (#407)

* Make implementation iaas-agnostic

* Rename storage resolver files

* Fix tests

* Remove instance storage resolver

* Don't use the aws pattern as default

* Refactor NVMe instance storage discovery and remove unused symlink patterns

* Enhance NVMe instance storage discovery with managed volume pattern support

* Fix unit tests

* Don't run windows unit tests when not supported

* Simplify FakeDevicePathResolver by removing unused fields and methods

* Wait for udev to settle before resolving EBS symlinks

* Add debug logs

* Import udev and add comment about why it's needed

* Fix missing closing bracket

* Update infrastructure/devicepathresolver/symlink_device_resolver.go

Co-authored-by: Ivaylo Ivanov <ivaylogi98@gmail.com>

* Update infrastructure/devicepathresolver/symlink_device_resolver_test.go

Co-authored-by: Ivaylo Ivanov <ivaylogi98@gmail.com>

* Fix lint identation

* Fix: skip unresolvable symlinks instead of returning error

ResolveSymlinksToDevices now logs a warning and continues when a
symlink cannot be resolved (e.g. stale/broken symlinks in
/dev/disk/by-id/). This prevents unnecessary deploy failures while
the count validation in discoverNVMeInstanceStorage still catches
any real mismatches.

Co-authored-by: Ivaylo Ivanov <ivaylogi98@gmail.com>

* Use the already constructed udev instance

Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

* fix: use %+v instead of %s for DiskSettings in error message

DiskSettings does not implement fmt.Stringer, so using %s produced
malformed %!s(...) output. Switch to %+v for actionable error messages.

* fix: handle timedOut in discoverIdentityInstanceStorage

* resolver: fail hard on unresolvable symlinks in ResolveSymlinksToDevices

Silently skipping a symlink that cannot be resolved leaves the managed
device exclusion set incomplete. If an EBS volume's by-id symlink is
broken, FilterDevices would not exclude it and the device could be
misidentified as instance storage, potentially causing data loss.

Return a wrapped error instead of continuing, so callers can propagate
the failure rather than proceeding with a partial exclusion set.

* test: update linux_platform test for hard-fail on broken symlinks

The platform-level test still expected the old skip-and-continue behavior.
Updated to assert that a broken managed volume symlink returns an error.

* fix: prioritize timedOut over err; assert nil result on resolver error

- discoverIdentityInstanceStorage: check timedOut before err so the
  explicit timeout error is never shadowed when both are set
- symlink_device_resolver_test: assert result is nil on failure to
  verify the fail-closed contract

---------

Co-authored-by: Ivaylo Ivanov <ivaylogi98@gmail.com>
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

v2.854.0

Toggle v2.854.0's commit message
Remove CentOS platform support

CentOS stemcells have been removed upstream, so this cleans up
the remaining CentOS-specific code from the agent:

- Delete platform/net/centos_net_manager.go and its tests
- Remove NewCentOSCertManager from platform/cert/cert_manager.go
- Remove CentOS test context from cert_manager_test.go
- Remove centos platform entry from platform/provider.go
- Clean up CentOS mention in routes_searcher comment

Fixes #339

v2.853.0

Toggle v2.853.0's commit message
Update vendored dependencies