In #3863, it was discovered that volumes from projected sources were
being backed up by restic when they should have been skipped. Restoring
these volumes triggers a known bug in restic.
In #3866, we started skipping volumes from a projected source, however
there will exist backups that were taken before this fix was introduced.
This change modifies the restore logic to skip the restore of any volume
that came from a projected source, allowing backups taken before #3866
to be restored successfully.
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
* Remove controllers and sleeps in API groups e2e tests
Signed-off-by: F. Gold <fgold@vmware.com>
* Print command in AfterEach(...) and check error
Signed-off-by: F. Gold <fgold@vmware.com>
* Make change ahead of PR3764 changes in main
Signed-off-by: F. Gold <fgold@vmware.com>
* Update go.{mod,sum} files
Signed-off-by: F. Gold <fgold@vmware.com>
* Run make update
Signed-off-by: F. Gold <fgold@vmware.com>
For internal builds of Velero, we need to be able to specify an
alternative Dockerfile which uses an alternative image registry to pull
the base images from. This change adapts our Makefile such that both the
main Dockerfile and build image Dockerfile can be overridden.
We have some special handling for the build image to only build when the
Dockerfile has changed. In this case, we check whether a custom
Dockerfile has been provided, and always rebuild in that case. For
custom build image Dockerfiles, use a fixed tag rather than the one
based on commit SHA of the original file.
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
Snapshot tests can be run with Ginkgo focus "Snapshot" and restic tests with Ginkgo focus "Restic".
Restic and volume snapshot tests can now be run simultaneously.
Added check for kibishii app start after restore.
Consolidated kibishii pod checks into waitForKibishiiPods.
Added WaitForPods function to e2e/tests/common.goSnapshot tests are skipped automatically on kind clusters.
Fixed issue where velero_utils InstallVeleroServer was looking for the Restic daemon set in the "velero" namespace only (was ignoring io.Namespace)
Signed-off-by: Dave Smith-Uchida <dsmithuchida@vmware.com>
* Improve readbility and formatting of pkg/restore/restore.go
Signed-off-by: F. Gold <fgold@vmware.com>
* Update paths to include API group versions
Signed-off-by: F. Gold <fgold@vmware.com>
* Use full word, 'resource' instead of 'resrc'
Signed-off-by: F. Gold <fgold@vmware.com>
The test for multiple credentials assumed that the plugin for the
additional BSL provider was already installed. This will not be the case
when performing a clean install of Velero between tests.
This adds a new utility function to add the plugins that are necessary
for the additional BSL provider. It doesn't check which plugins are
already installed, it will just attempt to install and if the stderr
contains the message that it is a duplicate plugin, we ignore the error
and continue. This could be improved by instpecting the output from
`velero plugin get` but I opted for a quicker solution given the
upcoming release.
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
* Use Credential from BSL for restic commands
This change introduces support for restic to make use of per-BSL
credentials. It makes use of the `credentials.FileStore` introduced in
PR #3442 to write the BSL credentials to disk. To support per-BSL
credentials for restic, the environment for the restic commands needs to
be modified for each provider to ensure that the credentials are
provided via the correct provider specific environment variables.
This change introduces a new function `restic.CmdEnv` to check the BSL
provider and create the correct mapping of environment variables for
each provider.
Previously, AWS and GCP could rely on the environment variables in the
Velero deployments to obtain the credentials file, but now these
environment variables need to be set with the path to the serialized
credentials file if a credential is set on the BSL.
For Azure, the credentials file in the environment was loaded and parsed
to set the environment variables for restic. Now, we check if the BSL
has a credential, and if it does, load and parse that file instead.
This change also introduces a few other small improvements. Now that we
are fetching the BSL to check for the `Credential` field, we can use the
BSL directly to get the `CACert` which means that we can remove the
`GetCACert` function. Also, now that we have a way to serialize secrets
to disk, we can use the `credentials.FileStore` to get a temp file for
the restic repo password and remove the `restic.TempCredentialsFile`
function.
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
* Add documentation for per-BSL credentials
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
* Address review feedback
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
* Address review comments
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
We are no longer adding the Credentials field to the VSL so this reverts
part the change that added it (#3409).
The original PR also added the `snapshot-location set` command. This
command only included options for setting the credential but is part of
the work for #2426. Due to this, the command has been left in place
(with the credentials option removed) but has been hidden.
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
This change adds an E2E test which exercises the mulitple credentials
feature added in #3489. The test creates a secret from the given
credentials and creates a BackupStorageLocation which uses those
credentials. A backup and restore is then performed to the default
BSL and to the newly created BSL.
This change adds new flags to the E2E test suite to configure the BSL
created and used in the test.
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
* Load credentials and pass to ObjectStorage plugins
Update NewObjectBackupStore to take a CredentialsGetter which can be
used to get the credentials for a BackupStorageLocation if it has been
configured with a Credential. If the BSL has a credential, use that
SecretKeySelector to fetch the secret, write the contents to a temp file
and then pass that file through to the plugin via the config map using
the key `credentialsFile`. This relies on the plugin being able to use
this new config field.
This does not yet handle VolumeSnapshotLocations or ResticRepositories.
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
* Address code reviews
Add godocs and comments.
Improve formatting and test names.
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
* Address code reviews
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
* Add uninstall cmd
- init fn to uninstall velero
- abstract dynamic client creation to a separate fn
- creates a separate client per unstructured resource
- add delete client for CRDs
- export appendUnstructured
- add uninstall command to main cmd
- export `podTemplateOption`
- uninstall resources in the reverse order of installation
- fallback to `velero` if no ns is provided during uninstall
- skip deletion if the resource doesn't exist
- handle resource not found error
- match log formatting with cli install logs
- add Delete fn to fake client
- fix import order
- add changelog
- add comment doc for CreateClient fn
Signed-off-by: Suraj Banakar <suraj@infracloud.io>
* Re-use uninstall code from test suite
- move helper functions out of test suite
- this is to prevent cyclic imports
- move uninstall helpers to uninstall cmd
- call them from test suite
- revert export of variables/fns from install code
- because not required anymore
Signed-off-by: Suraj Banakar <suraj@infracloud.io>
* Revert `PodTemplateOption` -> `podTemplateOption`
Signed-off-by: Suraj Banakar <suraj@infracloud.io>
* Use uninstall helper under VeleroUninstall
- as a wrapper
- fix import related errors in test suite
Signed-off-by: Suraj Banakar <suraj@infracloud.io>
* Validate CRDs against latest Kubernetes versions
Add Kubernetes v1.19 and v1.20 series images, and consolidate the job
into a single file to reduce repetition.
Signed-off-by: Nolan Brubaker <brubakern@vmware.com>
* Ignore job if the changes are only site/design
Signed-off-by: Nolan Brubaker <brubakern@vmware.com>
* Fix codespell error
Signed-off-by: Nolan Brubaker <brubakern@vmware.com>
* Cache Velero binary for reuse on workers
This will cache the Velero binary based on the PR number and a SHA256 of
the generated binary.
This way, the runners testing each version of Kubernetes do not need to
build it independently.
Signed-off-by: Nolan Brubaker <brubakern@vmware.com>
* Fix GitHub event access
Signed-off-by: Nolan Brubaker <brubakern@vmware.com>
* Wrap output path in quotes
Signed-off-by: Nolan Brubaker <brubakern@vmware.com>
* Move code checkout to build step
Signed-off-by: Nolan Brubaker <brubakern@vmware.com>
* Also cache go modules
Signed-off-by: Nolan Brubaker <brubakern@vmware.com>
* Fix syntax issues
Signed-off-by: Nolan Brubaker <brubakern@vmware.com>
* Download cached binary on each node
Signed-off-by: Nolan Brubaker <brubakern@vmware.com>
* Use cached go modules on main CI
Signed-off-by: Nolan Brubaker <brubakern@vmware.com>
HTTPS requests were failing due to the ca-certificates package not being
installed in the Tilt image.
This change takes the command to install this package from our main
Dockerfile (which also includes installing tzdata).
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
* Use pod namespace from backup when matching PVBs
In #3051, we introduced an additional check to ensure that a PVB matched
a particular pod by checking both the name and the namespace of the pod.
This caused an issue when using a namespace mapping on restore. In the
case where a namespace mapping is being used, the check for whether a
PVB matches a particular pod will fail as the PVB was created for the
original pod namespace and is not aware of the new namespace mapping
being used. This resulted in PVRs not being created for pods that were
being restored into new namespaces. The restic init containers were
being created to wait on the volume restore, however this would cause
the restored pods to block indefinitely as they would be waiting for a
volume restore that was not scheduled.
To fix this, we use the original namespace of the pod from the backup to
match the PVB to the pod being restored, not the new namespace where
the pod is being restored into.
Fixes#3467.
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
* Explain why the namespace mapping can't be used
Signed-off-by: Bridget McErlean <bmcerlean@vmware.com>
Bridget McErlean
codegold79andBridget McErlean
Scott SeagoandBridget McErlean
Carlisia ThompsonandBridget McErlean
Ashish AmarnathandBridget McErlean
David L. Smith-UchidaandGitHub
Nolan BrubakerandGitHub
Pranav GaikwadandGitHub
Suraj BanakarandGitHub