Karmic image stuck in "Waiting for EC2"

[Gustavo Niemeyer]

A Karmic server image in EC2 just got stuck in the following message during boot:

    * Waiting for EC2 meta-data service

Rebooting it didn't help. The machine successfully rebooted, but then was holding in the same position again.

Some details:

AMI: ami-1515f67c
Zone: us-east-1c
Type: m1.small
Lifecycle: spot
Kernel/Ramdisk: The default ones for the image

Note that the instance was started as a consequence of a spot request, but that shouldn't play an important role in theory, since once the request is satisfied the instance is pretty much a normal one. Scott has successfully booted a machine under the same circumstances and didn't face the same problem, proving that the issue is at least intermittent, even if related to spot requests.

The console output is attached.

[Scott Moser (smoser)]

Marking this as 'invalid' for linux-ec2. This is well out of kernel (user space messages in the log). Unless there was a kernel network driver issue, which I certainly wouldn't expect here.

[Scott Moser (smoser)]

Just for the record, Gustavo started instance and well over 40 minutes later captured logs and rebooted. This *should* have been enough time for ec2-wait-for-meta-data-service to timeout.