Cannot find /usr/local/tomcat/bin/setclasspath.sh
This file is needed to run this program

After trying, use the add command packaged with dockerfile

RUN unset CATALINA_HOME

Invalid after attempt

The prompt is that setclasspath.sh cannot be found, because this Tomcat restarts repeatedly

So use the command to copy a bin directory to the current folder

docker cp Docker_id:/usr/local/tomcat/bin ./

The result shows that setclasspath.sh exists, so where is the problem? Look for the error reporting code

In the catalina.sh script, the part causing the problem is as follows:

  if [ -r "$CATALINA_HOME"/bin/setclasspath.sh ]; then
    . "$CATALINA_HOME"/bin/setclasspath.sh
  else
    echo "Cannot find $CATALINA_HOME/bin/setclasspath.sh"
    echo "This file is needed to run this program"
  fi

The square brackets plus the -r command means to test whether the file is read-only. Similarly, – x tests whether the file is executable

In the problematic system, the – R command call in the container is abnormal.

Try to start a temporary tomcat8 authentication,

docker run -it --rm --entrypoint=/bin/bash tomcat:8

Note: executing the docker run command with the –rm command option is equivalent to executing docker rm -v after the container exits
execute command

root@f338debf92f6:/usr/local/tomcat# [[ -r /bin/bash ]]
root@f338debf92f6:/usr/local/tomcat# echo $?
1

Executed on a normal system

root@0083a80a9ec2:/usr/local/tomcat# [[ -r /bin/bash ]]
root@0083a80a9ec2:/usr/local/tomcat# echo $?
0

The command ‘$?’ indicates the exit status of the previous command or the return value of the function
for exit status, 0 indicates no error, and any other value indicates an error. In general, most commands return 0 upon successful execution and 1 upon failure. Some commands return other values, indicating different types of errors.

How to solve it
this is related to the faccessat2 system call. Due to the bug in runc, if your kernel does not support faccessat2, it will fail. There is an article saying that upgrading the kernel to 5.8 or above may work well, but I have tried it hard, because the kernel with the problem is 5.10
Method 1: update runc >= 1.0.0-rc93

Method 2: — privileged switch to run the container

Specific implementation of method 1:

View the original runc version

runc -v

Download version 1.0.0-rc95 of runc.amd64

Download address: releases · opencontainers/runc · GitHub

Name change and Execution Authority

mv runc.amd64 runc && chmod +x runc

View and back up the original runc

whereis runc

Some systems are in the /usr/bin/ directory, some systems are in the /usr/local/bin/ directory, the following commands are in the /usr/local/bin/ directory, the following commands are in the /usr/bin/ directory, change the directory yourself

mv /usr/local/bin/runc /usr/local/bin/runcbak

Overwrite the original runc

cp runc /usr/local/bin/runc

View the new version of runc

runc -v

Restart docker

systemctl restart docker

Method 2: concrete implementation

Docker run execution

docker run --privileged=true -p 8080:8080 tomcat:8

Add in docker-compose.yml

version: '2'
services:
    tomcat:
     container_name: tomcat
     image: tomcat:8
     ports:
         - '8080:8080'
     environment:
         - TZ=Asia/Shanghai
     privileged: true
     restart: always