EC2实例，“常规错误挂载文件系统”

Question

I have a AWS micro instance running Ubuntu 12.04 LTS and last night when I SSH in, I did apt-get update and it gave me an error (I don't recall which). So I thought I would give my instance a reboot. This morning, it says that my instance has failed an Instance Sstaus Check and I am unable to SSH into it. The bottom of my system log is below. Is there any way to save this and if not, anyway to save the data?

Thank you!

Loading, please wait...
[35914369.823672] udevd[81]: starting version 175
Begin: Loading essential drivers ... done.
Begin: Running /scripts/init-premount ... done.
Begin: Mounting root file system ... Begin: Running /scripts/local-top ... done.
Begin: Running /scripts/local-premount ... done.
[35914370.187877] EXT4-fs (xvda1): mounted filesystem with ordered data mode. Opts: (null)
Begin: Running /scripts/local-bottom ... done.
done.
Begin: Running /scripts/init-bottom ... done.
[35914373.347844] init: mountall main process (183) terminated with status 1
General error mounting filesystems.
A maintenance shell will now be started.
CONTROL-D will terminate this shell and reboot the system.
Give root password for maintenance
(or type Control-D to continue): 

Answer 1

It depends on how badly broken the filesystem is.

You can start a new instance in AWS and then attach the EBS volume to your new instance. That may help you recover the data.

Don't terminate the instance, otherwise you could lose the filesystem completely.

Answer 2

Always use the "Create AMI" option of a running instance before doing an apt-get update/upgrade or yum update/upgrade. This way, if your system fails to come up after a reboot (after the update), you can spin up a 'before' version (ie bootable) instance using the AMI you just created.

In this case, Ubuntu probably tried to install a new kernel and/or ram file system (ramfs) and since this is an AWS virtual machine with kernel and ramfs dependencies, the standard Ubuntu build probably did not meet those dependencies and your virtual machine is now toast.

As mentioned, if you need to recover data from the unbootable system, mount it's EBS volume to a working system. It may complain that it is in use. If so, and to keep the EBS volume, you must check the option that lets you preserve the volume before you terminate it. The default on termination of an instance is to destroy it's EBS volume because, the assumption is, you booted from an EBS backed AMI that you previously (or regularly) created.