| *** tmorin has joined #openstack-net-bgpvpn | 08:12 | |
| *** doude has joined #openstack-net-bgpvpn | 08:54 | |
| doude | tmorin Hi | 09:07 |
|---|---|---|
| tmorin | salut doude | 09:09 |
| tmorin | tiens j'avais une question pour toi hier: tu utilises quoi comme devstack ? | 09:10 |
| doude | tmorin we can speak french here ? | 09:16 |
| tmorin | sorry, I thought this was a private chat window popping up :) | 09:22 |
| doude | :) about the bug https://bugs.launchpad.net/bgpvpn/+bug/1733905, I can propose a fix but do you think I have to propose a db migration as the bug was only merged on the master branch? | 09:35 |
| openstack | Launchpad bug 1733905 in networking-bgpvpn "SQL integer type is to small to store BGP LOCAL_PREF value" [Undecided,Confirmed] | 09:35 |
| tmorin | since its not deployed I don't think we need a migration | 09:57 |
| tmorin | a simple fix will be enough | 09:57 |
| doude | ok | 09:58 |
| doude | btw tmorin, I also pushed my final version for the support RC extension for OSC CLI in review https://review.openstack.org/#/c/520021/ | 10:00 |
| tmorin | doude: yes, I saw this this morning | 10:04 |
| tmorin | great | 10:04 |
| tmorin | doude: I won't have time to look at this before next week | 10:04 |
| tmorin | doude: but I had a quick look, I'm wondering why you introduce all the project_id/tenant_id "translation" in this change | 10:05 |
| tmorin | doude: perhaps we only have project_id in port_associations ? | 10:07 |
| tmorin | doude: I'm wondering if the change introduced guarantees that this python-neutronclient will work with earlier releases (before neutron started to always include both project_id and tenant_id) -- what you did is possibly ok, I just haven't looked closely yet | 10:09 |
| doude | tmorin no I realize that if the resource have both 'tenant_id' and 'project_id' field, when you do a show on that resource you have both field in the command output | 10:09 |
| doude | good question | 10:10 |
| doude | I think it should work | 10:10 |
| doude | btw, why bgpvpn does not show the project ID field to non admin users? | 10:10 |
| tmorin | doude: shouldn't we just do the same as done elsewhere in python-neutronclient ? we need a good reason to do something specific to BGPVPN, right ? | 10:11 |
| *** matrohon has joined #openstack-net-bgpvpn | 10:11 | |
| tmorin | doude: I don't really know, we don't need to expose it, since the users only see their resources anyways | 10:11 |
| doude | https://github.com/openstack/networking-bgpvpn/blob/master/etc/neutron/policy.d/bgpvpn.conf#L8 | 10:11 |
| tmorin | we don't have any RBAC on BGPVPN resources at this point | 10:12 |
| doude | I look other neutron resource, and project_id is not hidden | 10:13 |
| doude | ie. http://paste.openstack.org/show/627203/ | 10:15 |
| tmorin | yes, true | 10:19 |
| tmorin | it seems we could let access to this field | 10:19 |
| *** bfernando has joined #openstack-net-bgpvpn | 11:59 | |
| doude | tmorin around? | 16:03 |
| tmorin | yes | 16:03 |
| tmorin | more or less | 16:03 |
| doude | just to say I've just have a look to the bug https://bugs.launchpad.net/bgpvpn/+bug/1733905 and I'm not able to reproduce it with unit tests framewark | 16:05 |
| openstack | Launchpad bug 1733905 in networking-bgpvpn "SQL integer type is to small to store BGP LOCAL_PREF value" [Undecided,Confirmed] | 16:05 |
| doude | it seems UT DB fixture does not hit the problem | 16:05 |
| doude | ok tmorin, after quick search it seems the UT framework uses a sqlite in memory data base (you can force to use a file thanks to flag OS_TEST_DBAPI_ADMIN_CONNECTION see https://github.com/openstack/neutron/blob/master/TESTING.rst#unit-tests) and sqlite3 integer is store on 8 octets so 2^64-1 | 16:36 |
| tmorin | interesting | 16:37 |
| doude | that seems hard to propose a test for that issue, is it an issue if I propose a patch without UT? | 16:37 |
| tmorin | no its ok | 16:38 |
| doude | ok let's go | 16:38 |
| tmorin | but we can keep the bug open | 16:38 |
| doude | which bug? | 16:38 |
| tmorin | with a note saying that the API tempest test for routes-control should have test with a local_pref > 2**31 | 16:38 |
| tmorin | you can cc lukasz rajewski to the LP bug, he will work on these tests | 16:39 |
| *** tmorin has quit IRC | 16:58 | |
| *** bfernando has quit IRC | 17:08 | |
| *** matrohon has quit IRC | 17:29 | |
| openstackgerrit | Édouard Thuleau proposed openstack/networking-bgpvpn master: Use SQL BigInteger type to store BGP LOCAL_PREF https://review.openstack.org/522631 | 18:29 |
| *** tmorin has joined #openstack-net-bgpvpn | 20:00 | |
| *** tmorin has quit IRC | 22:05 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!