Lost node from previous deployment seen as bootstrap, but is not functional
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Fuel for OpenStack |
Won't Fix
|
Medium
|
Alexandr Notchenko |
Bug Description
iso 4.0-22 Havana
Precondition:
1. Deploy env with Ubuntu on KVM (1 controller + 1 compute + 2 ceph + rados) = Neutron gre - (deployment was not success)
2. Delete failed env
3. Wait while slave nodes was discovered after deletion
4. Try to deploy simple env on Centos (1controller/cinder + compute + Nova Flat DHCP)
5. Deployment hung on installation centos (1 node - centos was successfully intalled - second one stay at bootstrap)
ssh om admin node and execute command cobbler list:
[root@nailgun ~]# cobbler list
distros:
bootstrap
centos-x86_64
ubuntu_
profiles:
bootstrap
centos-x86_64
ubuntu_
systems:
default
node-8
node-9
repos:
images:
mgmtclasses:
packages:
ssh on node - 9
And see that hostname is node-4 and ubuntu is installed in it
root@nailgun ~]# ssh node-9
Warning: the RSA host key for 'node-9' differs from the key for the IP address '10.108.0.7'
Offending key for IP in /root/.
Matching host key in /root/.
Are you sure you want to continue connecting (yes/no)? yes
Welcome to Ubuntu 12.04 LTS (GNU/Linux 3.8.0-31-generic x86_64)
* Documentation: https:/
Last login: Mon Nov 11 15:30:18 2013 from 10.108.0.2
root@node-4:~#
node ip is:
eth0 Link encap:Ethernet HWaddr 64:0e:dd:b6:94:67
inet addr:10.108.0.7 Bcast:10.108.0.255 Mask:255.255.255.0
inet6 addr: fe80::660e:
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:22794 errors:0 dropped:3714 overruns:0 frame:0
TX packets:11426 errors:0 dropped:0 overruns:0 carrier:0
RX bytes:51768264 (51.7 MB) TX bytes:2875390 (2.8 MB)
Also there is error in nailgun on PUT for this node
13-11-11 13:41:22 ERROR (logger) Traceback (most recent call last):
2013-11-11 13:41:23 ERROR (logger) Response code '500 Internal Server Error' for PUT /api/nodes/ from 10.108.0.7:40075
2013-11-11 13:41:23 ERROR
(logger) Response code '500 Internal Server Error' for PUT /api/nodes/ from 10.108.0.7:40075
2013-11-11 13:41:22 ERROR
(logger) Traceback (most recent call last):
File "/opt/nailgun/
return self.handle()
File "/opt/nailgun/
return self._delegate(fn, self.fvars, args)
File "/opt/nailgun/
return handle_class(cls)
File "/opt/nailgun/
return tocall(*args)
File "<string>", line 2, in PUT
File "/opt/nailgun/
data = func(*args, **kwargs)
File "/opt/nailgun/
db().commit()
File "/opt/nailgun/
self.
File "/opt/nailgun/
self.
File "/opt/nailgun/
self.
File "/opt/nailgun/
self.
File "/opt/nailgun/
flush_
File "/opt/nailgun/
rec.
File "/opt/nailgun/
uow
File "/opt/nailgun/
mapper, table, update)
File "/opt/nailgun/
execute(
File "/opt/nailgun/
params)
File "/opt/nailgun/
compiled_sql, distilled_params
File "/opt/nailgun/
context)
File "/opt/nailgun/
context)
File "/opt/nailgun/
cursor.
IntegrityError: (IntegrityError) null value in column "mac" violates not-null constraint
'UPDATE nodes SET meta=%(meta)s, mac=%(mac)s, ip=%(ip)s WHERE nodes.id = %(nodes_id)s' {'nodes_id': 2, 'mac': None, 'meta': '{"system": {"fqdn": "node-2.
Seems that node was not deleted properly
Changed in fuel: | |
milestone: | none → 4.0 |
Changed in fuel: | |
assignee: | Dmitry Pyzhov (lux-place) → Alexandr Notchenko (anotchenko) |
Changed in fuel: | |
status: | New → Confirmed |
status: | Confirmed → Triaged |
Changed in fuel: | |
status: | Invalid → Won't Fix |
Node failed to reboot after cluster deletion. Later it was successfully discovered as a new node.
First, we should alert user about 'new' pre-deployed nodes.
Second, we should not believe that mcollective on such node is able to reboot it.
We need a design for this use case.