deployment occasionally breaks with this error: 401 UNAUTHORIZED Nonce already used
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
MAAS |
Confirmed
|
Undecided
|
David Britton |
Bug Description
This happens periodically, but is very annoying when it does, since juju doesn't seem to retry it, and I can't do anything to fix it other than destroying the node. Back 6 months ago when I was using an older version of MAAS, I would get this Nonce error showing up during random juju
This is maas/juju-core from the UCA:
ubuntu@
ubuntu@
ubuntu@
environment: maas
machines:
"19":
agent-
gomaasapi: got error back from server: 401 UNAUTHORIZED (Nonce already used:
33410643))'
instance-id: pending
series: precise
services:
juju-gui:
charm: local:precise/
exposed: false
units:
juju-gui/0:
machine: "19"
ubuntu@
ubuntu@
ubuntu@
ubuntu@
ubuntu@
openstack-ha openstack-
ubuntu@
ubuntu@
error: invalid service name "juju-gui-2"
ubuntu@
ubuntu@
environment: maas
machines:
"20":
agent-state: pending
dns-name: yq8af.master
instance-id: /MAAS/api/
series: precise
services:
juju-gui2:
charm: local:precise/
exposed: false
units:
juju-gui2/0:
machine: "20"
ubuntu@
Here is a log of my interaction
ubuntu@
ubuntu@
environment: maas
machines:
"19":
agent-
gomaasapi: got error back from server: 401 UNAUTHORIZED (Nonce already used:
33410643))'
instance-id: pending
series: precise
services:
juju-gui:
charm: local:precise/
exposed: false
units:
juju-gui/0:
machine: "19"
ubuntu@
ubuntu@
environment: maas
machines:
"20":
agent-state: pending
dns-name: yq8af.master
instance-id: /MAAS/api/
series: precise
services:
juju-gui2:
charm: local:precise/
exposed: false
units:
juju-gui2/0:
machine: "20"
ubuntu@
assigned to me since we are still triaging.