you need load the kernel first

[lixiaolong]

upload image:
# wget https://dl.fedoraproject.org/pub/alt/atomic/stable/Fedora-Atomic-27-20180326.1/CloudImages/x86_64/images/Fedora-Atomic-27-20180326.1.x86_64.qcow2
# openstack image create --public --disk-format=qcow2 --container-format=bare --file=Fedora-Atomic-27-20180326.1.x86_64.qcow2 --property os_distro='fedora-atomic' fedora-atomic-latest

create template:
# openstack coe cluster template create swarm-cluster-template --image fedora-atomic-latest --external-network ext-net --fixed-network=demo-net --fixed-subnet=demo-subnet --floating-ip-disabled --dns-nameserver 8.8.8.8 --master-flavor 2 --flavor 2 --coe swarm --keypair mykey

create cluster:
# openstack coe cluster create swarm-cluster --cluster-template swarm-cluster-template --master-count 1 --node-count 1 --keypair mykey --timeout 10

get vnc:
# nova get-vnc-console xxxx novnc

in the vnc console,I found:
ERROR: cannot allocate kernel buffer
ERROR: you need load the kernel first

[lixiaolong (kong62)]

Has been resolved：

my zunclient version is 1.0.0
# pip uninstall python-zunclient
# pip install python-zunclient==1.2.0

[lixiaolong (kong62)]

I'm sorry, #1 I reply wrong. That's my bug for another one

[cally725 (christian-ally)]

I see the same problem with a CentOS-Generic qcow2 image have you been able to fix your problem ?

in the vnc console,I found:
 ERROR: cannot allocate kernel buffer
 ERROR: you need load the kernel first

[Sheel Rana (ranasheel2000)]

Hi cally725,

I too faced same issue.

Major reason for this is the RAM allocation to instance.
Kindly check how much ram is used in instance Flavor.

just a quick hint, if this is related:
  Normally while creating flavor, we specify "--ram 4", but it is not 4 GB rather 4 MB.