This commit add external-snapshotter container and
required rbac to support nfs snaphots.
Example volumesnaphotclass, volumesnapshot,
pvc-restore and pod-restore yamls are also added.
Signed-off-by: Rakshith R <rar@redhat.com>
CephFS CSI driver does not need permissions on Node,ConfigMap objects.
This PR also adjust the update->Patch for a couple of RBAC.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
Resizing is handled by the csi-resizer container, which needs to run in
the provisioner Pod. In addition to the container, the StorageClass also
needs to allow volume expansion.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
This should address the following failure when Pod Security Policies are
enabled:
> FailedCreate: Error creating: pods "csi-nfs-node-" is forbidden:
> PodSecurityPolicy: unable to admit pod: spec.containers[2].hostPort:
> Invalid value: 29653: Host port 29653 is not allowed to be used.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
This commit is added to use canary csi-provisioner image
to test different sc pvc-pvc cloning feature, which is not
yet present in released versions.
refer:
https://github.com/kubernetes-csi/external-provisioner/pull/699
Signed-off-by: Rakshith R <rar@redhat.com>
This commit make use of latest sidecars of livenessprobe and
node driver registrar in NFS driver deployment.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
This commit remove the clusterRole and Binding of cephfs node plugin
as the node RBAC is not needed for CephFS.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
When running the kubernetes cluster with one single privileged
PodSecurityPolicy which is allowing everything the nodeplugin
daemonset can fail to start. To be precise the problem is the
defaultAllowPrivilegeEscalation: false configuration in the PSP.
Containers of the nodeplugin daemonset won't start when they
have privileged: true but no allowPrivilegeEscalation in their
container securityContext.
Kubernetes will not schedule if this mismatch exists cannot set
allowPrivilegeEscalation to false and privileged to true:
Signed-off-by: Silvan Loser <silvan.loser@hotmail.ch>
Signed-off-by: Silvan Loser <33911078+losil@users.noreply.github.com>
This commit change the image registry URL for sidecars in the
NFS deployment from `k8s.gcr.io` to `registry.k8s.io` as
the migration is happening from former to the latter.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
This commit change the image registry URL for sidecars in the
RBD deployment from `k8s.gcr.io` to `registry.k8s.io` as
the migration is happening from former to the latter.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
This commit change the image registry URL for sidecars in the
CephFS deployment from `k8s.gcr.io` to `registry.k8s.io` as
the migration is happening from former to the latter.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
Kubernetes CSI now hosts the container-image for the NFS-nodeplugin in
the the k8s.gcr.io instead of the Microsoft registry.
See-also: kubernetes-csi/csi-driver-nfs@7b5b6f344
Signed-off-by: Niels de Vos <ndevos@redhat.com>
This argument in csi-snapshotter sidecar allows us to receive
snapshot-name/snapshot-namespace/snapshotcontent-name metadata in the
CreateSnapshot() request.
For ex:
csi.storage.k8s.io/volumesnapshot/name
csi.storage.k8s.io/volumesnapshot/namespace
csi.storage.k8s.io/volumesnapshotcontent/name
This is a useful information which can be used depend on the use case we
have at our driver. The features like adding metadata to snapshot image
can consume this based on the need.
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
add support to run rbd map and mount -t
commands with the nsenter.
complete design of pod/multus network
is added here https://github.com/rook/rook/
blob/master/design/ceph/multus-network.md#csi-pods
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
These deployment files are heavily based on the CephFS deployment.
Deploying an environment with these files work for me in minikube. This
should make it possible to add e2e testing as well.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The API is extended for generation of the NFS CSIDriver object. The
YAML file under deploy/ was created by `yamlgen`.
The contents of the csidriver.yaml file is heavily based on the upstream
CSIDriver from the Kubernetes csi-driver-nfs project.
Because ./tools/yamlgen uses the API, it gets copied under vendor/ .
This causes two copies of the API to be included in the repository, but
that can not be prevented, it seems.
See-also: https://github.com/kubernetes-csi/csi-driver-nfs/blob/master/deploy/csi-nfs-driverinfo.yaml
Signed-off-by: Niels de Vos <ndevos@redhat.com>
OIDC token file path has been modified from
`/var/run/secrets/token` to `/run/secrets/tokens`.
This has been done to ensure compliance with
FHS 3.0.
refer:
https://refspecs.linuxfoundation.org/FHS_3.0/fhs/ch05s13.html
Signed-off-by: Rakshith R <rar@redhat.com>
With Amazon STS and kubernetes cluster is configured with
OIDC identity provider, credentials to access Amazon KMS
can be fetched using oidc-token(serviceaccount token).
Each tenant/namespace needs to create a secret with aws region,
role and CMK ARN.
Ceph-CSI will assume the given role with oidc token and access
aws KMS, with given CMK to encrypt/decrypt DEK which will stored
in the image metdata.
Refer: https://docs.aws.amazon.com/STS/latest/APIReference/welcome.htmlResolves: #2879
Signed-off-by: Rakshith R <rar@redhat.com>
Mounts managed by ceph-fuse may get corrupted by e.g. the ceph-fuse process
exiting abruptly, or its parent container being terminated, taking down its
child processes with it.
This commit adds checks to NodeStageVolume and NodePublishVolume procedures
to detect whether a mountpoint in staging_target_path and/or target_path is
corrupted, and remount is performed if corruption is detected.
Signed-off-by: Robert Vasek <robert.vasek@cern.ch>
The CentOS 8 repository for Apache Arrow has been removed. This causes
container-image builds fail with the following error:
Errors during downloading metadata for repository 'apache-arrow-centos':
- Status code: 404 for https://apache.jfrog.io/artifactory/arrow/centos/8/x86_64/repodata/repomd.xml (IP: 54.190.66.70)
Error: Failed to download metadata for repo 'apache-arrow-centos': Cannot download repomd.xml: Cannot download repodata/repomd.xml: All mirrors were tried
The Ceph base image has `arrow/centos/8` configured, maybe Apache Arrow
offers a CentOS Stream 8 repository now? Once the Ceph container-image
has been updated, the repository can be enabled again.
Ceph-CSI does not depend on Apache Arrow, so there is no functional
change by disabling the repository.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
updating external resizer image version
from 1.3.0 to latest available release i.e
1.4.0
1.4.0 changelog link
https://github.com/kubernetes-csi/
external-resizer/blob/master/CHANGELOG/CHANGELOG-1.4.md
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
This commit updates sidecars to the latest available version
which is compatible with kubernetes 1.23 and csi spec 1.5
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
Deployments place all sockets for communicating with CSI components in
the shared `/csi` directory. The CSI-Addons socket was introduced
recently, but not configured to be in the same location (by default
placed in `/tmp`).
Signed-off-by: Niels de Vos <ndevos@redhat.com>
The generated files under the deploy/ directory contain an empty YAML
document that may cause confusion for some versions of kubectl. Dropping
the unneeded `---` start of the file for the header should make parsing
of the deployment artifacts a little less error prone.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
This initial version of yamlgen generates deploy/scc.yaml based on the
deployment artifact that is provided by the new api/deploy/ocp package.
Signed-off-by: Niels de Vos <ndevos@redhat.com>
cephfs deployment doesnot need extra permission like
privileged,Capabilities and remove unwanted volumes.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
rbd deployment doesnot need extra permission like
privileged,Capabilities and remove unwanted volumes.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
we dont need securityContext for the rbd provisioner
pod as its not doing any special operations like map
,unmap selinux etc.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
we dont need securityContext for the cephfs provisioner
pod as its not doing any special operations like mounts,
selinux etc.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Problem:
--------
1. rbd-nbd by default logs to /var/log/ceph/ceph-client.admin.log,
Unfortunately, container doesn't have /var/log/ceph directory hence
rbd-nbd is not logging now.
2. Rbd-nbd logs are not persistent across nodeplugin restarts.
Solution:
--------
Provide a host path so that log directory is made available, and the
logs persist on the hostnode across container restarts.
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
- mount host's /etc/selinux in node plugins
- process mount options in all code paths for cephfs volume options
Signed-off-by: Alexandre Lossent <alexandre.lossent@cern.ch>
This change resolves a typo for installing the CSIDriver
resource in Kubernetes clusters before 1.18,
where the apiVersion is incorrect.
See also:
https://kubernetes-csi.github.io/docs/csi-driver-object.html
[ndevos: replace v1betav1 in examples with v1beta1]
Signed-off-by: Thomas Kooi <t.j.kooi@avisi.nl>
Problem:
-------
For rbd nbd userspace mounter backends, after a restart of the nodeplugin
all the mounts will start seeing IO errors. This is because, for rbd-nbd
backends there will be a userspace mount daemon running per volume, post
restart of the nodeplugin pod, there is no way to restore the daemons
back to life.
Solution:
--------
The volume healer is a one-time activity that is triggered at the startup
time of the rbd nodeplugin. It navigates through the list of volume
attachments on the node and acts accordingly.
For now, it is limited to nbd type storage only, but it is flexible and
can be extended in the future for other backend types as needed.
From a few feets above:
This solves a severe problem for nbd backed csi volumes. The healer while
going through the list of volume attachments on the node, if finds the
volume is in attached state and is of type nbd, then it will attempt to
fix the rbd-nbd volumes by sending a NodeStageVolume request with the
required volume attributes like secrets, device name, image attributes,
and etc.. which will finally help start the required rbd-nbd daemons in
the nodeplugin csi-rbdplugin container. This will allow reattaching the
backend images with the right nbd device, thus allowing the applications
to perform IO without any interruptions even after a nodeplugin restart.
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
Nodeplugin needs below cluster roles:
persistentvolumes: get
volumeattachments: list, get
These additional permissions are needed by the volume healer. Volume healer
aims at fixing the volume health issues at the very startup time of the
nodeplugin. As part of its operations, volume healer has to run through
the list of volume attachments and understand details about each
persistentvolume.
The later commits will use these additional cluster roles.
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
