Can't found network on instance launch

Can you show whats in the networks table of the ovs_quantum database?

echo 'select * from networks;' | mysql -u <user> -p ovs_quantum

Hi,

There : https://gist.github.com/2431153. I also include nova networks and melange ip_blocks (all have infos with the related uuid)

Thanks

--- Nicolas

It looks like the project_id that it is using in validate_networks() (in nova/network/quantum/manager.py) isn't matching.. I think it should probably take the project_id from the networks table. I'll look more closely at it tonight and send you a patch if you're willing to try it out.

Yes, send me a patch and I try it. I want to help to get quantum working so I willing to try other patch if necessary.

OK.. this isn't the right fix long term (we're moving away from relying on the database) but it should work for your case for right now:

diff --git a/nova/network/quantum/manager.py b/nova/network/quantum/manager.py
index 3337094..71eea23 100644
--- a/nova/network/quantum/manager.py
+++ b/nova/network/quantum/manager.py
@@ -647,8 +647,9 @@ class QuantumManager(manager.FloatingIP, manager.FlatManager
         if networks is None:
             return

- project_id = context.project_id
         for (net_id, _i) in networks:
+ net = db.network_get_by_uuid(context.elevated(), net_id)
+ project_id = net['project_id']
             # TODO(bgh): At some point we should figure out whether or
             # not we want the verify_subnet_exists call to be optional.
             if not self.ipam.verify_subnet_exists(context, project_id,

Sorry, launchpad screwed up the indentation.. the two '+' lines should be indented four spaces in from the "for" line.

Hi,

With the patch, the VM start, but failed in "networking" state. Log of nova-compute here : https://gist.github.com/2473287

OK, new patch to try. If this one doesn't solve it then I'll need to spend some time testing it internally with melange. Thanks for trying this out!

diff --git a/nova/network/quantum/manager.py b/nova/network/quantum/manager.py
index 3337094..59aa69e 100644
--- a/nova/network/quantum/manager.py
+++ b/nova/network/quantum/manager.py
@@ -647,12 +647,21 @@ class QuantumManager(manager.FloatingIP, manager.FlatManag
         if networks is None:
             return

- project_id = context.project_id
         for (net_id, _i) in networks:
+ net = db.network_get_by_uuid(context.elevated(), net_id)
             # TODO(bgh): At some point we should figure out whether or
             # not we want the verify_subnet_exists call to be optional.
- if not self.ipam.verify_subnet_exists(context, project_id,
- net_id):
+ exists = False
+ # Try both the network project id and the context project id
+ for x in [net['project_id'], context.project_id]:
+ project_id = x
+ try:
+ if self.ipam.verify_subnet_exists(context, project_id,
+ net_id):
+ exists = True
+ except:
+ pass
+ if not exists:
                 raise exception.NetworkNotFound(network_id=net_id)
             is_tenant_net = self.q_conn.network_exists(project_id, net_id)
             is_provider_net = self.q_conn.network_exists(

Sadly, still the same error. The log of compute here : https://gist.github.com/2473802. (if you need other log, ask ! I'm also on irc with pseudo nicolas2b, for "real-time" debugging)

This question was expired because it remained in the 'Open' state without activity for the last 15 days.