*** jmlowe has quit IRC | 00:15 | |
*** jmlowe has joined #openstack-containers | 00:16 | |
*** jmlowe has quit IRC | 00:27 | |
*** jmlowe has joined #openstack-containers | 00:37 | |
*** jmlowe has quit IRC | 00:52 | |
*** ricolin has quit IRC | 01:32 | |
*** hongbin has joined #openstack-containers | 03:19 | |
*** ricolin has joined #openstack-containers | 03:36 | |
*** ykarel|away is now known as ykarel | 03:52 | |
*** ricolin has quit IRC | 04:31 | |
*** udesale has joined #openstack-containers | 04:45 | |
*** hongbin has quit IRC | 05:04 | |
*** ramishra has joined #openstack-containers | 05:05 | |
*** udesale has quit IRC | 05:31 | |
*** udesale has joined #openstack-containers | 05:32 | |
*** elenalindq has joined #openstack-containers | 06:12 | |
*** udesale has quit IRC | 06:23 | |
*** udesale has joined #openstack-containers | 06:25 | |
*** lpetrut has joined #openstack-containers | 07:09 | |
*** lpetrut has quit IRC | 07:10 | |
*** lpetrut has joined #openstack-containers | 07:11 | |
jrosser | cosmicsound: is your master node trying to connect to the internal rather than public keystone endpoint? | 07:36 |
---|---|---|
*** ykarel is now known as ykarel|lunch | 07:41 | |
*** pcaruana has joined #openstack-containers | 08:20 | |
*** ykarel|lunch is now known as ykarel | 08:46 | |
*** udesale_ has joined #openstack-containers | 08:49 | |
*** udesale has quit IRC | 08:52 | |
*** ricolin has joined #openstack-containers | 08:56 | |
brtknr | jrosser: +1 | 09:32 |
*** yolanda__ has joined #openstack-containers | 09:37 | |
*** yolanda has quit IRC | 09:40 | |
cosmicsound | jrosser , seems so | 09:40 |
jrosser | for any real deployment you need to ensure server_keystone_endpoint_type=public in heat.conf | 09:40 |
*** flwang1 has joined #openstack-containers | 09:42 | |
flwang1 | brtknr: ping | 09:42 |
brtknr | flwang1: pong | 09:42 |
flwang1 | brtknr: https://review.opendev.org/709419 | 09:42 |
brtknr | saw your heat patch, looks great! | 09:42 |
flwang1 | brtknr: thanks for your pushing, i'm trying to fix it from the heat side ;) | 09:42 |
brtknr | flwang1: definitely the best thing to do, given that heat does not allow null variables | 09:43 |
cosmicsound | jrosser , got this | 09:43 |
cosmicsound | region_name_for_services = RegionOne | 09:44 |
cosmicsound | server_keystone_endpoint_type = public | 09:44 |
cosmicsound | region_name = fr-rbx1 in heat.conf | 09:44 |
cosmicsound | [clients] | 09:44 |
cosmicsound | endpoint_type = internalURL | 09:44 |
brtknr | flwang1: are the heat folks open to it? | 09:45 |
flwang1 | brtknr: i had a chat with the Heat PTL, he is ok with the patch | 09:45 |
flwang1 | i just need more test and add unit test | 09:45 |
brtknr | will they will happy to backport to train? | 09:45 |
jrosser | cosmicsound: the logic with server_keystone_endpoint_type was that if it is defined it will use the service catalog to determine the endpoint url | 09:46 |
jrosser | so i would double check that your service catalog is right for the public keystone endpoint | 09:46 |
cosmicsound | I got this in cluster | 09:46 |
cosmicsound | https://mdb.uhlhost.net/uploads/3199a0939e9be44b/image.png | 09:47 |
cosmicsound | If I do not add the private-net and subnet the cluster gets a public ip | 09:47 |
cosmicsound | If I add them it does not | 09:47 |
brtknr | cosmicsound: do you have a router configured to act as a bridge between the public and the private-net? | 09:48 |
brtknr | if its not specified, magnum automatically creates a router | 09:49 |
brtknr | and the network | 09:49 |
cosmicsound | I have the default init-runtime | 09:49 |
flwang1 | brtknr: i'm not sure, but i think it should be ok, since it's a quite minor change | 09:49 |
cosmicsound | https://mdb.uhlhost.net/uploads/f52c375eae297867/image.png these are the only actions stuck, and they will remain so until it times out | 09:50 |
brtknr | flwang1: i found another bug | 09:50 |
cosmicsound | https://mdb.uhlhost.net/uploads/821b613a4e99d09b/image.png while this runs perfectly | 09:51 |
flwang1 | what? | 09:51 |
brtknr | did you see https://storyboard.openstack.org/#!/story/2007271 | 09:51 |
cosmicsound | and notice the public ip | 09:51 |
brtknr | flwang1: its quite major | 09:51 |
flwang1 | brtknr: that's a known issue | 09:51 |
cosmicsound | I can conect to machine via public ip, and it has internet access. | 09:52 |
brtknr | flwang1: is it? | 09:52 |
flwang1 | we had tried to fix it but we didn't work out an agreement about the design | 09:52 |
flwang1 | brtknr: https://review.opendev.org/#/c/621611/ | 09:52 |
brtknr | flwang1: how long has this problem been around? | 09:52 |
flwang1 | long time ago i would say | 09:53 |
cosmicsound | http://paste.openstack.org/show/789917/ from line 36 or so looks bad | 09:54 |
cosmicsound | https://mdb.uhlhost.net/uploads/1cfbee3ecc6a35a5/image.png my current network topology | 09:55 |
brtknr | flwang1: hmm i see | 09:55 |
brtknr | cosmicsound: hmm that is something odd, as you appear to have route to the public network | 09:57 |
brtknr | idk why your instance is not getting public ip | 09:57 |
brtknr | maybe check neutron logs? | 09:57 |
cosmicsound | Instance get public address | 09:57 |
brtknr | flwang1: i always assumed that the policy was merge | 09:58 |
flwang1 | brtknr: do you mean policy or labels? | 09:58 |
brtknr | i didnt realise all cluster template labels get dropped | 09:58 |
brtknr | if there are any labels on cluster | 09:59 |
flwang1 | let's get it fixed in this cycle | 09:59 |
cosmicsound | very weird is that Magnum API is not visible in horizon dashboard not sure this is normal behaviour | 09:59 |
cosmicsound | I tried several redeployments with kolla | 10:00 |
cosmicsound | Did not helped | 10:00 |
brtknr | flwang1: what kind of behaviour would you like to see? | 10:02 |
brtknr | merge or replace? | 10:03 |
brtknr | I'd prefer a simple merge policy, why would you go through the effort of specifying cluster template labels if they are just gonna get dropped | 10:06 |
cosmicsound | ./usr/local/lib/python3.8/site-packages/netaddr/strategy/__init__.py:189: SyntaxWarning: "is not" with a literal. Did you mean "!="? |. if word_sep is not '': | 10:09 |
brtknr | cosmicsound: can you reach keystone endpoint from the instance? | 10:12 |
cosmicsound | brtknr unfortunately no | 11:19 |
cosmicsound | i use a vlan for local kolla vip address | 11:20 |
cosmicsound | and we use openvswitch | 11:20 |
cosmicsound | will check why this happends | 11:20 |
*** udesale_ has quit IRC | 11:28 | |
jrosser | cosmicsound: is it your internal vip appearing in the logs in your master node? | 11:29 |
jrosser | if so that is your root cause | 11:29 |
cosmicsound | Yes it appears | 11:38 |
cosmicsound | we have at the moment only internal | 11:38 |
cosmicsound | while external vip fails completely to go online | 11:38 |
jrosser | the data inserted into the master node should reference the external vip, if it is in the service catalog | 11:41 |
*** pcaruana has quit IRC | 12:28 | |
*** pcaruana has joined #openstack-containers | 12:42 | |
*** dave-mccowan has joined #openstack-containers | 13:29 | |
*** dave-mccowan has quit IRC | 13:40 | |
*** dave-mccowan has joined #openstack-containers | 13:41 | |
*** ivve has joined #openstack-containers | 15:04 | |
*** ykarel is now known as ykarel|afk | 15:12 | |
*** dave-mccowan has quit IRC | 15:14 | |
*** ykarel|afk is now known as ykarel|away | 15:15 | |
*** lpetrut has quit IRC | 15:20 | |
*** dave-mccowan has joined #openstack-containers | 15:28 | |
*** dave-mccowan has quit IRC | 15:35 | |
*** jmlowe has joined #openstack-containers | 15:57 | |
*** ivve has quit IRC | 16:07 | |
*** jmlowe has quit IRC | 16:28 | |
*** jmlowe has joined #openstack-containers | 16:28 | |
*** jmlowe has quit IRC | 16:34 | |
*** jmlowe has joined #openstack-containers | 16:39 | |
*** pcaruana has quit IRC | 17:11 | |
*** pcaruana has joined #openstack-containers | 17:26 | |
*** jmlowe has quit IRC | 17:33 | |
*** jmlowe has joined #openstack-containers | 17:36 | |
*** jmlowe has quit IRC | 17:48 | |
*** jmlowe has joined #openstack-containers | 18:01 | |
*** jmlowe has quit IRC | 18:06 | |
*** jmlowe has joined #openstack-containers | 18:48 | |
*** jmlowe has quit IRC | 18:58 | |
*** jmlowe has joined #openstack-containers | 19:05 | |
*** pcaruana has quit IRC | 19:29 | |
*** jmlowe has quit IRC | 19:33 | |
*** jmlowe has joined #openstack-containers | 19:35 | |
*** flwang1 has quit IRC | 19:50 | |
*** jmlowe has quit IRC | 19:54 | |
*** jmlowe has joined #openstack-containers | 19:56 | |
*** elenalindq has quit IRC | 20:08 | |
*** elenalindq has joined #openstack-containers | 20:22 | |
*** jmlowe has quit IRC | 20:35 | |
*** jmlowe has joined #openstack-containers | 20:36 | |
*** jmlowe has quit IRC | 20:53 | |
*** jmlowe has joined #openstack-containers | 21:18 | |
*** jmlowe has quit IRC | 21:26 | |
*** jmlowe has joined #openstack-containers | 21:30 | |
*** elenalindq has quit IRC | 21:34 | |
*** jmlowe has quit IRC | 21:46 | |
*** jmlowe has joined #openstack-containers | 21:47 | |
cosmicsound | Anyone can point to a properly maintained magnum deployment docs? outside of openstack if possible some blog or such | 22:00 |
cosmicsound | https://mdb.uhlhost.net/uploads/cace6cdabe5c6550/image.png it seems the public network is not the way forward for this, need to find some solution to this. | 22:01 |
flwang | cosmicsound: what's your current problem? | 22:06 |
*** jmlowe has quit IRC | 22:14 | |
cosmicsound | is there a difference between overlay and overlay2 some templates were build with overlay2 . | 22:16 |
*** dave-mccowan has joined #openstack-containers | 22:22 | |
*** dave-mccowan has quit IRC | 22:27 | |
*** spiette has quit IRC | 22:28 | |
*** spiette has joined #openstack-containers | 22:31 | |
*** dave-mccowan has joined #openstack-containers | 22:35 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!