cgroups

package
v1.0.1 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: May 17, 2024 License: Apache-2.0 Imports: 19 Imported by: 0

Documentation

Index

Constants

View Source 
const (
	CgroupNamePrefix = "name="
)
View Source 
const (
	CgroupProcesses = "cgroup.procs"
)

Variables

View Source 
var (
	ErrV1NoUnified = errors.New("invalid configuration: cannot use unified on cgroup v1")
)
View Source 
var SyscontCgroupRoot string = "syscont-cgroup-root"
View Source 
var (
	// Set to true by fs unit tests
	TestMode bool
)

Functions

func ConvertBlkIOToCgroupV2Value 

func ConvertBlkIOToCgroupV2Value(blkIoWeight uint16) uint64

Since the OCI spec is designed for cgroup v1, in some cases there is need to convert from the cgroup v1 configuration to cgroup v2 the formula for BlkIOWeight is y = (1 + (x - 10) * 9999 / 990) convert linearly from [10-1000] to [1-10000]

func ConvertCPUSharesToCgroupV2Value 

func ConvertCPUSharesToCgroupV2Value(cpuShares uint64) uint64

Since the OCI spec is designed for cgroup v1, in some cases there is need to convert from the cgroup v1 configuration to cgroup v2 the formula for cpuShares is y = (1 + ((x - 2) * 9999) / 262142) convert from [2-262144] to [1-10000] 262144 comes from Linux kernel definition "#define MAX_SHARES (1UL << 18)"

func ConvertMemorySwapToCgroupV2Value 

func ConvertMemorySwapToCgroupV2Value(memorySwap, memory int64) (int64, error)

ConvertMemorySwapToCgroupV2Value converts MemorySwap value from OCI spec for use by cgroup v2 drivers. A conversion is needed since Resources.MemorySwap is defined as memory+swap combined, while in cgroup v2 swap is a separate value.

func CopyFile 

func CopyFile(source, dest string) error

func EnterPid 

func EnterPid(cgroupPaths map[string]string, pid int) error

func FindCgroupMountpoint 

func FindCgroupMountpoint(cgroupPath, subsystem string) (string, error)

https://www.kernel.org/doc/Documentation/cgroup-v1/cgroups.txt

func FindCgroupMountpointAndRoot 

func FindCgroupMountpointAndRoot(cgroupPath, subsystem string) (string, string, error)

func GetAllPids 

func GetAllPids(path string) ([]int, error)

GetAllPids returns all pids, that were added to cgroup at path and to all its subcgroups.

func GetAllSubsystems 

func GetAllSubsystems() ([]string, error)

GetAllSubsystems returns all the cgroup subsystems supported by the kernel

func GetHugePageSize 

func GetHugePageSize() ([]string, error)

func GetInitCgroup 

func GetInitCgroup(subsystem string) (string, error)

func GetInitCgroupPath 

func GetInitCgroupPath(subsystem string) (string, error)

func GetOwnCgroup 

func GetOwnCgroup(subsystem string) (string, error)

GetOwnCgroup returns the relative path to the cgroup docker is running in.

func GetOwnCgroupPath 

func GetOwnCgroupPath(subsystem string) (string, error)

func GetPids 

func GetPids(dir string) ([]int, error)

GetPids returns all pids, that were added to cgroup at path.

func IsCgroup2UnifiedMode 

func IsCgroup2UnifiedMode() bool

IsCgroup2UnifiedMode returns whether we are running in cgroup v2 unified mode.

func IsNotFound 

func IsNotFound(err error) bool

func NewNotFoundError 

func NewNotFoundError(sub string) error

func OpenFile 

func OpenFile(dir, file string, flags int) (*os.File, error)

OpenFile opens a cgroup file in a given dir with given flags. It is supposed to be used for cgroup files only.

func ParseCgroupFile 

func ParseCgroupFile(path string) (map[string]string, error)

ParseCgroupFile parses the given cgroup file, typically /proc/self/cgroup or /proc/<pid>/cgroup, into a map of subsystems to cgroup paths, e.g. 

"cpu": "/user.slice/user-1000.slice"
"pids": "/user.slice/user-1000.slice"

etc.

Note that for cgroup v2 unified hierarchy, there are no per-controller cgroup paths, so the resulting map will have a single element where the key is empty string ("") and the value is the cgroup path the <pid> is in.

func PathExists 

func PathExists(path string) bool

func ReadFile 

func ReadFile(dir, file string) (string, error)

ReadFile reads data from a cgroup file in dir. It is supposed to be used for cgroup files only.

func RemovePath 

func RemovePath(path string) error

RemovePath aims to remove cgroup path. It does so recursively, by removing any subdirectories (sub-cgroups) first.

func RemovePaths 

func RemovePaths(paths map[string]string) (err error)

RemovePaths iterates over the provided paths removing them. We trying to remove all paths five times with increasing delay between tries. If after all there are not removed cgroups - appropriate error will be returned.

func WriteCgroupProc 

func WriteCgroupProc(dir string, pid int) error

WriteCgroupProc writes the specified pid into the cgroup's cgroup.procs file

func WriteFile 

func WriteFile(dir, file, data string) error

WriteFile writes data to a cgroup file in dir. It is supposed to be used for cgroup files only.

Types

type BlkioStatEntry 

type BlkioStatEntry struct {
	Major uint64 `json:"major,omitempty"`
	Minor uint64 `json:"minor,omitempty"`
	Op    string `json:"op,omitempty"`
	Value uint64 `json:"value,omitempty"`
}

type BlkioStats 

type BlkioStats struct {
	// number of bytes tranferred to and from the block device
	IoServiceBytesRecursive []BlkioStatEntry `json:"io_service_bytes_recursive,omitempty"`
	IoServicedRecursive     []BlkioStatEntry `json:"io_serviced_recursive,omitempty"`
	IoQueuedRecursive       []BlkioStatEntry `json:"io_queue_recursive,omitempty"`
	IoServiceTimeRecursive  []BlkioStatEntry `json:"io_service_time_recursive,omitempty"`
	IoWaitTimeRecursive     []BlkioStatEntry `json:"io_wait_time_recursive,omitempty"`
	IoMergedRecursive       []BlkioStatEntry `json:"io_merged_recursive,omitempty"`
	IoTimeRecursive         []BlkioStatEntry `json:"io_time_recursive,omitempty"`
	SectorsRecursive        []BlkioStatEntry `json:"sectors_recursive,omitempty"`
}

type CPUSetStats 

type CPUSetStats struct {
	// List of the physical numbers of the CPUs on which processes
	// in that cpuset are allowed to execute
	CPUs []uint16 `json:"cpus,omitempty"`
	// cpu_exclusive flag
	CPUExclusive uint64 `json:"cpu_exclusive"`
	// List of memory nodes on which processes in that cpuset
	// are allowed to allocate memory
	Mems []uint16 `json:"mems,omitempty"`
	// mem_hardwall flag
	MemHardwall uint64 `json:"mem_hardwall"`
	// mem_exclusive flag
	MemExclusive uint64 `json:"mem_exclusive"`
	// memory_migrate flag
	MemoryMigrate uint64 `json:"memory_migrate"`
	// memory_spread page flag
	MemorySpreadPage uint64 `json:"memory_spread_page"`
	// memory_spread slab flag
	MemorySpreadSlab uint64 `json:"memory_spread_slab"`
	// memory_pressure
	MemoryPressure uint64 `json:"memory_pressure"`
	// sched_load balance flag
	SchedLoadBalance uint64 `json:"sched_load_balance"`
	// sched_relax_domain_level
	SchedRelaxDomainLevel int64 `json:"sched_relax_domain_level"`
}

type CgroupType 

type CgroupType int
const (
	Cgroup_v1_fs CgroupType = iota
	Cgroup_v1_systemd
	Cgroup_v2_fs
	Cgroup_v2_systemd
)

type CpuStats 

type CpuStats struct {
	CpuUsage       CpuUsage       `json:"cpu_usage,omitempty"`
	ThrottlingData ThrottlingData `json:"throttling_data,omitempty"`
}

type CpuUsage 

type CpuUsage struct {
	// Total CPU time consumed.
	// Units: nanoseconds.
	TotalUsage uint64 `json:"total_usage,omitempty"`
	// Total CPU time consumed per core.
	// Units: nanoseconds.
	PercpuUsage []uint64 `json:"percpu_usage,omitempty"`
	// CPU time consumed per core in kernel mode
	// Units: nanoseconds.
	PercpuUsageInKernelmode []uint64 `json:"percpu_usage_in_kernelmode"`
	// CPU time consumed per core in user mode
	// Units: nanoseconds.
	PercpuUsageInUsermode []uint64 `json:"percpu_usage_in_usermode"`
	// Time spent by tasks of the cgroup in kernel mode.
	// Units: nanoseconds.
	UsageInKernelmode uint64 `json:"usage_in_kernelmode"`
	// Time spent by tasks of the cgroup in user mode.
	// Units: nanoseconds.
	UsageInUsermode uint64 `json:"usage_in_usermode"`
}

CpuUsage denotes the usage of a CPU. All CPU stats are aggregate since container inception.

type HugetlbStats 

type HugetlbStats struct {
	// current res_counter usage for hugetlb
	Usage uint64 `json:"usage,omitempty"`
	// maximum usage ever recorded.
	MaxUsage uint64 `json:"max_usage,omitempty"`
	// number of times hugetlb usage allocation failure.
	Failcnt uint64 `json:"failcnt"`
}

type Manager 

type Manager interface {
	// Applies cgroup configuration to the process with the specified pid
	Apply(pid int) error

	// Returns the PIDs inside the cgroup set
	GetPids() ([]int, error)

	// Returns the PIDs inside the cgroup set & all sub-cgroups
	GetAllPids() ([]int, error)

	// Returns statistics for the cgroup set
	GetStats() (*Stats, error)

	// Toggles the freezer cgroup according with specified state
	Freeze(state configs.FreezerState) error

	// Destroys the cgroup set & all sub-cgroups
	Destroy() error

	// Path returns a cgroup path to the specified controller/subsystem.
	// For cgroupv2, the argument is unused and can be empty.
	Path(string) string

	// Sets the cgroup as configured.
	Set(container *configs.Config) error

	// GetPaths returns cgroup path(s) to save in a state file in order to restore later.
	//
	// For cgroup v1, a key is cgroup subsystem name, and the value is the path
	// to the cgroup for this subsystem.
	//
	// For cgroup v2 unified hierarchy, a key is "", and the value is the unified path.
	GetPaths() map[string]string

	// GetCgroups returns the cgroup data as configured.
	GetCgroups() (*configs.Cgroup, error)

	// GetFreezerState retrieves the current FreezerState of the cgroup.
	GetFreezerState() (configs.FreezerState, error)

	// Whether the cgroup path exists or not
	Exists() bool

	// sysbox-runc: creates a child cgroup that will serve as the cgroup root
	// exposed inside the system container. We don't need a corresponding
	// destroy method because the existing Destroy() method will destroy the
	// child cgroup.
	CreateChildCgroup(container *configs.Config) error

	// sysbox-runc: applies child cgroup configuration to the process with the specified
	// pid. Must be called after Apply() has been called because Apply() configures
	// internal state in the cgroup manager that ApplyChildCgroup() does not. This
	// awkwardness could be avoided if this interface had a separate Create() method as
	// currently Apply() serves as both create and apply.
	ApplyChildCgroup(pid int) error

	// sysbox-runc: same as GetPaths(), but returns child cgroup paths
	GetChildCgroupPaths() map[string]string

	// sysbox-runc: get the type of the cgroup manager
	GetType() CgroupType
}

type MemoryData 

type MemoryData struct {
	Usage    uint64 `json:"usage,omitempty"`
	MaxUsage uint64 `json:"max_usage,omitempty"`
	Failcnt  uint64 `json:"failcnt"`
	Limit    uint64 `json:"limit"`
}

type MemoryStats 

type MemoryStats struct {
	// memory used for cache
	Cache uint64 `json:"cache,omitempty"`
	// usage of memory
	Usage MemoryData `json:"usage,omitempty"`
	// usage of memory + swap
	SwapUsage MemoryData `json:"swap_usage,omitempty"`
	// usage of kernel memory
	KernelUsage MemoryData `json:"kernel_usage,omitempty"`
	// usage of kernel TCP memory
	KernelTCPUsage MemoryData `json:"kernel_tcp_usage,omitempty"`
	// usage of memory pages by NUMA node
	// see chapter 5.6 of memory controller documentation
	PageUsageByNUMA PageUsageByNUMA `json:"page_usage_by_numa,omitempty"`
	// if true, memory usage is accounted for throughout a hierarchy of cgroups.
	UseHierarchy bool `json:"use_hierarchy"`

	Stats map[string]uint64 `json:"stats,omitempty"`
}

type Mount 

type Mount struct {
	Mountpoint string
	Root       string
	Subsystems []string
}

func GetCgroupMounts 

func GetCgroupMounts(all bool) ([]Mount, error)

GetCgroupMounts returns the mounts for the cgroup subsystems. all indicates whether to return just the first instance or all the mounts. This function should not be used from cgroupv2 code, as in this case all the controllers are available under the constant unifiedMountpoint.

func (Mount) GetOwnCgroup 

func (m Mount) GetOwnCgroup(cgroups map[string]string) (string, error)

type NotFoundError 

type NotFoundError struct {
	Subsystem string
}

func (*NotFoundError) Error 

func (e *NotFoundError) Error() string

type PageStats 

type PageStats struct {
	Total uint64           `json:"total,omitempty"`
	Nodes map[uint8]uint64 `json:"nodes,omitempty"`
}

type PageUsageByNUMA 

type PageUsageByNUMA struct {
	// Embedding is used as types can't be recursive.
	PageUsageByNUMAInner
	Hierarchical PageUsageByNUMAInner `json:"hierarchical,omitempty"`
}

type PageUsageByNUMAInner 

type PageUsageByNUMAInner struct {
	Total       PageStats `json:"total,omitempty"`
	File        PageStats `json:"file,omitempty"`
	Anon        PageStats `json:"anon,omitempty"`
	Unevictable PageStats `json:"unevictable,omitempty"`
}

type PidsStats 

type PidsStats struct {
	// number of pids in the cgroup
	Current uint64 `json:"current,omitempty"`
	// active pids hard limit
	Limit uint64 `json:"limit,omitempty"`
}

type RdmaEntry 

type RdmaEntry struct {
	Device     string `json:"device,omitempty"`
	HcaHandles uint32 `json:"hca_handles,omitempty"`
	HcaObjects uint32 `json:"hca_objects,omitempty"`
}

type RdmaStats 

type RdmaStats struct {
	RdmaLimit   []RdmaEntry `json:"rdma_limit,omitempty"`
	RdmaCurrent []RdmaEntry `json:"rdma_current,omitempty"`
}

type Stats 

type Stats struct {
	CpuStats    CpuStats    `json:"cpu_stats,omitempty"`
	CPUSetStats CPUSetStats `json:"cpuset_stats,omitempty"`
	MemoryStats MemoryStats `json:"memory_stats,omitempty"`
	PidsStats   PidsStats   `json:"pids_stats,omitempty"`
	BlkioStats  BlkioStats  `json:"blkio_stats,omitempty"`
	// the map is in the format "size of hugepage: stats of the hugepage"
	HugetlbStats map[string]HugetlbStats `json:"hugetlb_stats,omitempty"`
	RdmaStats    RdmaStats               `json:"rdma_stats,omitempty"`
}

func NewStats 

func NewStats() *Stats

type ThrottlingData 

type ThrottlingData struct {
	// Number of periods with throttling active
	Periods uint64 `json:"periods,omitempty"`
	// Number of periods when the container hit its throttling limit.
	ThrottledPeriods uint64 `json:"throttled_periods,omitempty"`
	// Aggregate time the container was throttled for in nanoseconds.
	ThrottledTime uint64 `json:"throttled_time,omitempty"`
}

Source Files

View all Source files

Directories

Path Synopsis
devicefilter
Package devicefilter contains eBPF device filter program
 Package devicefilter contains eBPF device filter program

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.