rook: Bug flag provided but not defined (csi-snapshotter v8.2.0 + rook 1.16.0)
** Bug Report **
Container csi-snapshotter keep crashing
Expected behavior:
It should not
How to reproduce it (minimal and precise): Install rook v1.16.0 Define external cluster => two deployment using csi-snapshoter v8.2.0 fails.
Logs to submit:
flag provided but not defined: -enable-volume-group-snapshots 2024-12-23T16:30:54.823755853+01:00 Usage of /csi-snapshotter: 2024-12-23T16:30:54.823760058+01:00 -add_dir_header 2024-12-23T16:30:54.823762755+01:00 If true, adds the file directory to the header of the log messages 2024-12-23T16:30:54.823764924+01:00 -alsologtostderr 2024-12-23T16:30:54.823767013+01:00 log to standard error as well as files (no effect when -logtostderr=true) 2024-12-23T16:30:54.823769050+01:00 -csi-address string 2024-12-23T16:30:54.823771084+01:00 Address of the CSI driver socket. (default "/run/csi/socket") 2024-12-23T16:30:54.823773138+01:00 -extra-create-metadata 2024-12-23T16:30:54.823775201+01:00 If set, add snapshot metadata to plugin snapshot requests as parameters. 2024-12-23T16:30:54.823777219+01:00 -feature-gates value 2024-12-23T16:30:54.823779297+01:00 Comma-seprated list of key=value pairs that describe feature gates for alpha/experimental features. Options are: 2024-12-23T16:30:54.823785688+01:00 AllAlpha=true\|false (ALPHA - default=false) 2024-12-23T16:30:54.823787731+01:00 AllBeta=true\|false (BETA - default=false) 2024-12-23T16:30:54.823789725+01:00 CSIVolumeGroupSnapshot=true\|false (BETA - default=false) 2024-12-23T16:30:54.823791749+01:00 -groupsnapshot-name-prefix string 2024-12-23T16:30:54.823794217+01:00 Prefix to apply to the name of a created group snapshot (default "groupsnapshot") 2024-12-23T16:30:54.823796245+01:00 -groupsnapshot-name-uuid-length int 2024-12-23T16:30:54.823801303+01:00 Length in characters for the generated uuid of a created group snapshot. Defaults behavior is to NOT truncate. (default -1) 2024-12-23T16:30:54.823803547+01:00 -http-endpoint :8080 2024-12-23T16:30:54.823805830+01:00 The TCP network address where the HTTP server for diagnostics, including metrics and leader election health check, will listen (example: :8080). The default is empty string, which means the server is disabl
--
| ed. Only one of `--metrics-address` and `--http-endpoint` can be set. 2024-12-23T16:30:54.823808659+01:00 -kube-api-burst int 2024-12-23T16:30:54.823810861+01:00 Burst to use while communicating with the kubernetes apiserver. Defaults to 10. (default 10) 2024-12-23T16:30:54.823812890+01:00 -kube-api-qps float
Mon, Dec 23 2024 4:30:54 pm | QPS to use while communicating with the kubernetes apiserver. Defaults to 5.0. (default 5)
Mon, Dec 23 2024 4:30:54 pm | -kubeconfig string 2024-12-23T16:30:54.823819010+01:00 Absolute path to the kubeconfig file. Required only when running out of cluster. 2024-12-23T16:30:54.823821114+01:00 -leader-election 2024-12-23T16:30:54.823823157+01:00 Enables leader election. 2024-12-23T16:30:54.823825170+01:00 -leader-election-lease-duration duration 2024-12-23T16:30:54.823827507+01:00 Duration, in seconds, that non-leader candidates will wait to force acquire leadership. Defaults to 15 seconds. (default 15s) 2024-12-23T16:30:54.823829580+01:00 -leader-election-namespace string 2024-12-23T16:30:54.823831629+01:00 The namespace where the leader election resource exists. Defaults to the pod namespace if not set. 2024-12-23T16:30:54.823833691+01:00 -leader-election-renew-deadline duration 2024-12-23T16:30:54.823835730+01:00 Duration, in seconds, that the acting leader will retry refreshing leadership before giving up. Defaults to 10 seconds. (default 10s) 2024-12-23T16:30:54.823837831+01:00 -leader-election-retry-period duration 2024-12-23T16:30:54.823839895+01:00 Duration, in seconds, the LeaderElector clients should wait between tries of actions. Defaults to 5 seconds. (default 5s) 2024-12-23T16:30:54.823856198+01:00 -log_backtrace_at value 2024-12-23T16:30:54.823986406+01:00 when logging hits line file:N, emit a stack trace 2024-12-23T16:30:54.823996492+01:00 -log_dir string 2024-12-23T16:30:54.824003434+01:00 If non-empty, write log files in this directory (no effect when -logtostderr=true) 2024-12-23T16:30:54.824006877+01:00 -log_file string 2024-12-23T16:30:54.824010974+01:00 If non-empty, use this log file (no effect when -logtostderr=true) 2024-12-23T16:30:54.824014404+01:00 -log_file_max_size uint 2024-12-23T16:30:54.824018057+01:00 Defines the maximum size a log file can grow to (no effect when -logtostderr=true). Unit is megabytes. If the value is 0, the maximum file size is unlimited. (default 1800) 2024-12-23T16:30:54.824021
| 419+01:00 -logtostderr 2024-12-23T16:30:54.824024843+01:00 log to standard error instead of files (default true) 2024-12-23T16:30:54.824028104+01:00 -metrics-address :8080 2024-12-23T16:30:54.824031668+01:00 (deprecated) The TCP network address where the prometheus metrics endpoint will listen (example: :8080). The default is empty string, which means metrics endpoint is disabled. Only one of `--metrics-address` and `--http-endpoint` can be set.
Mon, Dec 23 2024 4:30:54 pm | -metrics-path /metrics
Mon, Dec 23 2024 4:30:54 pm | The HTTP path where prometheus metrics will be exposed. Default is /metrics. (default "/metrics") 2024-12-23T16:30:54.824059871+01:00 -node-deployment
Mon, Dec 23 2024 4:30:54 pm | Enables deploying the sidecar controller together with a CSI driver on nodes to manage snapshots for node-local volumes. 2024-12-23T16:30:54.824069145+01:00 -one_output 2024-12-23T16:30:54.824073182+01:00 If true, only write logs to their native severity level (vs also writing to each lower severity level; no effect when -logtostderr=true) 2024-12-23T16:30:54.824076529+01:00 -resync-period duration 2024-12-23T16:30:54.824079782+01:00 Resync interval of the controller. Default is 15 minutes (default 15m0s) 2024-12-23T16:30:54.824083185+01:00 -retry-interval-max duration 2024-12-23T16:30:54.824087218+01:00 Maximum retry interval of failed volume snapshot creation or deletion. Default is 5 minutes. (default 5m0s) 2024-12-23T16:30:54.824090568+01:00 -retry-interval-start duration 2024-12-23T16:30:54.824096451+01:00 Initial retry interval of failed volume snapshot creation or deletion. It doubles with each failure, up to retry-interval-max. Default is 1 second. (default 1s) 2024-12-23T16:30:54.824099891+01:00 -skip_headers 2024-12-23T16:30:54.824103387+01:00 If true, avoid header prefixes in the log messages 2024-12-23T16:30:54.824106521+01:00 -skip_log_headers 2024-12-23T16:30:54.824109772+01:00 If true, avoid headers when opening log files (no effect when -logtostderr=true) 2024-12-23T16:30:54.824113575+01:00 -snapshot-name-prefix string 2024-12-23T16:30:54.824116942+01:00 Prefix to apply to the name of a created snapshot (default "snapshot") 2024-12-23T16:30:54.824120241+01:00 -snapshot-name-uuid-length int 2024-12-23T16:30:54.824123562+01:00 Length in characters for the generated uuid of a created snapshot. Defaults behavior is to NOT truncate. (default -1)
Mon, Dec 23 2024 4:30:54 pm | -stderrthreshold value 2024-12-23T16:30:54.824130505+01:00 logs at or above this threshold go to stderr when writing to files and stderr (no effect when -logtostderr=true or -alsologtostderr=true) (default 2)
Mon, Dec 23 2024 4:30:54 pm | -timeout duration 2024-12-23T16:30:54.824137105+01:00 The timeout for any RPCs to the CSI driver. Default is 1 minute. (default 1m0s) 2024-12-23T16:30:54.824140446+01:00 -v value 2024-12-23T16:30:54.824144138+01:00 number for the log level verbosity 2024-12-23T16:30:54.824155403+01:00 -version
Mon, Dec 23 2024 4:30:54 pm | Show version. 2024-12-23T16:30:54.824162281+01:00 -vmodule value 2024-12-23T16:30:54.824165653+01:00 comma-separated list of pattern=N settings for file-filtered logging 2024-12-23T16:30:54.824168914+01:00 -worker-threads int 2024-12-23T16:30:54.824172249+01:00 Number of worker threads. (default 10)
https://github.com/kubernetes-csi/external-snapshotter/pull/1194
If I’m reading this right the feature on v8.2.0 was moved from --enable-volume-group-snapshots to feature-gate to --feature-gates=CSIVolumeGroupSnapshot=true
So to fix this atm you have to edit the deployment to change the option on the csi-snapshotter
Environment:
- Rook version (use
rook versioninside of a Rook Pod): v1.16.0 - latest charts - Storage backend version (e.g. for ceph do
ceph -v): N/A - Kubernetes version (use
kubectl version): v1.28 - Kubernetes cluster type (e.g. Tectonic, GKE, OpenShift): Rancher Governement
About this issue
- Original URL
- State: closed
- Created 2 years ago
- Reactions: 1
- Comments: 26 (5 by maintainers)
Commits related to this issue
- removing "CSI_ENABLE_VOLUME_GROUP_SNAPSHOT" due to issue https://github.com/rook/rook/issues/15219 — committed to SimonStiil/rke2-flux by SimonStiil 2 years ago
- removing "CSI_ENABLE_VOLUME_GROUP_SNAPSHOT" due to issue https://github.com/rook/rook/issues/15219 — committed to SimonStiil/rke2-flux by SimonStiil 2 years ago
@travisn i went on PTO before the blog was released, I have updated https://github.com/rook/rook/pull/15037/ to match the API and also to update Rook document to install Beta CRD.
Rook enables the VGS feature only when the required CRDs are present. As stated in external-snapshotter the Alpha is only for testing and experimenting purposes and beta might get backward compatibility support. Rook doesn’t enable this feature by default unless the user installed the CRD.
Rook checks only the CRD and not for the container arguments as its not what Rook does today and has a plan for it, but this will be a new support coming in csi-operator. Rook currently only relies on the CRD’s to enable the feature. it will be user’s responsibility to update the sidecar images if they are using custom one or use the default sidecar that comes with Rook.