OpenStack Mitaka instance with nova-vmware (vsphere 6.0) hangs while booting at "... random: nonblocking pool is initialized"

[Larry Michel]

I have seen it recreated consistently with the Xenial vmdk cloud image from http://cloud-images.ubuntu.com/releases/xenial/release/.

I also attempted to qemu-convert the QCOW2 version to vmdk and issue was still reproducible. Note that the hangs eventually seems to eventually time out. But when I tried to access both a trusty instance (which I had previously sshed into) and xenial instance could not be reached through the floating IP.

With the Trusty instance, the vmdk was created by converting the QCOW2 image to vmdk using qemu-convert. It's not clear whether what other issue caused it to stop being reachable, but upon rebooting it, that instance also started to hit "random: nonblocking pool is initialized" consistently.

I then attempted to deploy another Trusty instance after powering off all other instances, and it was reachable through the floating IP.

I have attached the console screen shot as it hits the issue.

[Date Huang (tjjh89017)]

`console= tty1 console=ttyS0` boot options cause this result.
If you remove the options from `grub.cfg` and `/etc/default/grub.d/50-cloudimg-settings.cfg`
All will be fine.

[Alexander Hall (compuguy1088)]

I've seen this also appear on ESXi 6.5 U1, but I thought it was related to errors of the floppy image not mounting properly.

[David (davidjaquier)]

Have the same issue with cloudstack running on the top of esxi 6.5.

Is there a way to remove `console= tty1 console=ttyS0` from the grub configuration, from the OVA, without deploying a virtual machine in an other hypervisor?

I tried to mount the vmdk with guestmount on ubuntu or with fuse on mac, without success.