Pod IO Stress

Introduction

• This fault causes disk stress on the application pod. It aims to verify the resiliency of applications that share this disk resource for ephemeral or persistent storage purposes.

Fault execution flow chart

Uses

View the uses of the fault

Disk Pressure or CPU hogs is another very common and frequent scenario we find in kubernetes applications that can result in the eviction of the application replica and impact its delivery. Such scenarios that can still occur despite whatever availability aids K8s provides. These problems are generally referred to as "Noisy Neighbour" problems.

Stressing the disk with continuous and heavy IO for example can cause degradation in reads written by other microservices that use this shared disk for example modern storage solutions for Kubernetes use the concept of storage pools out of which virtual volumes/devices are carved out. Another issue is the amount of scratch space eaten up on a node which leads to the lack of space for newer containers to get scheduled (kubernetes too gives up by applying an "eviction" taint like "disk-pressure") and causes a wholesale movement of all pods to other nodes.

Prerequisites

info

• Ensure that Kubernetes Version > 1.16.

Default Validations

note

The application pods should be in running state before and after chaos injection.

Fault Tunables

Check the Fault Tunables

Optional Fields

Variables	Description	Notes
FILESYSTEM_UTILIZATION_PERCENTAGE	Specify the size as percentage of free space on the file system	Default to 10%
FILESYSTEM_UTILIZATION_BYTES	Specify the size in GigaBytes(GB). FILESYSTEM_UTILIZATION_PERCENTAGE & FILESYSTEM_UTILIZATION_BYTES are mutually exclusive. If both are provided, FILESYSTEM_UTILIZATION_PERCENTAGE is prioritized.
NUMBER_OF_WORKERS	It is the number of IO workers involved in IO disk stress	Default to 4
TOTAL_CHAOS_DURATION	The time duration for chaos (seconds)	Default to 120s
VOLUME_MOUNT_PATH	Fill the given volume mount path
LIB	The chaos lib used to inject the chaos	Default to litmus. Available litmus and pumba.
LIB_IMAGE	Image used to run the stress command	Default to litmuschaos/go-runner:latest
TARGET_PODS	Comma separated list of application pod name subjected to pod io stress chaos	If not provided, it will select target pods randomly based on provided appLabels
PODS_AFFECTED_PERC	The Percentage of total pods to target	Defaults to 0 (corresponds to 1 replica), provide numeric value only
CONTAINER_RUNTIME	container runtime interface for the cluster	Defaults to docker, supported values: docker, containerd and crio for litmus and only docker for pumba LIB
SOCKET_PATH	Path of the containerd/crio/docker socket file	Defaults to /var/run/docker.sock
RAMP_TIME	Period to wait before and after injection of chaos in sec	Eg. 30
SEQUENCE	It defines sequence of chaos execution for multiple target pods	Default value: parallel. Supported: serial, parallel

Fault Examples

Common and Pod specific tunables

Refer the common attributes and Pod specific tunable to tune the common tunables for all fault and pod specific tunables.

Filesystem Utilization Percentage

It stresses the FILESYSTEM_UTILIZATION_PERCENTAGE percentage of total free space available in the pod.

Use the following example to tune this:

# stress the i/o of the targeted pod with FILESYSTEM_UTILIZATION_PERCENTAGE of total free space 
# it is mutually exclusive with the FILESYSTEM_UTILIZATION_BYTES.
# if both are provided then it will use FILESYSTEM_UTILIZATION_PERCENTAGE for stress
apiVersion: litmuschaos.io/v1alpha1
kind: ChaosEngine
metadata:
  name: engine-nginx
spec:
  engineState: "active"
  annotationCheck: "false"
  appinfo:
    appns: "default"
    applabel: "app=nginx"
    appkind: "deployment"
  chaosServiceAccount: litmus-admin
  experiments:
  - name: pod-io-stress
    spec:
      components:
        env:
        # percentage of free space of file system, need to be stressed
        - name: FILESYSTEM_UTILIZATION_PERCENTAGE
          value: '10' #in GB
        - name: TOTAL_CHAOS_DURATION
          VALUE: '60'

Filesystem Utilization Bytes

It stresses the FILESYSTEM_UTILIZATION_BYTES GB of the i/o of the targeted pod. It is mutually exclusive with the FILESYSTEM_UTILIZATION_PERCENTAGE ENV. If FILESYSTEM_UTILIZATION_PERCENTAGE ENV is set then it will use the percentage for the stress otherwise, it will stress the i/o based on FILESYSTEM_UTILIZATION_BYTES ENV.

Use the following example to tune this:

# stress the i/o of the targeted pod with given FILESYSTEM_UTILIZATION_BYTES
# it is mutually exclusive with the FILESYSTEM_UTILIZATION_PERCENTAGE.
# if both are provided then it will use FILESYSTEM_UTILIZATION_PERCENTAGE for stress
apiVersion: litmuschaos.io/v1alpha1
kind: ChaosEngine
metadata:
  name: engine-nginx
spec:
  engineState: "active"
  annotationCheck: "false"
  appinfo:
    appns: "default"
    applabel: "app=nginx"
    appkind: "deployment"
  chaosServiceAccount: litmus-admin
  experiments:
  - name: pod-io-stress
    spec:
      components:
        env:
        # size of io to be stressed
        - name: FILESYSTEM_UTILIZATION_BYTES
          value: '1' #in GB
        - name: TOTAL_CHAOS_DURATION
          VALUE: '60'

Container Runtime Socket Path

It defines the CONTAINER_RUNTIME and SOCKET_PATH ENV to set the container runtime and socket file path.

• CONTAINER_RUNTIME: It supports docker, containerd, and crio runtimes. The default value is docker.
• SOCKET_PATH: It contains path of docker socket file by default(/var/run/docker.sock). For other runtimes provide the appropriate path.

Use the following example to tune this:

## provide the container runtime and socket file path
apiVersion: litmuschaos.io/v1alpha1
kind: ChaosEngine
metadata:
  name: engine-nginx
spec:
  engineState: "active"
  annotationCheck: "false"
  appinfo:
    appns: "default"
    applabel: "app=nginx"
    appkind: "deployment"
  chaosServiceAccount: litmus-admin
  experiments:
  - name: pod-io-stress
    spec:
      components:
        env:
        # runtime for the container
        # supports docker, containerd, crio
        - name: CONTAINER_RUNTIME
          value: 'docker'
        # path of the socket file
        - name: SOCKET_PATH
          value: '/var/run/docker.sock'
        - name: TOTAL_CHAOS_DURATION
          VALUE: '60'

Mount Path

The volume mount path, which needs to be filled. It can be tuned with VOLUME_MOUNT_PATH ENV.

Use the following example to tune this:

# provide the volume mount path, which needs to be filled
apiVersion: litmuschaos.io/v1alpha1
kind: ChaosEngine
metadata:
  name: engine-nginx
spec:
  engineState: "active"
  annotationCheck: "false"
  appinfo:
    appns: "default"
    applabel: "app=nginx"
    appkind: "deployment"
  chaosServiceAccount: litmus-admin
  experiments:
  - name: pod-io-stress
    spec:
      components:
        env:
        # path need to be stressed/filled
        - name: VOLUME_MOUNT_PATH
          value: '/some-dir-in-container'
        - name: TOTAL_CHAOS_DURATION
          VALUE: '60'

Workers For Stress

The worker's count for the stress can be tuned with NUMBER_OF_WORKERS ENV.

Use the following example to tune this:

# number of workers for the stress
apiVersion: litmuschaos.io/v1alpha1
kind: ChaosEngine
metadata:
  name: engine-nginx
spec:
  engineState: "active"
  annotationCheck: "false"
  appinfo:
    appns: "default"
    applabel: "app=nginx"
    appkind: "deployment"
  chaosServiceAccount: litmus-admin
  experiments:
  - name: pod-io-stress
    spec:
      components:
        env:
        # number of io workers 
        - name: NUMBER_OF_WORKERS
          value: '4'
        - name: TOTAL_CHAOS_DURATION
          VALUE: '60'

Pumba Chaos Library

It specifies the Pumba chaos library for the chaos injection. It can be tuned via LIB ENV. The defaults chaos library is litmus.

Use the following example to tune this:

# use the pumba lib for io stress
apiVersion: litmuschaos.io/v1alpha1
kind: ChaosEngine
metadata:
  name: engine-nginx
spec:
  engineState: "active"
  annotationCheck: "false"
  appinfo:
    appns: "default"
    applabel: "app=nginx"
    appkind: "deployment"
  chaosServiceAccount: litmus-admin
  experiments:
  - name: pod-io-stress
    spec:
      components:
        env:
        # name of lib
        # it supports litmus and pumba lib
        - name: LIB
          value: 'pumba'
        - name: TOTAL_CHAOS_DURATION
          VALUE: '60'