Interaction between nova and melange : ip fixed not found

asked 2012-02-28 09:54:02 -0500

Hi all,

I use the trunk version of Nova, Quantum (with the OVS plugin) and Melange. I created networks, everything seems to be right.

I have two questions : - the first VM I boot takes always a wrong IP address (for instance 172.16.0.0). However, when I boot a second VM, this one takes a good IP (for instance 172.16.0.2). Do you know why this can happened ? - I have an error regarding an fixed IP not found. Effectively, when I check the nova database, the fixed_ip table is empty but as I am using quantum and melange and their tables seems to be nicely filled. Do you have an idea about this issue ? This is a copy/paste of the error: 2012-02-28 10:45:53 DEBUG nova.rpc.common [-] received {u'_context_roles': [u'admin'], u'_context_request_id': u'req-461788a6-3570-4fa9-8620-6705eb69243c', u│·· '_context_read_deleted': u'no', u'args': {u'address': u'172.16.0.2'}, u'_context_auth_token': None, u'_context_strategy': u'noauth', u'_context_is_admin': Tr│·· ue, u'_context_project_id': None, u'_context_timestamp': u'2012-02-28T09:45:53.484445', u'_context_user_id': None, u'method': u'lease_fixed_ip', u'_context_r│·· emote_address': None} from (pid=8844) _safe_log /usr/local/src/nova/nova/rpc/common.py:144 │·· 2012-02-28 10:45:53 DEBUG nova.rpc.common [req-461788a6-3570-4fa9-8620-6705eb69243c None None] unpacked context: {'request_id': u'req-461788a6-3570-4fa9-8620│·· -6705eb69243c', 'user_id': None, 'roles': [u'admin'], 'timestamp': '2012-02-28T09:45:53.484445', 'is_admin': True, 'auth_token': None, 'project_id': None, 'r│·· emote_address': None, 'read_deleted': u'no', 'strategy': u'noauth'} from (pid=8844) unpack_context /usr/local/src/nova/nova/rpc/amqp.py:187 │·· 2012-02-28 10:45:53 DEBUG nova.network.manager [req-461788a6-3570-4fa9-8620-6705eb69243c None None] Leased IP |172.16.0.2| from (pid=8844) lease_fixed_ip /us│··r/local/src/nova/nova/network/manager.py:1186 │·· 2012-02-28 10:45:53 ERROR nova.rpc.common [-] Exception during message handling │··(nova.rpc.common): TRACE: Traceback (most recent call last): │·· (nova.rpc.common): TRACE: File "/usr/local/src/nova/nova/rpc/amqp.py", line 250, in _process_data │··(nova.rpc.common): TRACE: rval = node_func(context=ctxt, *node_args) │··(nova.rpc.common): TRACE: File "/usr/local/src/nova/nova/network/manager.py", line 1187, in lease_fixed_ip │··(nova.rpc.common): TRACE: fixed_ip = self.db.fixed_ip_get_by_address(context, address) │·· (nova.rpc.common): TRACE: File "/usr/local/src/nova/nova/db/api.py", line 473, in fixed_ip_get_by_address │··(nova.rpc.common): TRACE: return IMPL.fixed_ip_get_by_address(context, address) │··(nova.rpc.common): TRACE: File "/usr/local/src/nova/nova/db/sqlalchemy/api.py", line 119, in wrapper │·· (nova.rpc.common): TRACE: return f(args, **kwargs) │··(nova.rpc.common): TRACE: File "/usr/local/src/nova/nova/db/sqlalchemy/api.py", line 1131, in fixed_ip_get_by_address │·· (nova.rpc.common): TRACE: raise exception.FixedIpNotFoundForAddress(address=address) │·· (nova.rpc.common): TRACE: FixedIpNotFoundForAddress: Fixed ip not found for address 172.16.0.2. │·· (nova.rpc.common): TRACE:

Thank you very much for your help.

Regards, Jérôme

edit retag flag offensive close merge delete

4 answers

Sort by » oldest newest most voted
0

answered 2012-02-28 10:54:57 -0500

I have just posted this question on the netstack and openstack mailing list.

Jérôme

edit flag offensive delete link more
0

answered 2012-03-02 08:30:11 -0500

Hi,

Here is the full thread : http://www.mail-archive.com/openstack@lists.launchpad.net/msg07974.html (http://www.mail-archive.com/openstack...)

Thanks for your help.

Jérôme

edit flag offensive delete link more
0

answered 2012-02-28 16:13:25 -0500

Here is a dirty fix for the problem mentionned aboved. http://paste.openstack.org/show/5326/

edit flag offensive delete link more
0

answered 2012-02-29 10:08:33 -0500

A solution for question number one (the first VM I boot takes always a wrong IP address (for instance 172.16.0.0). However, when I boot a second VM, this one takes a good IP (for instance 172.16.0.2). Do you know why this can happened ?). This solution is copied/pasted from the openstack/netstack mailing list.

Thanks to Jason!

The default melange policy allows assignment of the network address and synthesise a gateway address (if it is not specified). It will not hand out the gateway address. The "fix" is to create an ip policy that restricts the octect 0. I think the syntax is something like

melange policy create -t {tennant} name={block_name} desc={policy_name} (This should return the policy_id for the next command)

melange unusable_ip_octet create -t {tennant} policy_id={policy_id} octect=0

melange ip_block update -t {tennant} id={block_id} policy_id={policy_id}

edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower

Stats

Asked: 2012-02-28 09:54:02 -0500

Seen: 44 times

Last updated: Mar 02 '12