mount process should exit on volume destroy and reformat.

[zhoucheng361]

What happened:
https://github.com/juicedata/juicefs/actions/runs/8180665344/job/22369132377
Test Code:

    for i in {1..3}; do 
        echo "round $i"
        ./juicefs format $META_URL volume-$i
        ./juicefs mount -d $META_URL /tmp/myjfs$i_$j --no-usage-report
        cd /tmp/myjfs$i_$j
        bash -c 'for k in {1..300}; do echo abc>$k; sleep 0.2; done' || true & 
        cd -
        sleep 3
        uuid=$(./juicefs status $META_URL | grep UUID | cut -d '"' -f 4) 
        ./juicefs destroy --force $META_URL $uuid
        ./juicefs format $META_URL new-volume-$i 
        sleep 15   
        ps -ef | grep juicefs
        pidof juicefs && exit 1
        uuid=$(./juicefs status $META_URL | grep UUID | cut -d '"' -f 4) 
        ./juicefs destroy --force $META_URL $uuid
    done

Log:

runner      4332       1  0 00:20 ?        00:00:00 ./juicefs mount -d redis://localhost/1 /tmp/myjfs --no-usage-report
runner      4470    4332  1 00:20 ?        00:00:00 /home/runner/work/juicefs/juicefs/juicefs mount -d redis://localhost/1 /tmp/myjfs --no-usage-report
runner      4559    4305  0 00:20 ?        00:00:00 grep juicefs
4470 4332

What you expected to happen:

How to reproduce it (as minimally and precisely as possible):

Anything else we need to know?

Environment:

• JuiceFS version (use juicefs --version) or Hadoop Java SDK version:
• Cloud provider or hardware configuration running JuiceFS:
• OS (e.g cat /etc/os-release):
• Kernel (e.g. uname -a):
• Object storage (cloud provider and region, or self maintained):
• Metadata engine info (version, cloud provider managed or self maintained):
• Network connectivity (JuiceFS to metadata engine, JuiceFS to object storage):
• Others: