Instance starting but non reachable: check 1/2 fails

Question

Hi, we are experiencing a big issue: after trying to make the VM backup of 3 instances of our tenant, using the AWS docs procedure. At one moment we need to stop the export of running instance on the backup S3 bucket and delete the instance. We restore the two instance form the snapshots and one is good, the other one is not function as well: it start, but after 7 min gave back the unreacheable message and fail 1/2 checks.
I f I go on the Log console I could see that it can't find a boot device connected, but the disk and the volume are mounted and operating. The mounting point are set on dev/xvda as in the past.
We control every thing but nothing seem to be wrong.
We must create a new one and try to rebuild the content but we need to restart the oldest.
At the same time, we think that we could use the staging server that is on t2.micro instance and we try to promote at t3.xlarge but, when we tried, the instance doesn't start and fail check 1/2.
Also if we try to demote the actual prod. server from t3.xlarge to t2.micro.
One a firtst time I think that the problem must be related to a configuration or so on tha block the access to the instance and related resources, but I cannot find the solution. At my eyes seems to be a problem on allocating resources or something like that. But I cannot find anything different and we don't touch any configuration from the tried-backup.
Hope on you, some ideas or advice to solve as fast as possibile?
All the best!

Leo K · Answer

In general terms, "can't find a boot device" means, when you've already checked that the correct drive has the proper device name, as you did, that the disk is corrupted or missing data, such that it either cannot be mounted at all or doesn't have the contents needed even to try to start the operating system. You mentioned that the disks were mounted, but I suspect you mean that the EBS volumes are attached to the EC2 instance at the AWS level, but I don't see how they could be mounted at the operating system level, if BIOS/UEFI can't even find its boot device to try to start from.

You could create a temporary EC2 instance in the same AZ where the problematic EBS volume resides. Make sure you can log on to the new instance. Then attach the problematic EBS volume to the new instance. You can try mounting it there and see if any of its contents are accessible. With luck, you might be able at least to recover the data, configurations, etc. you need, even if the original server can no longer be restored to life. Given your microscopic instance sizes, I expect you're using Linux, so if the disk contents are mostly healthy but the OS just won't boot, you could also try copying most of the programs and data directly to a new server's disk and see if you can get your application/system to work on the new server.

Instance starting but non reachable: check 1/2 fails

Relevant content