tencent cloud

Tencent Kubernetes Engine

소식 및 공지 사항
릴리스 노트
제품 릴리스 기록
제품 소개
제품 장점
제품 아키텍처
시나리오
제품 기능
리전 및 가용존
빠른 시작
신규 사용자 가이드
표준 클러스터를 빠르게 생성
Demo
클라우드에서 컨테이너화된 애플리케이션 배포 Check List
TKE 표준 클러스터 가이드
Tencent Kubernetes Engine(TKE)
클러스터 관리
네트워크 관리
스토리지 관리
Worker 노드 소개
Kubernetes Object Management
워크로드
클라우드 네이티브 서비스 가이드
Tencent Managed Service for Prometheus
TKE Serverless 클러스터 가이드
TKE 클러스터 등록 가이드
실습 튜토리얼
Serverless 클러스터
네트워크
로그
모니터링
유지보수
DevOps
탄력적 스케일링
자주 묻는 질문
클러스터
TKE Serverless 클러스터
유지보수
서비스
이미지 레지스트리
원격 터미널
문서Tencent Kubernetes Engine

Use Systemtap to Identify Pod Exceptions

포커스 모드
폰트 크기
마지막 업데이트 시간: 2024-12-13 14:48:39
This article describes how to use SystemTap to troubleshoot pod issues.

Preparations

Different operating systems have different methods for installing SystemTap and its dependencies. Pick one that suits you.

Ubuntu

1. Run the following command to install SystemTap:
apt install -y systemtap
2. Run the following command to check for dependencies:
stap-prep
The following is a sample result:
Please install linux-headers-4.4.0-104-generic
You need package linux-image-4.4.0-104-generic-dbgsym but it does not seem to be available
 Ubuntu -dbgsym packages are typically in a separate repository
 Follow https://wiki.ubuntu.com/DebuggingProgramCrash to add this repository
apt install -y linux-headers-4.4.0-104-generic
3. The above result shows that you need to install dbgsym, which is not in the existing sources. Run the following command to add the third-party source:
 sudo apt-key adv --keyserver keyserver.ubuntu.com --recv-keys C8CAB6595FDFF622

 codename=$(lsb_release -c | awk  '{print $2}')
 sudo tee /etc/apt/sources.list.d/ddebs.list << EOF
 deb http://ddebs.ubuntu.com/ ${codename}      main restricted universe multiverse
 deb http://ddebs.ubuntu.com/ ${codename}-security main restricted universe multiverse
 deb http://ddebs.ubuntu.com/ ${codename}-updates  main restricted universe multiverse
 deb http://ddebs.ubuntu.com/ ${codename}-proposed main restricted universe multiverse
 EOF

 sudo apt-get update
4. Run the following command after adding the source:
stap-prep
The following is a sample result:
Please install linux-headers-4.4.0-104-generic
Please install linux-image-4.4.0-104-generic-dbgsym
5. Run the following command to install the prompted packages:
apt install -y linux-image-4.4.0-104-generic-dbgsym
apt install -y linux-headers-4.4.0-104-generic

CentOS

1. Run the following command to install SystemTap:
yum install -y systemtap
2. For the purpose of this article, we assume that debuginfo is not added. Add the following to /etc/yum.repos.d/CentOS-Debug.repo and save.
[debuginfo]
name=CentOS-$releasever - DebugInfo
baseurl=http://debuginfo.centos.org/$releasever/$basearch/
gpgcheck=0
enabled=1
protect=1
priority=1
3. Run the following command to check for dependencies and install them:
Note:
The following command installs kernel-debuginfo.
stap-prep
4. Run the following command to check if the node has multiple versions of kernel-devel installed:
rpm -qa | grep kernel-devel
The returned result is as follows:
kernel-devel-3.10.0-327.el7.x86_64
kernel-devel-3.10.0-514.26.2.el7.x86_64
kernel-devel-3.10.0-862.9.1.el7.x86_64
If there are multiple versions, keep the one that corresponds to the kernel version. For example, if the current kernel version is 3.10.0-862.9.1.el7.x86_64, delete all version except kernel-devel-3.10.0-862.9.1.el7.x86_64.
Note:
You can use uname -r to view the kernel version.
Make sure kernel-debuginfo and kernel-devel are both installed and their versions correspond to the kernel version.
rpm -e kernel-devel-3.10.0-327.el7.x86_64 kernel-devel-3.10.0-514.26.2.el7.x86_64

Problem Analysis

You can use SystemTap to monitor a process in order to troubleshoot pod issues. This is how it works:
1. SystemTap translates the script into C code and calls gcc to compile the code into the Linux kernel module. It then uses modprobe to load the module into the kernel.
2. It uses the script to create kernel hooks and identify the causes of pod issues using the signals captured by the hooks.

Troubleshooting

Step 1: obtain the pids of the containers that restarted automatically in the pod due to exceptions

1. Run the following command to obtain the Container ID:
kubectl describe pod <pod name>
The returned result is as follows:
......
Container ID:  docker://5fb8adf9ee62afc6d3f6f3d9590041818750b392dff015d7091eaaf99cf1c945
......
Last State:     Terminated
 Reason:       Error
 Exit Code:    137
 Started:      Thu, 05 Sep 2019 19:22:30 +0800
 Finished:     Thu, 05 Sep 2019 19:33:44 +0800
2. 
Run the following command to query the pid of the main container process using the obtained Container ID:

docker inspect -f "{{.State.Pid}}" 5fb8adf9ee62afc6d3f6f3d9590041818750b392dff015d7091eaaf99cf1c945
The returned result is as follows:
7942

Step 2: narrow the scope using the container exit code

Use the Exit Code in the result of Step 1 to obtain the status code of the last container exit. For the purpose of this article, we will use 137 as an example. The analysis is as follows:
If the process was killed by an external signal, the exit code should be between 129 and 255.
An exit code of 137 indicates that the process was killed by SIGKILL. However, we still cannot determine the reason why the process exited.

Step 3: use the SystemTap script to identify the reason

Assuming the issue is reproducible, you can use a SystemTap to troubleshoot the problem.
1. Create a file called sg.stp. Add the following content and save.
global target_pid = 7942
probe signal.send{
 if (sig_pid == target_pid) {
     printf("%s(%d) send %s to %s(%d)\\n", execname(), pid(), sig_name, pid_name, sig_pid);
     printf("parent of sender: %s(%d)\\n", pexecname(), ppid())
     printf("task_ancestry:%s\\n", task_ancestry(pid2task(pid()), 1));
 }
}
Note:
Substitute pid with the value of the main container process pid obtained in Step 2. For the purpose of this article, we will use 7942 as an example:
2. Run the following command to execute the script:
stap sg.stp
When the container process is killed, the script captures the event and outputs the following:
pkill(23549) send SIGKILL to server(7942)
parent of sender: bash(23495)
task_ancestry:swapper/0(0m0.000000000s)=>systemd(0m0.080000000s)=>vGhyM0(19491m2.579563677s)=>sh(33473m38.074571885s)=>bash(33473m38.077072025s)=>bash(33473m38.081028267s)=>bash(33475m4.817798337s)=>pkill(33475m5.202486630s)

Solution

By observing task_ancestry, you can see the parent processes of the stopped process. In the example above, you can see a strange process called vGhyM0. This usually indicates that there is a trojan in the system. Take the necessary steps to clean it so your containers can function properly.
다음 주제: TKE Serverless 클러스터 업데이트(2023년)

도움말 및 지원

문제 해결에 도움이 되었나요?

피드백