Make 'test.sh' functional

[dashohoxha]

The function of test.sh is not just to create and destroy a virtual interface. It should also start and stop the ltsp-server, the dhcp-server and the ltsp-client.

[d78ui98]

based on above comment and testing plan These are the changes that I plan on making in test.sh

first I set DEVICE=ltsp-test-if

then this these changes

case $1 in
    start )

        echo "creating virtual interface.."
        modprobe dummy
        ip link add ${DEVICE} type dummy
        ip link set ${DEVICE} up
        ip addr add ${NETWORK}.100/24 brd + dev ${DEVICE}
        sed -i settings.sh -e "/^LAN_IF/ c LAN_IF=\"${DEVICE}\"" 

        echo "destroying already present ltsp-server.."
        vagrant destroy -f
        (cd dhcp/ && vagrant destroy -f)
        if [ ${STANDALONE,,} != "yes" ] ; then
            echo "starting dhcp server.."
            (cd dhcp/ && vagrant up)
        fi

        echo "starting ltsp-server and ltsp-client"
        vagrant up && ./client.sh

        ;;  
    stop )

        echo "destroying virtual interface.."
        ip addr del ${NETWORK}.100/24 brd + dev ${DEVICE}
        ip link delete ${DEVICE} type dummy
        sed -i settings.sh -e '/^LAN_IF/ c LAN_IF="" '  
        rmmod dummy

        echo "destroying ltsp-client"
        VBoxManage controlvm ltsp-client poweroff 2>/dev/null
        VBoxManage unregistervm ltsp-client --delete 2>/dev/null

        echo "halt and destroying ltsp-server and dhcp-server"
        vagrant halt
        vagrant destroy -f
        (cd dhcp/ vagrant halt && vagrant destroy -f)
        ;;
    * )
        echo "error: invalid arguments provided"
        help
        ;;
esac 

vagrant destroy -f -f flag is used so that it does not prompt yes or no (cd dhcp/ && vagrant destroy -f) brackets are used so that the directory is not changed

[dashohoxha]

You can also keep DEVICE=ltsptest01.

You can set the option -x on #/bin/bash -x and then eliminate all those echo lines.

[dashohoxha]

(cd dhcp/ && vagrant destroy -f) brackets are used so that the directory is not changed\

You can also do:

    cd dhcp/
    vagrant destroy -f
    cd -

[d78ui98]

sure we can do that. I thought of using (cd dhcp/ && vagrant destroy -f) because it is a one liner

cd dhcp/ vagrant destroy -f cd -

we can do this as well. It looks cleaner. will push the changes after one final test.

[d78ui98]

Testing suggests we should remove this root check block

if [[ $UID != 0 ]]; then
    echo "error: use sudo or run script as root user"
    exit 1
fi

Since root is not a regular user in ubuntu. Meaning I can not log in from root. So it does not make sense to create vagrant boxes with root user. If we do it with root user they will then we wont be able to control them with normal user. this video showes the problem - https://youtu.be/GSpfOm4OfRo 2:06

will be using sudo whever required and vagrant box related commands will be executed with normal user.

[akash0x53]

This happened because you're running vagrant up with root hence the permission issues. You can run vagrant command with non-root user, with su <non-root user> -c command

[akash0x53]

You can also check linked-clones option to speed up vagrant testing.

https://www.vagrantup.com/docs/virtualbox/configuration.html#linked-clones

[d78ui98]

You can run vagrant command with non-root user, with su <non-root user> -c command

If we do this way then we would have to add su <non-root user> -c command to 14 lines. Also extra line to get non-root user from machine. However if we do it with sudo we will have to add sudo to just 7 lines. Using sudo is less number of characters. Making code more efficient. It is also familiar which increases readability of our code.

So it makes sense to use sudo :)