up vote 3 down vote favorite
1
share [g+] share [fb]

I have an instance which is stuck at this boot prompt (looking at the instance log from the management console): 

The disk drive for /mydisk is not ready yet or not present
Continue to wait; or Press S to skip mounting or M for manual recovery

I added an EBS volume, mounted it, then deleted the EBS volume and forgot to remove the entry from /etc/fstab and rebooted the machine. Now it prompts me that at boot time, and I am unable to SSH into the instance.

What should I do to resolve this issue?

link|improve this question
75% accept rate
feedback

1 Answer

up vote 6 down vote accepted

I don't know an 'easy' solution to your problem, but there is a somewhat convoluted (depending on your root device type) solution.

If your instance has an EBS root, the solution is simple enough: stop the instance, detach the EBS drive, launch another instance and attach the EBS volume to it. Edit your fstab file as needed, detach the EBS volume and re-attach it to the original instance.

If your instance has an S3-backed root (i.e. instance-store), the solution is a good bit harder. You can download the data from your AMI (using ec2-download-bundle), and then extract that data into a single file (using ec2-unbundle). You can then mount the image, make the necessary change, and rebundle the image (with ec2-bundle-vol, overriding the default volume to bundle with the -v flag). That should give you a new, viable AMI, identical to your old one, except for the modifications you make. Alternatively, you could copy the content of the image file to an EBS volume using dd - and then convert to an EBS root backed instance.

link|improve this answer
Thankfully, my instance has an EBS root! Thanks a lot for the answer. I will try it out and confirm that it works. – ErJab Oct 15 '11 at 4:15
1  
ErJab: I wrote an article describing how to do this with EBS boot: alestic.com/2011/02/ec2-fix-ebs-root – Eric Hammond Oct 15 '11 at 20:18
Would it be prudent to also add nofail to /etc/fstab in case you make a typo? Maybe not, maybe "errors" (as described in fstab(5)) don't hold up the boot process. – Adam Monsen Feb 10 at 22:15
@AdamMonsen: that is good practice for some drives - e.g. ephemeral drives, since the boot process will exit to recovery mode if a disk is unavailable. I usually use the option nobootwait - haven't tried with nofail - I believe there are some subtle differences between them. – cyberx86 Feb 10 at 23:47
feedback

Your Answer

 
or
required, but never shown
discard

Not the answer you're looking for? Browse other questions tagged or ask your own question.