ceph-csi/cmd/cephcsi.go
Niels de Vos 31648c8feb 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>
2019-08-13 14:43:29 +00:00

157 lines
4.3 KiB
Go

/*
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"
"os"
"path"
"path/filepath"
"strings"
"github.com/ceph/ceph-csi/pkg/cephfs"
"github.com/ceph/ceph-csi/pkg/rbd"
"github.com/ceph/ceph-csi/pkg/util"
"k8s.io/klog"
)
const (
rbdType = "rbd"
cephfsType = "cephfs"
rbdDefaultName = "rbd.csi.ceph.com"
cephfsDefaultName = "cephfs.csi.ceph.com"
)
var (
// common flags
vtype = flag.String("type", "", "driver type [rbd|cephfs]")
endpoint = flag.String("endpoint", "unix://tmp/csi.sock", "CSI endpoint")
driverName = flag.String("drivername", "", "name of the driver")
nodeID = flag.String("nodeid", "", "node id")
instanceID = flag.String("instanceid", "", "Unique ID distinguishing this instance of Ceph CSI among other"+
" instances, when sharing Ceph clusters across CSI instances for provisioning")
metadataStorage = flag.String("metadatastorage", "", "metadata persistence method [node|k8s_configmap]")
pluginPath = flag.String("pluginpath", "/var/lib/kubelet/plugins/", "the location of cephcsi plugin")
pidLimit = flag.Int("pidlimit", 0, "the PID limit to configure through cgroups")
// rbd related flags
containerized = flag.Bool("containerized", true, "whether run as containerized")
// cephfs related flags
volumeMounter = flag.String("volumemounter", "", "default volume mounter (possible options are 'kernel', 'fuse')")
mountCacheDir = flag.String("mountcachedir", "", "mount info cache save dir")
)
func init() {
klog.InitFlags(nil)
if err := flag.Set("logtostderr", "true"); err != nil {
klog.Exitf("failed to set logtostderr flag: %v", err)
}
flag.Parse()
}
func getType() string {
if vtype == nil || *vtype == "" {
a0 := path.Base(os.Args[0])
if strings.Contains(a0, rbdType) {
return rbdType
}
if strings.Contains(a0, cephfsType) {
return cephfsType
}
return ""
}
return *vtype
}
func getDriverName() string {
// was explicitly passed a driver name
if driverName != nil && *driverName != "" {
return *driverName
}
// select driver name based on volume type
switch getType() {
case rbdType:
return rbdDefaultName
case cephfsType:
return cephfsDefaultName
default:
return ""
}
}
func main() {
klog.Infof("Driver version: %s and Git version: %s", util.DriverVersion, util.GitCommit)
var cp util.CachePersister
driverType := getType()
if driverType == "" {
klog.Fatalln("driver type not specified")
}
dname := getDriverName()
err := util.ValidateDriverName(dname)
if err != nil {
klog.Fatalln(err) // calls exit
}
csipluginPath := filepath.Join(*pluginPath, dname)
if *metadataStorage != "" {
cp, err = util.CreatePersistanceStorage(
csipluginPath, *metadataStorage, *pluginPath)
if err != nil {
os.Exit(1)
}
}
// the driver may need a higher PID limit for handling all concurrent requests
if pidLimit != nil && *pidLimit != 0 {
currentLimit, err := util.GetPIDLimit()
if err != nil {
klog.Errorf("Failed to get the PID limit, can not reconfigure: %v", err)
} else {
klog.Infof("Initial PID limit is set to %d", currentLimit)
err = util.SetPIDLimit(*pidLimit)
if err != nil {
klog.Errorf("Failed to set new PID limit to %d: %v", *pidLimit, err)
} else {
s := ""
if *pidLimit == -1 {
s = " (max)"
}
klog.Infof("Reconfigured PID limit to %d%s", *pidLimit, s)
}
}
}
klog.Infof("Starting driver type: %v with name: %v", driverType, dname)
switch driverType {
case rbdType:
driver := rbd.NewDriver()
driver.Run(dname, *nodeID, *endpoint, *instanceID, *containerized, cp, driverType)
case cephfsType:
driver := cephfs.NewDriver()
driver.Run(dname, *nodeID, *endpoint, *volumeMounter, *mountCacheDir, *instanceID, csipluginPath, cp, driverType)
default:
klog.Fatalln("invalid volume type", vtype) // calls exit
}
os.Exit(0)
}