ceph-csi/cmd/cephcsi.go

213 lines
7.1 KiB
Go
Raw Normal View History

/*
Copyright 2019 The Ceph-CSI Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
*/
package main
import (
"flag"
"fmt"
"os"
"runtime"
"time"
"github.com/ceph/ceph-csi/internal/cephfs"
"github.com/ceph/ceph-csi/internal/liveness"
"github.com/ceph/ceph-csi/internal/rbd"
"github.com/ceph/ceph-csi/internal/util"
klog "k8s.io/klog/v2"
)
const (
rbdType = "rbd"
cephfsType = "cephfs"
livenessType = "liveness"
rbdDefaultName = "rbd.csi.ceph.com"
cephfsDefaultName = "cephfs.csi.ceph.com"
livenessDefaultName = "liveness.csi.ceph.com"
pollTime = 60 // seconds
probeTimeout = 3 // seconds
)
var (
conf util.Config
)
func init() {
// common flags
flag.StringVar(&conf.Vtype, "type", "", "driver type [rbd|cephfs|liveness]")
flag.StringVar(&conf.Endpoint, "endpoint", "unix://tmp/csi.sock", "CSI endpoint")
flag.StringVar(&conf.DriverName, "drivername", "", "name of the driver")
flag.StringVar(&conf.NodeID, "nodeid", "", "node id")
flag.StringVar(&conf.InstanceID, "instanceid", "", "Unique ID distinguishing this instance of Ceph CSI among other"+
Make CephFS plugin stateless reusing RADOS based journal scheme This is a part of the stateless set of commits for CephCSI. This commit removes the dependency on config maps to store cephFS provisioned volumes, and instead relies on RADOS based objects and keys, and required CSI VolumeID encoding to detect the provisioned volumes. Changes: - Provide backward compatibility to provisioned volumes by older plugin versions (1.0.0 or older) - Remove Create/Delete support for statically provisioned volumes (fixes #382) - Added namespace support to RADOS OMaps and used the same to store RADOS CSI objects and keys in the CephFS metadata pool - Added support to mention fsname for CephFS provisioning (fixes #359) - Changed field name in CSI Identifier to 'location', to denote a pool or fscid - Updated mounter cache to use new scheme - Required Helm manifests are updated - Required documentation and other manifests are updated - Made driver option 'metadatastorage' as optional, as fresh installs do not need to specify the same Testing done: - Create/Mount/Delete PVC - Create/Delete 5 PVCs - Mount version 1.0.0 PVC - Delete version 1.0.0 PV - Mount Statically defined PV/PVC/Pod - Mount Statically defined version 1.0.0 PV/PVC/Pod - Delete Statically defined version 1.0.0 PV/PVC/Pod - Node restart when mounted to test mountcache - Use InstanceID other than 'default' - RBD basic round of tests, as namespace is added to OMaps - csitest against ceph-fs plugin - NOTE: CephFS plugin still does not detect and address already created volumes but of a different size - Test not providing any value to the metadata storage parameter Signed-off-by: ShyamsundarR <srangana@redhat.com>
2019-05-28 19:03:18 +00:00
" instances, when sharing Ceph clusters across CSI instances for provisioning")
flag.IntVar(&conf.PidLimit, "pidlimit", 0, "the PID limit to configure through cgroups")
flag.BoolVar(&conf.IsControllerServer, "controllerserver", false, "start cephcsi controller server")
flag.BoolVar(&conf.IsNodeServer, "nodeserver", false, "start cephcsi node server")
flag.StringVar(&conf.DomainLabels, "domainlabels", "", "list of kubernetes node labels, that determines the topology"+
" domain the node belongs to, separated by ','")
// cephfs related flags
flag.BoolVar(&conf.ForceKernelCephFS, "forcecephkernelclient", false, "enable Ceph Kernel clients on kernel < 4.17 which support quotas")
// liveness/grpc metrics related flags
flag.IntVar(&conf.MetricsPort, "metricsport", 8080, "TCP port for liveness/grpc metrics requests")
flag.StringVar(&conf.MetricsPath, "metricspath", "/metrics", "path of prometheus endpoint where metrics will be available")
flag.DurationVar(&conf.PollTime, "polltime", time.Second*pollTime, "time interval in seconds between each poll")
flag.DurationVar(&conf.PoolTimeout, "timeout", time.Second*probeTimeout, "probe timeout in seconds")
flag.BoolVar(&conf.EnableGRPCMetrics, "enablegrpcmetrics", false, "[DEPRECATED] enable grpc metrics")
flag.StringVar(&conf.HistogramOption, "histogramoption", "0.5,2,6",
"[DEPRECATED] Histogram option for grpc metrics, should be comma separated value, ex:= 0.5,2,6 where start=0.5 factor=2, count=6")
flag.UintVar(&conf.RbdHardMaxCloneDepth, "rbdhardmaxclonedepth", 8, "Hard limit for maximum number of nested volume clones that are taken before a flatten occurs")
flag.UintVar(&conf.RbdSoftMaxCloneDepth, "rbdsoftmaxclonedepth", 4, "Soft limit for maximum number of nested volume clones that are taken before a flatten occurs")
flag.UintVar(&conf.MaxSnapshotsOnImage, "maxsnapshotsonimage", 450, "Maximum number of snapshots allowed on rbd image without flattening")
flag.BoolVar(&conf.SkipForceFlatten, "skipforceflatten", false,
"skip image flattening if kernel support mapping of rbd images which has the deep-flatten feature")
flag.BoolVar(&conf.Version, "version", false, "Print cephcsi version information")
klog.InitFlags(nil)
if err := flag.Set("logtostderr", "true"); err != nil {
klog.Exitf("failed to set logtostderr flag: %v", err)
}
flag.Parse()
}
func getDriverName() string {
// was explicitly passed a driver name
if conf.DriverName != "" {
return conf.DriverName
}
// select driver name based on volume type
switch conf.Vtype {
case rbdType:
return rbdDefaultName
case cephfsType:
return cephfsDefaultName
case livenessType:
return livenessDefaultName
default:
return ""
}
}
func main() {
if conf.Version {
fmt.Println("Cephcsi Version:", util.DriverVersion)
fmt.Println("Git Commit:", util.GitCommit)
fmt.Println("Go Version:", runtime.Version())
fmt.Println("Compiler:", runtime.Compiler)
fmt.Printf("Platform: %s/%s\n", runtime.GOOS, runtime.GOARCH)
if kv, err := util.GetKernelVersion(); err == nil {
fmt.Println("Kernel:", kv)
}
os.Exit(0)
}
util.DefaultLog("Driver version: %s and Git version: %s", util.DriverVersion, util.GitCommit)
Make CephFS plugin stateless reusing RADOS based journal scheme This is a part of the stateless set of commits for CephCSI. This commit removes the dependency on config maps to store cephFS provisioned volumes, and instead relies on RADOS based objects and keys, and required CSI VolumeID encoding to detect the provisioned volumes. Changes: - Provide backward compatibility to provisioned volumes by older plugin versions (1.0.0 or older) - Remove Create/Delete support for statically provisioned volumes (fixes #382) - Added namespace support to RADOS OMaps and used the same to store RADOS CSI objects and keys in the CephFS metadata pool - Added support to mention fsname for CephFS provisioning (fixes #359) - Changed field name in CSI Identifier to 'location', to denote a pool or fscid - Updated mounter cache to use new scheme - Required Helm manifests are updated - Required documentation and other manifests are updated - Made driver option 'metadatastorage' as optional, as fresh installs do not need to specify the same Testing done: - Create/Mount/Delete PVC - Create/Delete 5 PVCs - Mount version 1.0.0 PVC - Delete version 1.0.0 PV - Mount Statically defined PV/PVC/Pod - Mount Statically defined version 1.0.0 PV/PVC/Pod - Delete Statically defined version 1.0.0 PV/PVC/Pod - Node restart when mounted to test mountcache - Use InstanceID other than 'default' - RBD basic round of tests, as namespace is added to OMaps - csitest against ceph-fs plugin - NOTE: CephFS plugin still does not detect and address already created volumes but of a different size - Test not providing any value to the metadata storage parameter Signed-off-by: ShyamsundarR <srangana@redhat.com>
2019-05-28 19:03:18 +00:00
if conf.Vtype == "" {
logAndExit("driver type not specified")
}
dname := getDriverName()
err := util.ValidateDriverName(dname)
if err != nil {
logAndExit(err.Error())
}
provisioners: add reconfiguring of PID limit The container runtime CRI-O limits the number of PIDs to 1024 by default. When many PVCs are requested at the same time, it is possible for the provisioner to start too many threads (or go routines) and executing 'rbd' commands can start to fail. In case a go routine can not get started, the process panics. The PID limit can be changed by passing an argument to kubelet, but this will affect all pids running on a host. Changing the parameters to kubelet is also not a very elegant solution. Instead, the provisioner pod can change the configuration itself. The pod is running in privileged mode and can write to /sys/fs/cgroup where the limit is configured. With this change, the limit is configured to 'max', just as if there is no limit at all. The logs of the csi-rbdplugin in the provisioner pod will reflect the change it makes when starting the service: $ oc -n rook-ceph logs -c csi-rbdplugin csi-rbdplugin-provisioner-0 .. I0726 13:59:19.737678 1 cephcsi.go:127] Initial PID limit is set to 1024 I0726 13:59:19.737746 1 cephcsi.go:136] Reconfigured PID limit to -1 (max) .. It is possible to pass a different limit on the commandline of the cephcsi executable. The following flag has been added: --pidlimit=<int> the PID limit to configure through cgroups This accepts special values -1 (max) and 0 (default, do not reconfigure). Other integers will be the limit that gets configured in cgroups. Signed-off-by: Niels de Vos <ndevos@redhat.com>
2019-07-26 12:36:43 +00:00
// the driver may need a higher PID limit for handling all concurrent requests
if conf.PidLimit != 0 {
currentLimit, pidErr := util.GetPIDLimit()
if pidErr != nil {
klog.Errorf("Failed to get the PID limit, can not reconfigure: %v", pidErr)
provisioners: add reconfiguring of PID limit The container runtime CRI-O limits the number of PIDs to 1024 by default. When many PVCs are requested at the same time, it is possible for the provisioner to start too many threads (or go routines) and executing 'rbd' commands can start to fail. In case a go routine can not get started, the process panics. The PID limit can be changed by passing an argument to kubelet, but this will affect all pids running on a host. Changing the parameters to kubelet is also not a very elegant solution. Instead, the provisioner pod can change the configuration itself. The pod is running in privileged mode and can write to /sys/fs/cgroup where the limit is configured. With this change, the limit is configured to 'max', just as if there is no limit at all. The logs of the csi-rbdplugin in the provisioner pod will reflect the change it makes when starting the service: $ oc -n rook-ceph logs -c csi-rbdplugin csi-rbdplugin-provisioner-0 .. I0726 13:59:19.737678 1 cephcsi.go:127] Initial PID limit is set to 1024 I0726 13:59:19.737746 1 cephcsi.go:136] Reconfigured PID limit to -1 (max) .. It is possible to pass a different limit on the commandline of the cephcsi executable. The following flag has been added: --pidlimit=<int> the PID limit to configure through cgroups This accepts special values -1 (max) and 0 (default, do not reconfigure). Other integers will be the limit that gets configured in cgroups. Signed-off-by: Niels de Vos <ndevos@redhat.com>
2019-07-26 12:36:43 +00:00
} else {
util.DefaultLog("Initial PID limit is set to %d", currentLimit)
err = util.SetPIDLimit(conf.PidLimit)
provisioners: add reconfiguring of PID limit The container runtime CRI-O limits the number of PIDs to 1024 by default. When many PVCs are requested at the same time, it is possible for the provisioner to start too many threads (or go routines) and executing 'rbd' commands can start to fail. In case a go routine can not get started, the process panics. The PID limit can be changed by passing an argument to kubelet, but this will affect all pids running on a host. Changing the parameters to kubelet is also not a very elegant solution. Instead, the provisioner pod can change the configuration itself. The pod is running in privileged mode and can write to /sys/fs/cgroup where the limit is configured. With this change, the limit is configured to 'max', just as if there is no limit at all. The logs of the csi-rbdplugin in the provisioner pod will reflect the change it makes when starting the service: $ oc -n rook-ceph logs -c csi-rbdplugin csi-rbdplugin-provisioner-0 .. I0726 13:59:19.737678 1 cephcsi.go:127] Initial PID limit is set to 1024 I0726 13:59:19.737746 1 cephcsi.go:136] Reconfigured PID limit to -1 (max) .. It is possible to pass a different limit on the commandline of the cephcsi executable. The following flag has been added: --pidlimit=<int> the PID limit to configure through cgroups This accepts special values -1 (max) and 0 (default, do not reconfigure). Other integers will be the limit that gets configured in cgroups. Signed-off-by: Niels de Vos <ndevos@redhat.com>
2019-07-26 12:36:43 +00:00
if err != nil {
klog.Errorf("Failed to set new PID limit to %d: %v", conf.PidLimit, err)
provisioners: add reconfiguring of PID limit The container runtime CRI-O limits the number of PIDs to 1024 by default. When many PVCs are requested at the same time, it is possible for the provisioner to start too many threads (or go routines) and executing 'rbd' commands can start to fail. In case a go routine can not get started, the process panics. The PID limit can be changed by passing an argument to kubelet, but this will affect all pids running on a host. Changing the parameters to kubelet is also not a very elegant solution. Instead, the provisioner pod can change the configuration itself. The pod is running in privileged mode and can write to /sys/fs/cgroup where the limit is configured. With this change, the limit is configured to 'max', just as if there is no limit at all. The logs of the csi-rbdplugin in the provisioner pod will reflect the change it makes when starting the service: $ oc -n rook-ceph logs -c csi-rbdplugin csi-rbdplugin-provisioner-0 .. I0726 13:59:19.737678 1 cephcsi.go:127] Initial PID limit is set to 1024 I0726 13:59:19.737746 1 cephcsi.go:136] Reconfigured PID limit to -1 (max) .. It is possible to pass a different limit on the commandline of the cephcsi executable. The following flag has been added: --pidlimit=<int> the PID limit to configure through cgroups This accepts special values -1 (max) and 0 (default, do not reconfigure). Other integers will be the limit that gets configured in cgroups. Signed-off-by: Niels de Vos <ndevos@redhat.com>
2019-07-26 12:36:43 +00:00
} else {
s := ""
if conf.PidLimit == -1 {
provisioners: add reconfiguring of PID limit The container runtime CRI-O limits the number of PIDs to 1024 by default. When many PVCs are requested at the same time, it is possible for the provisioner to start too many threads (or go routines) and executing 'rbd' commands can start to fail. In case a go routine can not get started, the process panics. The PID limit can be changed by passing an argument to kubelet, but this will affect all pids running on a host. Changing the parameters to kubelet is also not a very elegant solution. Instead, the provisioner pod can change the configuration itself. The pod is running in privileged mode and can write to /sys/fs/cgroup where the limit is configured. With this change, the limit is configured to 'max', just as if there is no limit at all. The logs of the csi-rbdplugin in the provisioner pod will reflect the change it makes when starting the service: $ oc -n rook-ceph logs -c csi-rbdplugin csi-rbdplugin-provisioner-0 .. I0726 13:59:19.737678 1 cephcsi.go:127] Initial PID limit is set to 1024 I0726 13:59:19.737746 1 cephcsi.go:136] Reconfigured PID limit to -1 (max) .. It is possible to pass a different limit on the commandline of the cephcsi executable. The following flag has been added: --pidlimit=<int> the PID limit to configure through cgroups This accepts special values -1 (max) and 0 (default, do not reconfigure). Other integers will be the limit that gets configured in cgroups. Signed-off-by: Niels de Vos <ndevos@redhat.com>
2019-07-26 12:36:43 +00:00
s = " (max)"
}
util.DefaultLog("Reconfigured PID limit to %d%s", conf.PidLimit, s)
provisioners: add reconfiguring of PID limit The container runtime CRI-O limits the number of PIDs to 1024 by default. When many PVCs are requested at the same time, it is possible for the provisioner to start too many threads (or go routines) and executing 'rbd' commands can start to fail. In case a go routine can not get started, the process panics. The PID limit can be changed by passing an argument to kubelet, but this will affect all pids running on a host. Changing the parameters to kubelet is also not a very elegant solution. Instead, the provisioner pod can change the configuration itself. The pod is running in privileged mode and can write to /sys/fs/cgroup where the limit is configured. With this change, the limit is configured to 'max', just as if there is no limit at all. The logs of the csi-rbdplugin in the provisioner pod will reflect the change it makes when starting the service: $ oc -n rook-ceph logs -c csi-rbdplugin csi-rbdplugin-provisioner-0 .. I0726 13:59:19.737678 1 cephcsi.go:127] Initial PID limit is set to 1024 I0726 13:59:19.737746 1 cephcsi.go:136] Reconfigured PID limit to -1 (max) .. It is possible to pass a different limit on the commandline of the cephcsi executable. The following flag has been added: --pidlimit=<int> the PID limit to configure through cgroups This accepts special values -1 (max) and 0 (default, do not reconfigure). Other integers will be the limit that gets configured in cgroups. Signed-off-by: Niels de Vos <ndevos@redhat.com>
2019-07-26 12:36:43 +00:00
}
}
}
if conf.EnableGRPCMetrics || conf.Vtype == livenessType {
// validate metrics endpoint
conf.MetricsIP = os.Getenv("POD_IP")
if conf.MetricsIP == "" {
klog.Warning("missing POD_IP env var defaulting to 0.0.0.0")
conf.MetricsIP = "0.0.0.0"
}
err = util.ValidateURL(&conf)
if err != nil {
logAndExit(err.Error())
}
}
util.DefaultLog("Starting driver type: %v with name: %v", conf.Vtype, dname)
switch conf.Vtype {
case rbdType:
validateCloneDepthFlag(&conf)
validateMaxSnaphostFlag(&conf)
driver := rbd.NewDriver()
driver.Run(&conf)
case cephfsType:
driver := cephfs.NewDriver()
driver.Run(&conf)
case livenessType:
liveness.Run(&conf)
}
os.Exit(0)
}
func validateCloneDepthFlag(conf *util.Config) {
// keeping hardlimit to 14 as max to avoid max image depth
if conf.RbdHardMaxCloneDepth == 0 || conf.RbdHardMaxCloneDepth > 14 {
logAndExit("rbdhardmaxclonedepth flag value should be between 1 and 14")
}
if conf.RbdSoftMaxCloneDepth > conf.RbdHardMaxCloneDepth {
logAndExit("rbdsoftmaxclonedepth flag value should not be greater than rbdhardmaxclonedepth")
}
}
func validateMaxSnaphostFlag(conf *util.Config) {
// maximum number of snapshots on an image are 510 [1] and 16 images in
// a parent/child chain [2],keeping snapshot limit to 500 to avoid issues.
// [1] https://github.com/torvalds/linux/blob/master/drivers/block/rbd.c#L98
// [2] https://github.com/torvalds/linux/blob/master/drivers/block/rbd.c#L92
if conf.MaxSnapshotsOnImage == 0 || conf.MaxSnapshotsOnImage > 500 {
logAndExit("maxsnapshotsonimage flag value should be between 1 and 500")
}
}
func logAndExit(msg string) {
klog.Errorln(msg)
os.Exit(1)
}