Ubuntu：“达到目标关闭”，可以安全地拔掉插头吗？

Question 1

尝试添加nolapic参数grub.cfg

跑步

sudo update-grub

https://help.ubuntu.com/community/BootOptions

帮助过我

在这里尝试了这些参数noapic nolapic noacpi irqpoll noapm nodma nomce

否则它没有帮助。解决了问题-重新编译initramfs：

sudo update-initramfs -u

Answer

尝试添加nolapic参数grub.cfg

跑步

sudo update-grub

https://help.ubuntu.com/community/BootOptions

帮助过我

在这里尝试了这些参数noapic nolapic noacpi irqpoll noapm nodma nomce

否则它没有帮助。解决了问题-重新编译initramfs：

sudo update-initramfs -u

Question 2

我认为这个问题与安装硬盘/固态硬盘有关。

我正在做的是，在 crontab 中我有一个 @reboot startup.shshell 脚本，该脚本由主用户（uid 1000）执行，并且该脚本安装 HDD（它曾经是 HDD，现在是 SSD）：

sudo mount /dev/sdb1 /media/ssd1

然后我所做的是创建一个shutdown.sh如下所示的脚本：

#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#       Ensure that this is desired
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
echo ""
echo "---- do you really want to shutdown/reboot? ----"
echo ""
read -n 1 -p "no, yes? (N/y) " ans;
case $ans in
  y|Y) printf "\n\nok, preparing for shutdown/reboot\n\n";;
    *) printf "\nok, exited\n\n";       exit;;
esac

#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#       DOCKER
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
echo ''
echo '--| stopping docker portainer |--';
echo ''
docker stop portainer

... stop more containers

#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#       Python
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
echo ''
echo '--| killing all python processes |--';
echo ''
sudo pkill python

#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#       MongoDB
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
echo ''
echo '--| exiting mongodb instances |--';
echo ''
mongod --dbpath /media/ssd1/data/mongodb/network/wiredTiger --shutdown

... stop more databases

echo ''
echo '--| deleting old mongodb logs |--';
echo ''
exec rm -rf {} \;"
sudo sh -c "find /media/ssd1/data/mongodb -type f -name 'log.txt.*' -exec du -ch {} + | sort -h"
sudo sh -c "find /media/ssd1/data/mongodb -type f -name 'log.txt.*' -exec rm -rf {} \;"

#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#       Sync
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
echo ''
echo '--| syncing discs |--';
sync

#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#       SSD
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
echo ''
echo '--| unmounting ssd1 and hdd5-1 |--';
echo ''
sudo umount /media/ssd1/

#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
#       Ask for next step
#~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
echo ""
echo "---- ok, done. what next? ----"
echo ""
read -n 1 -p "exit, reboot, or shutdown? (E/r/s) " ans;
case $ans in
  r|R) printf "\n\nok, rebooting\n\n";    sudo reboot </dev/null;;
  s|S) printf "\n\nok, powering off\n\n"; sudo poweroff </dev/null;;
    *) printf "\nok, exited\n\n";       exit;;
esac

这样我就可以关闭和重新启动系统。因此，可能与未明确卸载驱动器以及在关闭/重新启动之前未发出同步有关。

正如评论中提到的，我注意到一个容器 (RabbitMQ) 正在生成一个巨大的日志文件（增长到大约 40 GB）。也许同步此文件导致了麻烦，它试图在几乎断电的状态下同步它。但我不确定，因为我几乎从一开始就遇到了这个问题，而且据我所知，我最初并没有运行 RabbitMQ。

Answer