deploy cinder,glance with ceph failed when has no monitor node on kolla host
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
kolla |
Won't Fix
|
Undecided
|
jackning |
Bug Description
kolla version: stable/mitaka
First, i deployed ceph success.
Then, i have deployed cinder,glance using ceph, but there has two results.
One: if there has one monitor node on kolla host, deploy succeed.
Two: if there has no monitor node on kolla host, they are separately, deploy failed, the relative task skipping.
for example:
TASK: [glance | Creating the Glance service and endpoint] *******
skipping: [ceph1] => (item={'interface': 'admin', 'url': u'http://
skipping: [ceph1] => (item={'interface': 'internal', 'url': u'http://
skipping: [ceph1] => (item={'interface': 'public', 'url': u'http://
......
......
TASK: [glance | Creating Glance database] *******
skipping: [ceph1]
TASK: [glance | Reading json from variable] *******
skipping: [ceph1]
fatal: [localhost] => One or more undefined variables: 'dict object' has no attribute 'stdout'
description: | updated |
description: | updated |
description: | updated |
description: | updated |
summary: |
- deploy cinder,glance with ceph failed when ceph monitor and kolla aren't - on same node. + deploy cinder,glance with ceph failed when has no monitor node on kolla + host |
Changed in kolla: | |
assignee: | nobody → jackning (ningyougang) |
which kolla version are u using?
and i do not think this could be happen.
if will be helpful if u can provide more log with `-vvv` parameter and you inventory file.