1
0
mirror of https://github.com/ceph/ceph-csi.git synced 2024-12-22 04:50:23 +00:00
ceph-csi/docs/design/proposals/intree-migrate.md

130 lines
6.2 KiB
Markdown
Raw Normal View History

# In-tree storage plugin to CSI Driver Migration
Prior to CSI, Kubernetes provided a powerful volume plugin system. These volume
plugins were “in-tree” meaning their code was part of the core Kubernetes code
and shipped with the core Kubernetes binaries. The CSI migration effort enables
the replacement of existing in-tree storage plugins such as kubernetes.io/rbd
with a corresponding CSI driver. This will be an opt-in feature turned on at
cluster creation time that will redirect in-tree plugin operations to a
corresponding CSI Driver. The main mechanism we will use to migrate plugins is
redirecting in-tree operation calls to the CSI Driver instead of the in-tree
driver, the external components will pick up these in-tree PV's and use a
translation library to translate to CSI Source. For more information on CSI
migration effort
refer [design doc](https://github.com/kubernetes/community/blob/master/contributors/design-proposals/storage/csi-migration.md)
## RBD in-tree plugin to CSI migration
This feature is tracked in Kubernetes for v1.23 as alpha under
[RBD in-tree migration KEP](https://github.com/kubernetes/enhancements/issues/2923)
. Below are the changes identified in the design discussions to enable CSI
migration for RBD plugin:
### ClusterID field in the migration request
Considering the `clusterID` field is a required one for CSI, but in-tree
StorageClass has `monitors` field as a required thing the clusterID will be sent
from the migration library based on the monitors field as like any other CSI
request, Kubernetes storage admin supposed to create a clusterID based on the
monitors hash ( ex: `#md5sum <<<
"<monaddress:port>"`) in the CSI config map and keep the monitors under this
configuration before enabling the migration. While CSI driver receive the volume
ID, it will look at the configmap and figure out the monitors to do the
operations like create volume. Thus, CSI operations are unchanged or untouched
wrt to the clusterID field. In other words, for Ceph CSI this is yet another
request coming in from a client with required parameters. This is an opaque/NOOP
change from the Ceph CSI side for its operations, however mentioning it here as
a reference.
### New Volume ID for existing volume operations
The existing volume will have a volume ID passed in similar to the following
format:
```
mig_mons-<monitors-hash>_image-<imageUUID>_<pool-hash>
```
Details on the hash values:
* Monitors Hash: this carry a hash value (md5sum) which will be acted as the
`clusterID` for the operations in this context.
* ImageUUID: this is the unique UUID generated by Kubernetes for the created
volume.
* PoolHash: this is an encoded string of pool name.
This volume handle is introduced as the existing in-tree volume does not have
the volume ID format (
ex: `0001-0020-b7f67366bb43f32e07d8a261a7840da9 -0000000000000002-c867ff35-3f04-11ec-b823-0242ac110003`)
which Ceph CSI use by default. However, Ceph CSI need certain information like
clusterID, pool, Image name to perform deletion and expansion of the volumes.
The existing in-tree
volume's `NodeStageVolume, NodePublishVolume, NodeUnstageVolume, NodeUnpublishVolume`
operations should work without any issues as those will be tracked as static
volumes for the CSI driver. For `DeleteVolume and ControllerExpandVolume`
operations, from above migration specific volume ID, the CSI driver can derive
the required details and connect to the backend cluster ( as it carry
information like monitors, pool and image name) then perform deletion of the
image or perform resize operation.
### Migration secret for CSI operations
The in-tree secret has a data field called `key` which is the base64 admin
secret key. The ceph CSI driver currently expect the secret to contain data
field `UserKey` for the equivalent. The CSI driver also expect the `UserID`
field which is not available in the in-tree secret by default. This
missing `userID` will be filled (if the username differ than `admin`) in the
migration secret as `adminId` field in the migration request, ie:
`key` field value will be picked up from the migration secret to `UserKey`
field.
`adminId` field value will be picked up from the migration secret to `UserID`
field
if `adminId` field is nil or not set, `UserID` field will be filled with default
value ie `admin`. The above logic get activated only when the secret is a
migration secret, otherwise skipped to the normal workflow as we have today.
CreateVolume, DeleteVolume, NodeStage and ExpandVolume has to validate the
secret and use for its operations:
## Helper functions
Few helper/utility functions will be introduced as part of this effort to handle
migration specific logic to it.
internal/rbd/migration.go
* isMigrationVolID() - Identify the volume ID is of above-mentioned form of
migration volume ID.
* parseMigrationVolID() - Parse the migration volume ID and populate the
required information like `pool`, `image`, `clusterID`..etc. into
`migrationVolID` structure.
* deleteMigratedVolume() - Deletes the volume from the cluster by making use of
the `migrationVolID` structure fields like `pool, image, clusterID`.
internal/util/credentials.go
* isMigrationSecret() - Identify the passed in secret is a migration secret
based on the map values (ex: `key` field) of request secret.
* NewUserCredentialsWithMigration() - This take request secret map as input and
validate whether it is migration secret with the help of isMigrationSecret(),
and return credentials by calling NewUserCredentials().
* ParseAndSetSecretMapFromMigSecret() - Parse the secret map from the migration
request and return a map with adjusted secret fields for the CSI driver to
continue performing secret specific operations.
## CephFS in-tree plugin to CSI migration
This is yet to be done and tracked
at [CephFS in-tree migration KEP](https://github.com/kubernetes/enhancements/issues/2924)
### Additional Reference
[Kubernetes 1.17 Feature: Kubernetes In-Tree to CSI Volume Migration Moves to Beta](https://kubernetes.io/blog/2019/12/09/kubernetes-1-17-feature-csi-migration-beta/)
[Tracker Issue in Ceph CSI](https://github.com/ceph/ceph-csi/issues/2509)
[In-tree storage plugin to CSI Driver Migration KEP](https://github.com/kubernetes/enhancements/issues/625)