sxid

Documentation

Overview

Package sxid provides the NVIDIA SXID error details.

Index

• Constants
• Variables
• func ExtractNVSwitchSXid(line string) int
• func ExtractNVSwitchSXidDeviceUUID(line string) string
• type Detail
  • func GetDetail(id int) (*Detail, bool)
  • func (d Detail) JSON() ([]byte, error)
• type DmesgError
  • func ParseDmesgErrorJSON(data []byte) (*DmesgError, error)
  • func ParseDmesgErrorYAML(data []byte) (*DmesgError, error)
  • func ParseDmesgLogLine(time metav1.Time, line string) (DmesgError, error)
  • func (de *DmesgError) JSON() ([]byte, error)
  • func (de *DmesgError) YAML() ([]byte, error)

Constants

const (
	// e.g.,
	// [111111111.111] nvidia-nvswitch3: SXid (PCI:0000:05:00.0): 12028, Non-fatal, Link 32 egress non-posted PRIV error (First)
	// [131453.740743] nvidia-nvswitch0: SXid (PCI:0000:00:00.0): 20034, Fatal, Link 30 LTSSM Fault Up
	//
	// ref.
	// "D.4 Non-Fatal NVSwitch SXid Errors"
	// https://docs.nvidia.com/datacenter/tesla/pdf/fabric-manager-user-guide.pdf
	RegexNVSwitchSXidDmesg = `SXid.*?: (\d+),`

	// Regex to extract PCI device ID from NVSwitch SXid messages
	RegexNVSwitchSXidDeviceUUID = `SXid \((PCI:[0-9a-fA-F:\.]+)\)`
)

Variables

var (
	CompiledRegexNVSwitchSXidDmesg      = regexp.MustCompile(RegexNVSwitchSXidDmesg)
	CompiledRegexNVSwitchSXidDeviceUUID = regexp.MustCompile(RegexNVSwitchSXidDeviceUUID)
)

Functions

func ExtractNVSwitchSXid

func ExtractNVSwitchSXid(line string) int

Extracts the nvidia NVSwitch SXid error code from the dmesg log line. Returns 0 if the error code is not found. https://docs.nvidia.com/datacenter/tesla/pdf/fabric-manager-user-guide.pdf

func ExtractNVSwitchSXidDeviceUUID

func ExtractNVSwitchSXidDeviceUUID(line string) string

ExtractNVSwitchSXidDeviceUUID extracts the PCI device ID from the dmesg log line. Returns empty string if the device ID is not found.

Types

type Detail

type Detail struct {
	DocumentVersion string `json:"documentation_version"`

	SXid        int    `json:"sxid"`
	Name        string `json:"name"`
	Description string `json:"description"`

	// SuggestedActionsByGPUd is the suggested actions by GPUd.
	SuggestedActionsByGPUd *common.SuggestedActions `json:"suggested_actions_by_gpud,omitempty"`
	// CriticalErrorMarkedByGPUd is true if the GPUd marks this SXid as a critical error.
	// You may use this field to decide whether to alert or not.
	CriticalErrorMarkedByGPUd bool `json:"critical_error_marked_by_gpud"`

	PotentialFatal bool   `json:"potential_fatal"`
	AlwaysFatal    bool   `json:"always_fatal"`
	Impact         string `json:"impact"`
	Recovery       string `json:"recovery"`
	OtherImpact    string `json:"other_impact"`
}

Defines the SXid error information that is static. ref. https://docs.nvidia.com/datacenter/tesla/pdf/fabric-manager-user-guide.pdf

func GetDetail

func GetDetail(id int) (*Detail, bool)

Returns the error if found. Otherwise, returns false.

func (Detail) JSON

func (d Detail) JSON() ([]byte, error)

type DmesgError

type DmesgError struct {
	DeviceUUID string         `json:"device_uuid"`
	Detail     *Detail        `json:"detail"`
	LogItem    query_log.Item `json:"log_item"`
}

func ParseDmesgErrorJSON

func ParseDmesgErrorJSON(data []byte) (*DmesgError, error)

func ParseDmesgErrorYAML

func ParseDmesgErrorYAML(data []byte) (*DmesgError, error)

func ParseDmesgLogLine

func ParseDmesgLogLine(time metav1.Time, line string) (DmesgError, error)

func (*DmesgError) JSON

func (de *DmesgError) JSON() ([]byte, error)

func (*DmesgError) YAML

func (de *DmesgError) YAML() ([]byte, error)