I am testing Fedora-Cloud-Base-27-20171004.n.0.ppc64le.qcow2 in a local environment. When I run the image the first time using a cdrom with user-data and meta-data, the VM start smoothly and I get the prompt quickly.
But when I run restart this image a second time without the cdrom, then I get a lot of timeouts because cloud-init try to reach a metadata local address that does not exist. This make the prompt very long to have.
On the first VM start I used:
/usr/bin/qemu-system-ppc64 -name cloud-host -enable-kvm -M pseries -smp 4 -m 4G -nographic -nodefaults -monitor stdio -serial pty -device virtio-net-pci,netdev=net10133,mac=c0:ff:ee:00:00:93 -netdev bridge,br=br0,id=net10133 -hda cloud.qcow2 -cdrom cloud01-cidata.iso -S
where cloud01-cidata.iso was created by "genisoimage" using my meta-data file and my user-data file. This work fine.
On the second start, I removed the cdrom as I think all info are already in the image now.
/usr/bin/qemu-system-ppc64 -name cloud-host -enable-kvm -M pseries -smp 4 -m 4G -nographic -nodefaults -monitor stdio -serial pty -device virtio-net-pci,netdev=net10133,mac=c0:ff:ee:00:00:93 -netdev bridge,br=br0,id=net10133 -hda cloud.qcow2 -S
But then I got several timeouts messages like:
Oct 05 12:55:22 atomic01.toulouse-stg.fr.ibm.com cloud-init: 2017-10-05 12:55:22,182 - url_helper.py[WARNING]: Calling 'http://169.254.169.254/2009-04-04/meta-data/instance-id' failed [50/120s]: request error [HTTPConnectionPool(host='169.254.169.254', port=80): Max retries exceeded with url: /2009-04-04/meta-data/instance-id (Caused by ConnectTimeoutError(<urllib3.connection.HTTPConnection object at 0x7fffb88a2a20>, 'Connection to 169.254.169.254 timed out. (connect timeout=50.0)'))]
This leads to several minutes init before getting the prompt.
Created attachment 1334883 [details]
Sorry for the noise, I forgot to create the file /etc/cloud/cloud-init.disabled to not run cloud-init on my second reboot without the cdrom. I don't have timeout problems anymore.