chaos-mesh / chaosd Goto Github PK

View Code? Open in Web Editor NEW

122.0 14.0 63.0 27.74 MB

A Chaos Engineering toolkit.

License: Apache License 2.0

Makefile 1.10% Go 91.16% Shell 5.50% C 0.19% Dockerfile 0.37% Python 1.68%

chaos-engineering fault-injection chaos chaos-mesh chaos-testing

chaosd's Introduction

chaosd

chaosd is an easy-to-use Chaos Engineering tool used to inject failures to a physical node.

Document

For details about the introduction and usage of chaosd, refer to the documentation.

Types of fault

You can use Chaosd to simulate the following fault types:

Process: Injects faults into the processes. Operations such as killing the process or stopping the process are supported.
Network: Injects faults into the network of physical machines. Operations such as increasing network latency, losing packets, and corrupting packets are supported.
Stress: Injects stress on the CPU or memory of the physical machines.
Disk: Injects faults into disks of the physical machines. Operations such as increasing disk load of reads and writes, and filling disks are supported.
Host: Injects faults into the physical machine. Operations such as shutdown the physical machine are supported.

For details about the introduction and usage of each fault type, refer to the related documentation.

Work modes

You can use Chaosd in the following modes:

Command-line mode: Run Chaosd directly as a command-line tool to inject and recover faults.
Service mode: Run Chaosd as a service in the background, to inject and recover faults by sending HTTP requests.

Prerequisites

Before deploying chaosd, make sure the following items have been installed:

tc
ipset
iptables
stress-ng (required when install chaosd by building from source code)
byteman(required when install chaosd by building from source code)

Install

You can either build directly from the source or download the binary to finish the installation.

Build from source code

Build chaosd:
```
make chaosd
```
Build or download tools related to Chaosd:
```
make chaos-tools
```
Put Chaosd into PATH:
```
mv ./bin /usr/local/chaosd
export PATH=$PATH:/usr/local/chaosd
```

Download binary

Download the latest unstable version by executing the command below:

curl -fsSL -o chaosd-latest-linux-amd64.tar.gz https://mirrors.chaos-mesh.org/chaosd-latest-linux-amd64.tar.gz

If you want to download the release version, you can replace the latest in the above command with the version number. For example, download v1.1.1 by executing the command below:

curl -fsSL -o chaosd-v1.1.1-linux-amd64.tar.gz https://mirrors.chaos-mesh.org/chaosd-v1.1.1-linux-amd64.tar.gz

Then uncompress the archived file：

tar zxvf chaosd-latest-linux-amd64.tar.gz

Put Chaosd into PATH:

mv ./chaosd-latest-linux-amd64 /usr/local/chaosd
export PATH=$PATH:/usr/local/chaosd

chaosd's People

Contributors

Stargazers

Watchers

chaosd's Issues

build docker image

build docker image for chaosd, and the image should contain stress-ng, iptables, ipset and other tools.

support kill vm

chaosd need to support kill vm, can refer to https://github.com/pingcap/tidb-operator/blob/845a716d41ad1e1ae2579048e50f8a9516e975ab/tests/pkg/fault-trigger/manager/vm_virsh.go#L45

stress-ng exit when chaod exit with interrupt signal

seems that the go.uber.org/fx not handle the SIGINT signal, when using kill to exit the chaosd, the stress-ng process is still alive.

disk attack: support inject write/read payload for specified duration

For example, I want to inject a write payload while an experiment, and the experiment need to run for a long time. I need to write a shell script like the below:

#!/bin/bash

for((i=1;i<=100;i++));
do
    ./chaosd attack disk add-payload write -p /chaosd_data -n 2 --size 100G
    rm /home/root/chaosd_data
done

I can't set the size with a large value, because the storage capacity is limited.

Maybe we should add duration config for disk attack, and chaosd will write and delete automatically.

disk fill: num must not be zero

run chaosd with server mode, and execute command below:

curl -X POST 127.0.0.1:31767/api/attack/disk -H "Content-Type:application/json" -d '{"action":"fill","path":"/tmp/test", "fill-by-fallocate":false,"percent":"50"}'

get error:

{"error":true,"message":"error.api.invalid_request: error.core.attack_config_validation_error: attack config validation failed, cause: num must not be zero","code":"error.api.invalid_request","full_text":"error.api.invalid_request: error.core.attack_config_validation_error: attack config validation failed, cause: num must not be zero\n at github.com/chaos-mesh/chaosd/pkg/server/httpserver.(*httpServer).createDiskAttack()\n\t/__w/chaosd/chaosd/pkg/server/httpserver/server.go:216\n at github.com/gin-gonic/gin.(*Context).Next()\n\t/github/home/go/pkg/mod/github.com/gin-gonic/[email protected]/context.go:161\n at github.com/chaos-mesh/chaosd/pkg/server/utils.MWHandleErrors.func1()\n\t/__w/chaosd/chaosd/pkg/server/utils/error.go:47\n at github.com/gin-gonic/gin.(*Context).Next()\n\t/github/home/go/pkg/mod/github.com/gin-gonic/[email protected]/context.go:161\n at github.com/gin-gonic/gin.RecoveryWithWriter.func1()\n\t/github/home/go/pkg/mod/github.com/gin-gonic/[email protected]/recovery.go:83\n at github.com/gin-gonic/gin.(*Context).Next()\n\t/github/home/go/pkg/mod/github.com/gin-gonic/[email protected]/context.go:161\n at github.com/gin-gonic/gin.LoggerWithConfig.func1()\n\t/github/home/go/pkg/mod/github.com/gin-gonic/[email protected]/logger.go:241\n at github.com/gin-gonic/gin.(*Context).Next()\n\t/github/home/go/pkg/mod/github.com/gin-gonic/[email protected]/context.go:161\n at github.com/gin-gonic/gin.(*Engine).handleHTTPRequest()\n\t/github/home/go/pkg/mod/github.com/gin-gonic/[email protected]/gin.go:409\n at github.com/gin-gonic/gin.(*Engine).ServeHTTP()\n\t/github/home/go/pkg/mod/github.com/gin-gonic/[email protected]/gin.go:367\n at net/http.serverHandler.ServeHTTP()\n\t/__t/go/1.16.2/x64/src/net/http/server.go:2887\n at net/http.(*conn).serve()\n\t/__t/go/1.16.2/x64/src/net/http/server.go:1952\n at runtime.goexit()\n\t/__t/go/1.16.2/x64/src/runtime/asm_amd64.s:1371"}

support shutdown

chaosd needs to support shutdown the host machine which the chaosd executed on.

just execute the shutdown command in chaosd is OK.

support DNS type under Network attack

This feature maybe includes:

update the address of the DNS server to a wrong value
map a host with a wrong address

This can be implement by modify the file /etc/hosts and /etc/resolv.conf

support recover all network experiment

users may forget to recover network experiment, we can add a new flag to recover all network experiment

请问下带简单web界面的工具有github地址吗？类似chaos-mesh的那种。

Thank you!

network attack support partition

partition can use loss 100% to implement this function, but loss use tc, it works on the transport layer.

It's better to use iptables to implement the partition function.

print chaos mesh's log when log level is debug

now chaosd will print the log in chaos mesh, most of them only used when debug
maybe we can add a judgement before here, execute it only when log level is debug

chaosd/cmd/main.go

Line 68 in 693b6f7

ctrl.SetLogger(ctrlzap.New(ctrlzap.UseDevMode(true)))

[discussion] could we replace stress-ng with other tools written by golang

when users try to inject stress by chaosd, users must install stress-ng, and stress-ng require some dependencies, just like:

warning: stress-ng-0.07.29-2.el7.aarch64.rpm: Header V3 RSA/SHA256 Signature, key ID 352c64e5: NOKEY
error: Failed dependencies:
	ld-linux-aarch64.so.1()(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	ld-linux-aarch64.so.1(GLIBC_2.17)(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	libaio.so.1()(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	libaio.so.1(LIBAIO_0.1)(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	libaio.so.1(LIBAIO_0.4)(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	libbsd.so.0()(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	libbsd.so.0(LIBBSD_0.0)(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	libbsd.so.0(LIBBSD_0.3)(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	libbsd.so.0(LIBBSD_0.5)(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	libcrypt.so.1(GLIBC_2.17)(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	libm.so.6(GLIBC_2.17)(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	libpthread.so.0(GLIBC_2.17)(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	librt.so.1(GLIBC_2.17)(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	libsctp.so.1()(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64
	libsctp.so.1(VERS_1)(64bit) is needed by stress-ng-0.07.29-2.el7.aarch64

could we use another tool written by golang?

add some test for network chaos

We need to add some tests for network chaos, like integration test or unit test, can refer to https://github.com/chaos-mesh/chaosd/tree/main/test/integration_test

Chaosd can't build in macos because of disk attack

version: fd4d07f

make chaosd
GO15VENDOREXPERIMENT="1" CGO_ENABLED=1 GOOS="" GOARCH="" go build -ldflags '-s -w -X 'github.com/chaos-mesh/chaosd/pkg/version.buildDate=2021-04-20T03:21:49Z' -X 'github.com/chaos-mesh/chaosd/pkg/version.gitCommit=ab8818565272ba03f0a1bd4f55e3a78fe1a06d30' -X 'github.com/chaos-mesh/chaosd/pkg/version.gitVersion=v0.9.0-10+ab8818565272ba'' -tags "" -o bin/chaosd ./cmd/main.go
# github.com/chaos-mesh/chaosd/pkg/server/chaosd
pkg/server/chaosd/disk.go:136:21: s.Frsize undefined (type syscall.Statfs_t has no field or method Frsize)
make: *** [chaosd] Error 2

Implement store pkg and store experiment record to DB

We use store pkg to store experiment records and use them to recover experiment.

At present, we have defined the struct of store pkg and then we need to store these records to DB.

We can run the following commands to test chaosd:

$ make image-chaosd
$ docker run -it pingcap/chaosd:latest sh
$ chaosd  --help 
A command line client to run chaos experiment

Usage:
  chaos [command]

Available Commands:
  attack      Attack related commands
  help        Help about any command
  recover     Recover a chaos experiment
  server      run Chaosd Server

Flags:
  -h, --help         help for chaos
  -u, --url string   chaosd address (default "http://127.0.0.1:31767")

Use "chaos [command] --help" for more information about a command. 

chaosd attack process stop -p xxx 
xxxx uid: xxxx 

chaosd recover xxxx

support bandwidth limit in network attack

same with https://chaos-mesh.org/docs/simulate-network-chaos-on-kubernetes/#bandwidth-example

return 404 code when delete not exist experiment

now will get an error when delete not exist experiment:

$ curl -X DELETE 127.0.0.1:31767/api/attack/1234
{"error":true,"message":"error.api.internal_server_error: record not found","code":"error.api.internal_server_error","full_text":"error.api.internal_server_error: record not found\n at github.com/chaos-mesh/chaosd/pkg/server/httpserver.handleError()\n\t/home/xiang/chaosd/pkg/server/httpserver/server.go:227\n at github.com/chaos-mesh/chaosd/pkg/server/httpserver.(*httpServer).recoverAttack()\n\t/home/xiang/chaosd/pkg/server/httpserver/server.go:216\n at github.com/gin-gonic/gin.(*Context).Next()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/context.go:161\n at github.com/chaos-mesh/chaosd/pkg/server/utils.MWHandleErrors.func1()\n\t/home/xiang/chaosd/pkg/server/utils/error.go:46\n at github.com/gin-gonic/gin.(*Context).Next()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/context.go:161\n at github.com/gin-gonic/gin.RecoveryWithWriter.func1()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/recovery.go:83\n at github.com/gin-gonic/gin.(*Context).Next()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/context.go:161\n at github.com/gin-gonic/gin.LoggerWithConfig.func1()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/logger.go:241\n at github.com/gin-gonic/gin.(*Context).Next()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/context.go:161\n at github.com/gin-gonic/gin.(*Engine).handleHTTPRequest()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/gin.go:409\n at github.com/gin-gonic/gin.(*Engine).ServeHTTP()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/gin.go:367\n at net/http.serverHandler.ServeHTTP()\n\t/usr/local/go/src/net/http/server.go:2887\n at net/http.(*conn).serve()\n\t/usr/local/go/src/net/http/server.go:1952\n at runtime.goexit()\n\t/usr/local/go/src/runtime/asm_amd64.s:1371"}

it's better to return 404 status code.

support install chaosd by yum/apt-get

yum install chaosd
apt-get install chaosd

add config validate and adjustment for server mode

Users need to specify all configs when injecting failure with server mode.

For example, below is a command for network attack:

$ curl -X POST "127.0.0.1:31767/api/attack/network" -H "Content-Type: application/json" -d '{"device": "eth0", "ipaddress": "172.16.4.4", "action": "duplicate", "percent": "50", "correlation": "0"}'

The correlation is 0, but we can not ignore it, otherwise, it will get an error:

curl -X POST "127.0.0.1:31767/api/attack/network" -H "Content-Type: application/json" -d '{"device": "eth0", "ipaddress": "172.16.4.4", "action": "duplicate", "percent": "50"}'
{"error":true,"message":"error.api.internal_server_error: strconv.ParseFloat: parsing \"\": invalid syntax","code":"error.api.internal_server_error","full_text":"error.api.internal_server_error: strconv.ParseFloat: parsing \"\": invalid syntax\n at github.com/chaos-mesh/chaosd/pkg/server/httpserver.(*httpServer).createNetworkAttack()\n\t/home/xiang/chaosd/pkg/server/httpserver/server.go:131\n at github.com/gin-gonic/gin.(*Context).Next()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/context.go:161\n at github.com/chaos-mesh/chaosd/pkg/server/utils.MWHandleErrors.func1()\n\t/home/xiang/chaosd/pkg/server/utils/error.go:46\n at github.com/gin-gonic/gin.(*Context).Next()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/context.go:161\n at github.com/gin-gonic/gin.RecoveryWithWriter.func1()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/recovery.go:83\n at github.com/gin-gonic/gin.(*Context).Next()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/context.go:161\n at github.com/gin-gonic/gin.LoggerWithConfig.func1()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/logger.go:241\n at github.com/gin-gonic/gin.(*Context).Next()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/context.go:161\n at github.com/gin-gonic/gin.(*Engine).handleHTTPRequest()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/gin.go:409\n at github.com/gin-gonic/gin.(*Engine).ServeHTTP()\n\t/home/xiang/go/pkg/mod/github.com/gin-gonic/[email protected]/gin.go:367\n at net/http.serverHandler.ServeHTTP()\n\t/usr/local/go/src/net/http/server.go:2887\n at net/http.(*conn).serve()\n\t/usr/local/go/src/net/http/server.go:1952\n at runtime.goexit()\n\t/usr/local/go/src/runtime/asm_amd64.s:1371"}

So we should add adjustments for some config, set default values for them when they are not specified, and add validations for them, just like the command mode

chaosd/pkg/core/network_command.go

Line 51 in 0ace2ad

func (n *NetworkCommand) Validate() error {

chaosd/cmd/chaosd/ctl/command/network_command.go

Line 54 in 0ace2ad

 cmd.Flags().StringVarP(&nFlag.Correlation, "correlation", "c", "0", "correlation is percentage (10 is 10%)") 

ignore useless log of go.uber.org/fx with command mode

since pr #31 merged, there are many logs of go.uber.org/fx when run with command mode:

./bin/chaosd attack disk fill -p /tmp/temp -s 100
2021/04/14 02:19:58 [Fx] PROVIDE        *config.Config <= github.com/chaos-mesh/chaosd/cmd/chaosd/ctl/command.glob..func1()
2021/04/14 02:19:58 [Fx] PROVIDE        *dbstore.DB <= github.com/chaos-mesh/chaosd/pkg/store/dbstore.NewDBStore()
2021/04/14 02:19:58 [Fx] PROVIDE        core.ExperimentStore <= github.com/chaos-mesh/chaosd/pkg/store/experiment.NewStore()
2021/04/14 02:19:58 [Fx] PROVIDE        core.ExperimentRunStore <= github.com/chaos-mesh/chaosd/pkg/store/experiment.NewRunStore()
2021/04/14 02:19:58 [Fx] PROVIDE        core.IPSetRuleStore <= github.com/chaos-mesh/chaosd/pkg/store/network.NewIPSetRuleStore()
2021/04/14 02:19:58 [Fx] PROVIDE        core.IptablesRuleStore <= github.com/chaos-mesh/chaosd/pkg/store/network.NewIptablesRuleStore()
2021/04/14 02:19:58 [Fx] PROVIDE        core.TCRuleStore <= github.com/chaos-mesh/chaosd/pkg/store/network.NewTCRuleStore()
2021/04/14 02:19:58 [Fx] PROVIDE        *chaosd.Server <= github.com/chaos-mesh/chaosd/pkg/server/chaosd.NewServer()
2021/04/14 02:19:58 [Fx] PROVIDE        *httpserver.httpServer <= github.com/chaos-mesh/chaosd/pkg/server/httpserver.NewServer()
2021/04/14 02:19:58 [Fx] PROVIDE        crclients.ContainerRuntimeInfoClient <= github.com/chaos-mesh/chaosd/pkg/crclient.NewNodeCRClient()
2021/04/14 02:19:58 [Fx] PROVIDE        int <= os.Getpid()
2021/04/14 02:19:58 [Fx] PROVIDE        *chaosdaemon.DaemonServer <= github.com/chaos-mesh/chaos-mesh/pkg/chaosdaemon.NewDaemonServerWithCRClient()
2021/04/14 02:19:58 [Fx] PROVIDE        scheduler.Scheduler <= github.com/chaos-mesh/chaosd/pkg/scheduler.NewScheduler()
2021/04/14 02:19:58 [Fx] PROVIDE        *core.DiskCommand <= github.com/chaos-mesh/chaosd/cmd/chaosd/ctl/command.NewDiskAttackCommand.func1()
2021/04/14 02:19:58 [Fx] PROVIDE        fx.Lifecycle <= go.uber.org/fx.New.func1()
2021/04/14 02:19:58 [Fx] PROVIDE        fx.Shutdowner <= go.uber.org/fx.(*App).shutdowner-fm()
2021/04/14 02:19:58 [Fx] PROVIDE        fx.DotGraph <= go.uber.org/fx.(*App).dotGraph-fm()
2021/04/14 02:19:58 [Fx] INVOKE         github.com/chaos-mesh/chaosd/cmd/chaosd/ctl/command.processDiskAttack()
[2021/04/14 02:19:58.372 +00:00] [INFO] [disk.go:128]
Fill file /tmp/temp successfully, uid: e60f93df-c88c-410b-808d-a8a013f8dbcd

can we shield these logs？

support IO chaos on disk

Note: It is different from the IO chaos in Chaos Mesh

This feature includes the following capabilities：

emulation of real read and write on disk
fill the disk

JVM attack support inject fault to MySQL

include latency, throw exception on specified table

Support running duration settings

For example, the network attack may partition all host, so that can't connect to it and can't recover it anymore. So need to add a running duration settings, network attack will recover automatically.

Add time skew support in chaosd.

Feature : There already has a time skew tool called watchmaker in chaos-mesh . I think it is a good idea to merge it into chaosd.
Since watchmaker seems have not completed the function of recovery ,we also need to complete it.

support JVM chaos

refer to chaos-mesh/chaos-mesh#1264 and https://github.com/chaosblade-io/chaosblade-exec-jvm

disk payload support use multiple threads

disk read/write payload should support to set the thread number, so that Chaosd can read/write in multiple threads.

version `GLIBC_2.28' not found

./chaosd --help
./chaosd: /lib64/libc.so.6: version `GLIBC_2.28' not found (required by ./chaosd)

disk read-payload attack will exist after read all the file

For example, file temp's size is 100M, but we expect to read 200M data by running the command below:

./bin/chaosd attack disk add-payload read --path ./temp --size 200

the command will exit when reading at the end of this file, maybe it is better to read this file again so that the Chaosd read 200M data.

Distribute binaries via Homebrew

What

I'm macOS user and I will be happy if I can install and manage the binary via Homebrew.
Many tools release it via Homebrew and it's very friendly because we can easily install or uninstall easily.

Ideas

$ brew tap chaos-mesh/tap
$ brew install chaosctl

support arm64 platform

chaosd can build and run under arm64 platform

support scheduler

Just like the experiments of Chaos Mesh, chaosd also needs to support scheduler. The config of the scheduler in Chaos Mesh looks like this:

duration: '10s'
  scheduler:
    cron: '@every 15s'

The experiment was run every 15 seconds for 10 seconds according to the example config.

support file attack

Includes:

JVM attack support inject fault to RocketMQ

include latency, throw exception

stress mem bug

How to reproduce

POC

./bin/chaosd attack stress mem -s 10%

Output

stress memory option: [][2021/09/16 18:27:28.221 +08:00] [INFO] [stress.go:67] ["stressors normalize"] [arguments=" --vm 1 --vm-keep--vm-bytes 10%"]
[2021/09/16 18:27:28.223 +08:00] [INFO] [stress.go:83] ["Start stress-ng process successfully"] [command="/usr/local/sbin/stress-ng --vm 1 --vm-keep--vm-bytes 10%"] [Pid=240670]
Attack stress mem successfully, uid: 72493bd6-5a0d-4e81-9860-50ac1b0cefde

you'll notice that the stress command is wrong

/usr/local/sbin/stress-ng --vm 1 --vm-keep--vm-bytes

should be

/usr/local/sbin/stress-ng --vm 1 --vm-keep --vm-bytes

a SPACE between --vm-keep and --vm-bytes is missing

Root Cause

chaosd depend on github.com/chaos-mesh/chaos-mesh/api/v1alpha1 v0.0.0-20210329070828-9be168b2b489
which has bug and is fixed on this commit

Solution

update dependency github.com/chaos-mesh/chaos-mesh/api/v1alpha1 after the fixed commit.
I have not idea why the dependency version is v0.0.0, so I don't make a PR :-(

Hope that will be help.

error.api.internal_server_error: strconv.ParseFloat: parsing \"\": invalid syntax

run chaosd with server mode, and then run the command:

curl -X POST 127.0.0.1:31767/api/attack/network -H "Content-Type:application/json" -d '{"action":"duplicate","ip-address":"172.16.4.4","device":"eth0","percent":"50"}'

will get error:

error.api.internal_server_error: strconv.ParseFloat: parsing \"\": invalid syntax

jvm chaosd throw exception `AttachNotSupportedException`

Use chaosd attach agent to java process, and get a error:

Error: A JNI error has occurred, please check your installation and try again
Exception in thread "main" java.lang.NoClassDefFoundError: com/sun/tools/attach/AttachNotSupportedException
        at java.lang.Class.getDeclaredMethods0(Native Method)
        at java.lang.Class.privateGetDeclaredMethods(Class.java:2701)
        at java.lang.Class.privateGetMethodRecursive(Class.java:3048)
        at java.lang.Class.getMethod0(Class.java:3018)
        at java.lang.Class.getMethod(Class.java:1784)
        at sun.launcher.LauncherHelper.validateMainClass(LauncherHelper.java:650)
        at sun.launcher.LauncherHelper.checkAndLoadMain(LauncherHelper.java:632)
Caused by: java.lang.ClassNotFoundException: com.sun.tools.attach.AttachNotSupportedException
        at java.net.URLClassLoader.findClass(URLClassLoader.java:382)
        at java.lang.ClassLoader.loadClass(ClassLoader.java:418)
        at sun.misc.Launcher$AppClassLoader.loadClass(Launcher.java:352)
        at java.lang.ClassLoader.loadClass(ClassLoader.java:351)
        ... 7 more

Java version: v1.8.0

Network Attack: support simulate network reorder

Feature Request

Describe the feature you'd like:

At present, Chaosd supports simulate network delay, network duplicate, network loss, and network corruption.
But simulated network reordering is not supported. Network reordering is a common real-world event and Chaosd should support it.

Note:

to use reordering, a delay option must be specified, which is required by tc.

add some test for process chaos

We need to add some tests for process chaos, like integration test or unit test, can refer to https://github.com/chaos-mesh/chaosd/tree/main/test/integration_test

Does anyone start developing the dashboard? If not, I could do that.

I can complete all the functions of chaosd server mode in a dashboard.

disk attack: create file and write data to it automatically when the action is read-payload

when action is read-payload, users must specify the path of the file to read.

Chaosd can create a file and write data automatically, and then users don't need to specify the path.

glibc-2.14 not found when run chaosd on centos 6

when I run chaosd on centos 6, error occurred

chaosd: /lib64/libc.so.6: version GLIBC_2.14' not found (required by ./chaosd)

glibc 2.12 was bounded with centos 6. so I can't update the version of glibc

also, I failed compile chaosd on centos 6

the file's size not match the size set by disk write payload

create PhysicalMachineChaos in Chaos Mesh, the YAML config is below:

kind: PhysicalMachineChaos
apiVersion: chaos-mesh.org/v1alpha1
metadata:
  namespace: default
  name: write-payload
spec:
  action: disk-write-payload
  address:
    - http://172.16.112.130:31767
  uid: c11fa67a-ee03-4fab-9e85-77fdece0b1ff
  disk-write-payload:
    size: 100MB
    path: /tmp/test3
    payload_process_num: 5

the file's size should be 100MB, but in the host, the file's size is 19922944:

 ls -l /tmp/test3
-rw-r--r-- 1 root root 19922944 Oct 27 03:56 /tmp/test3

HealthCheck API should always be exported by http server

Now, health-check API will return failure if we send http request to chaosd when chaosd using mTLS mode.
https://github.com/chaos-mesh/chaosd/blob/main/pkg/server/httpserver/server.go#L94
https://github.com/chaos-mesh/chaosd/blob/main/pkg/server/httpserver/auth.go

stress chaos support stress on CPU and Memory at one experiment

We can create stress chaos with CPU type or memory type:

chaosd attack stress cpu...
chaosd attack stress memory...

But we can not create one chaos experiment with both CPU and Memory, we need to refine it.

Inject network chaos failed when iptables version lower than 1.4.20

chaosd attack network delay -d eth0 -i 10.8.198.117 -p tcp -s 7005 -l 3s

2021-12-09T15:36:51.507+0800    INFO    chaos-daemon-server     flush ipset     {"request": "ipsets:{name:\"chaos-11ed9986-f0e4-41\"  cidrs:\"10.8.198.117/32\"}"}
2021-12-09T15:36:51.509+0800    INFO    background-process-manager      build command   {"command": "ipset create chaos-11ed9986-f0e4-41old hash:net"}
2021-12-09T15:36:51.519+0800    INFO    chaos-daemon-server     create ipset    {"command": "/usr/sbin/ipset create chaos-11ed9986-f0e4-41old hash:net"}
2021-12-09T15:36:51.878+0800    INFO    background-process-manager      build command   {"command": "ipset add chaos-11ed9986-f0e4-41old 10.8.198.117/32"}
2021-12-09T15:36:51.878+0800    INFO    chaos-daemon-server     add CIDR to ipset       {"command": "/usr/sbin/ipset add chaos-11ed9986-f0e4-41old 10.8.198.117/32"}
2021-12-09T15:36:51.885+0800    INFO    background-process-manager      build command   {"command": "ipset rename chaos-11ed9986-f0e4-41old chaos-11ed9986-f0e4-41"}
2021-12-09T15:36:51.886+0800    INFO    chaos-daemon-server     rename ipset    {"command": "/usr/sbin/ipset rename chaos-11ed9986-f0e4-41old chaos-11ed9986-f0e4-41"}
2021-12-09T15:36:51.897+0800    INFO    chaos-daemon-server     Set iptables chains     {"request": ""}
2021-12-09T15:36:51.898+0800    INFO    background-process-manager      build command   {"command": "iptables -w -N CHAOS-INPUT"}
2021-12-09T15:36:52.062+0800    ERROR   chaos-daemon-server     error while initializing iptables       {"error": "error code: exit status 2, msg: iptables v1.4.7: option `-w' requires an argument\nTry `iptables -h' or 'iptables --help' for more information.\n"}
github.com/go-logr/zapr.(*zapLogger).Error
        /home/vagrant/go/pkg/mod/github.com/go-logr/[email protected]/zapr.go:128
github.com/chaos-mesh/chaos-mesh/pkg/chaosdaemon.(*DaemonServer).SetIptablesChains
        /home/vagrant/go/pkg/mod/github.com/chaos-mesh/[email protected]/pkg/chaosdaemon/iptables_server.go:45
github.com/chaos-mesh/chaosd/pkg/server/chaosd.(*Server).applyIptables
        /home/vagrant/chaosd/pkg/server/chaosd/network.go:129
github.com/chaos-mesh/chaosd/pkg/server/chaosd.networkAttack.Attack
        /home/vagrant/chaosd/pkg/server/chaosd/network.go:78
github.com/chaos-mesh/chaosd/pkg/server/chaosd.(*Server).ExecuteAttack
        /home/vagrant/chaosd/pkg/server/chaosd/attack.go:105
github.com/chaos-mesh/chaosd/cmd/attack.commonNetworkAttackFunc
        /home/vagrant/chaosd/cmd/attack/network.go:248
reflect.Value.call
        /usr/local/go/src/reflect/value.go:476
reflect.Value.Call
        /usr/local/go/src/reflect/value.go:337
go.uber.org/dig.defaultInvoker
        /home/vagrant/go/pkg/mod/go.uber.org/[email protected]/dig.go:284
go.uber.org/dig.(*Container).Invoke
        /home/vagrant/go/pkg/mod/go.uber.org/[email protected]/dig.go:439
go.uber.org/fx.(*App).executeInvokes
        /home/vagrant/go/pkg/mod/go.uber.org/[email protected]/app.go:692
go.uber.org/fx.New
        /home/vagrant/go/pkg/mod/go.uber.org/[email protected]/app.go:471
github.com/chaos-mesh/chaosd/pkg/utils.FxNewAppWithoutLog
        /home/vagrant/chaosd/pkg/utils/utils.go:27
github.com/chaos-mesh/chaosd/cmd/attack.NewNetworkDelayCommand.func1
        /home/vagrant/chaosd/cmd/attack/network.go:65
github.com/spf13/cobra.(*Command).execute
        /home/vagrant/go/pkg/mod/github.com/spf13/[email protected]/command.go:854
github.com/spf13/cobra.(*Command).ExecuteC
        /home/vagrant/go/pkg/mod/github.com/spf13/[email protected]/command.go:958
github.com/spf13/cobra.(*Command).Execute
        /home/vagrant/go/pkg/mod/github.com/spf13/[email protected]/command.go:895
main.main
        /home/vagrant/chaosd/cmd/main.go:77
runtime.main
        /usr/local/go/src/runtime/proc.go:225
Error: error code: exit status 2, msg: iptables v1.4.7: option `-w' requires an argument
Try `iptables -h' or 'iptables --help' for more information.

support deploy chaosd by TiUP

Now we can only deploy Chaosd manually. If you have a lot of servers, it takes a lot of work to deploy it.
TiUP is a cluster operation and maintenance tool, we can use it to deploy Chaosd.

有完整的arm安装包吗？

测试环境是内网ARM架构，下载源码尝试make，发现需要下载一大堆依赖包

GO15VENDOREXPERIMENT="1" CGO_ENABLED=1 GOOS="" GOARCH="" go build -ldflags '-s -w -X 'github.com/chaos-mesh/chaosd/pkg/version.buildDate=2021-12-07T03:46:43Z'' -tags "" -o bin/chaosd ./cmd/main.go
go: github.com/alecthomas/[email protected]: Get "https://proxy.golang.org/github.com/alecthomas/template/@v/v0.0.0-20190718012654-fb15b899a751.mod": dial tcp: lookup proxy.golang.org on [::1]:53: read udp [::1]:45337->[::1]:53: read: connection refused
go: downloading sigs.k8s.io/controller-runtime v0.4.0
go: downloading github.com/alecthomas/template v0.0.0-20190718012654-fb15b899a751
go: downloading go.uber.org/zap v1.15.0
go: downloading github.com/spf13/cobra v1.1.1
go: downloading go.uber.org/fx v1.13.1
go: downloading github.com/pingcap/log v0.0.0-20200117041106-d28c14d3b1cd
go: downloading github.com/swaggo/swag v1.6.7
go: downloading github.com/olekukonko/tablewriter v0.0.4
go: downloading github.com/pingcap/errors v0.11.5-0.20190809092503-95897b64e011
go: downloading github.com/chaos-mesh/chaos-mesh v0.9.1-0.20210525104133-41e37dd1ac16
go: downloading github.com/spf13/pflag v1.0.5
go: downloading github.com/pkg/errors v0.9.1
go: downloading github.com/robfig/cron/v3 v3.0.1
go: downloading github.com/go-logr/zapr v0.1.0
go: downloading gorm.io/gorm v1.20.7
go: downloading github.com/joomcode/errorx v1.0.1
go: downloading github.com/shirou/gopsutil v0.0.0-20180427012116-c95755e4bcd7
go: downloading github.com/hashicorp/go-multierror v1.1.0
go: downloading github.com/gin-gonic/gin v1.6.3
go: downloading github.com/mitchellh/go-ps v0.0.0-20170309133038-4fdf99ab2936
go: downloading github.com/chaos-mesh/chaos-mesh/api/v1alpha1 v0.0.0-20210519061540-67fe258c7dec
go: downloading gorm.io/driver/sqlite v1.1.4
go: downloading github.com/google/uuid v1.1.2
go: downloading k8s.io/apimachinery v0.17.1-beta.0
go: downloading github.com/alecthomas/units v0.0.0-20151022065526-2efee857e7cf
go: downloading k8s.io/api v0.17.0
go: downloading google.golang.org/grpc v1.33.2
go: github.com/alecthomas/[email protected]: Get "https://proxy.golang.org/github.com/alecthomas/template/@v/v0.0.0-20190718012654-fb15b899a751.mod": dial tcp: lookup proxy.golang.org on [::1]:53: read udp [::1]:45337->[::1]:53: read: connection refused
make: *** [Makefile:62：chaosd] 错误 1