| *** vilobhmm111 has quit IRC | 00:00 | |
| *** devkulkarni has quit IRC | 00:04 | |
| *** jerrygb_ has joined #openstack-containers | 00:05 | |
| *** jerrygb_ has quit IRC | 00:08 | |
| *** eghobo has joined #openstack-containers | 00:08 | |
| *** jerrygb has quit IRC | 00:09 | |
| *** Tango has quit IRC | 00:13 | |
| *** rebase33 has joined #openstack-containers | 00:16 | |
| *** Guest18166 is now known as dims_ | 00:17 | |
| *** rebase33 has quit IRC | 00:17 | |
| *** suro-patz has quit IRC | 00:20 | |
| *** ganeshna has quit IRC | 00:21 | |
| *** ganeshna has joined #openstack-containers | 00:21 | |
| *** devkulkarni has joined #openstack-containers | 00:27 | |
| *** dimtruck is now known as zz_dimtruck | 00:28 | |
| openstackgerrit | Eli Qiao proposed openstack/magnum: Fix container action debug log not accurate https://review.openstack.org/226194 | 00:29 |
|---|---|---|
| openstackgerrit | Eli Qiao proposed openstack/magnum: Functional: Split python client functional testing case https://review.openstack.org/223907 | 00:33 |
| openstackgerrit | Eli Qiao proposed openstack/magnum: Add functional test cases(containers actions) for swarm https://review.openstack.org/226125 | 00:33 |
| openstackgerrit | Eli Qiao proposed openstack/magnum: Add functional test cases for swarm https://review.openstack.org/221633 | 00:33 |
| *** julim has quit IRC | 00:39 | |
| *** mfalatic has quit IRC | 00:40 | |
| *** achanda has quit IRC | 00:41 | |
| *** devkulkarni has quit IRC | 00:46 | |
| *** vilobhmm11 has joined #openstack-containers | 00:47 | |
| *** suro-patz has joined #openstack-containers | 00:48 | |
| *** suro-patz1 has joined #openstack-containers | 00:49 | |
| *** suro-patz has quit IRC | 00:52 | |
| *** sthillma has quit IRC | 00:53 | |
| *** suro-patz1 has quit IRC | 01:03 | |
| *** banix has quit IRC | 01:09 | |
| *** banix has joined #openstack-containers | 01:10 | |
| *** zz_dimtruck is now known as dimtruck | 01:10 | |
| *** achanda has joined #openstack-containers | 01:11 | |
| *** puranamr has joined #openstack-containers | 01:15 | |
| *** suro-patz has joined #openstack-containers | 01:25 | |
| *** hongbin has joined #openstack-containers | 01:29 | |
| *** ganeshna has quit IRC | 01:36 | |
| *** suro-patz has quit IRC | 01:38 | |
| *** ganeshna has joined #openstack-containers | 01:39 | |
| *** shuquan has joined #openstack-containers | 01:39 | |
| *** puranamr has quit IRC | 01:40 | |
| *** jerrygb has joined #openstack-containers | 01:44 | |
| *** jerrygb has quit IRC | 01:45 | |
| *** jerrygb has joined #openstack-containers | 01:45 | |
| *** harshs has quit IRC | 01:53 | |
| *** shuquan has quit IRC | 01:54 | |
| *** kebray has joined #openstack-containers | 01:56 | |
| *** shuquan has joined #openstack-containers | 01:56 | |
| *** shuquan has quit IRC | 01:57 | |
| *** dims_ has quit IRC | 01:57 | |
| *** erkules_ has joined #openstack-containers | 01:57 | |
| *** dims_ has joined #openstack-containers | 01:57 | |
| *** shuquan has joined #openstack-containers | 01:58 | |
| *** erkules has quit IRC | 01:59 | |
| *** banix has quit IRC | 02:09 | |
| *** dims_ has quit IRC | 02:10 | |
| *** dims_ has joined #openstack-containers | 02:13 | |
| *** dims__ has joined #openstack-containers | 02:18 | |
| *** dims_ has quit IRC | 02:18 | |
| *** dims__ has quit IRC | 02:23 | |
| *** eghobo has quit IRC | 02:24 | |
| *** unicell has quit IRC | 02:33 | |
| *** madhuri has joined #openstack-containers | 02:37 | |
| *** ganeshna has quit IRC | 02:37 | |
| *** ganeshna has joined #openstack-containers | 02:40 | |
| madhuri | hongbin: Hi | 02:48 |
| madhuri | hongbin: http://paste.openstack.org/show/473668/ | 02:48 |
| eliqiao | hi, seems gate's broken | 02:48 |
| eliqiao | madhuri: does port 8080 open? | 02:49 |
| madhuri | eliqiao: It should be open | 02:50 |
| madhuri | Let me check | 02:50 |
| *** harshs has joined #openstack-containers | 02:51 | |
| madhuri | eliqiao: Not opened | 02:54 |
| madhuri | But I guess we don't have to manually add it | 02:54 |
| eliqiao | ya. | 02:54 |
| madhuri | Do you face the same issue on master? | 02:55 |
| madhuri | I will add a rule for it and then check | 02:55 |
| eliqiao | madhuri: I don't try kube yet. | 02:55 |
| madhuri | Ok I will try opening the port | 02:55 |
| eliqiao | currently, gate's broken, devstack can not be installed. | 02:55 |
| eliqiao | I see this error http://logs.openstack.org/94/226194/2/check/gate-functional-dsvm-magnum/7c5e4f7/logs/devstacklog.txt.gz#_2015-09-23_01_37_17_426 | 02:56 |
| eliqiao | create_barbican_accounts was called 2 times... | 02:59 |
| hongbin | madhuri: yo | 03:00 |
| madhuri | hongbin: Hi | 03:00 |
| madhuri | Please check the paster URL above | 03:00 |
| madhuri | eliqiao: hongbin 8080 is open in the secgroup-rules | 03:01 |
| madhuri | But actually seems close | 03:01 |
| hongbin | You are using master? | 03:01 |
| hongbin | and could you confirm the image you are using (it should be *-5) | 03:02 |
| madhuri | Yes I am using master | 03:02 |
| madhuri | fedora-atomic-5 | 03:03 |
| hongbin | It looks the kube-apiserver is not running | 03:03 |
| hongbin | does the etcd running? | 03:03 |
| madhuri | hongbin No it is not running | 03:05 |
| hongbin | Could you paste two things | 03:05 |
| hongbin | 1. The etcd log | 03:05 |
| madhuri | Yes sure | 03:05 |
| hongbin | 2. the etcd config file | 03:05 |
| madhuri | 2. http://paste.openstack.org/show/473669/ | 03:07 |
| madhuri | I can't find the etcd logs | 03:08 |
| hongbin | sudo journalctl -u etcd --no-pager | 03:09 |
| hongbin | (If I type everything correct ) | 03:09 |
| *** shuquan has quit IRC | 03:10 | |
| *** shuquan has joined #openstack-containers | 03:11 | |
| madhuri | hongbin: -- Logs begin at Tue 2015-09-22 17:22:16 UTC, end at Tue 2015-09-22 19:30:54 UTC. -- | 03:12 |
| madhuri | No logs I guess | 03:12 |
| hongbin | ...... | 03:12 |
| hongbin | how about restart the etcd | 03:12 |
| *** david-lyle has joined #openstack-containers | 03:13 | |
| madhuri | hongbin: even after starting etcd, kubectl fails | 03:16 |
| hongbin | but etcd succeed? | 03:16 |
| madhuri | Yes | 03:18 |
| hongbin | any luck by reboot the vm | 03:18 |
| madhuri | Let me try | 03:19 |
| madhuri | Have you tested fedora-atomic-5? | 03:19 |
| hongbin | Yes I do | 03:19 |
| hongbin | So as the gate | 03:20 |
| eliqiao | hongbin: I checked gate, and found barbican plugin were called 2 times | 03:20 |
| hongbin | eliqiao: not sure about that | 03:21 |
| eliqiao | #link http://logs.openstack.org/94/226194/2/check/gate-functional-dsvm-magnum/7c5e4f7/logs/devstacklog.txt.gz#_2015-09-23_01_23_45_342 | 03:21 |
| eliqiao | and it cause devstak failed to setup openstack. | 03:21 |
| hongbin | yuanying: do you have any idea for the two barbican, check eliqiao's link | 03:21 |
| eliqiao | yuanying: it leads this error http://logs.openstack.org/94/226194/2/check/gate-functional-dsvm-magnum/7c5e4f7/logs/devstacklog.txt.gz#_2015-09-23_01_37_17_426 | 03:22 |
| hongbin | madhuri: I expect all the processes go back after rebooting, does it? | 03:23 |
| madhuri | hongbin: after restarting, still etcd was not running | 03:23 |
| hongbin | so strange | 03:23 |
| madhuri | I manually rebooted it and then the same error | 03:23 |
| hongbin | but restart the etcd will bring it back? | 03:24 |
| hongbin | strange | 03:24 |
| madhuri | Does it work well with flannel network-driver? | 03:24 |
| hongbin | Not sure for the flannel staffs | 03:24 |
| openstackgerrit | Vilobh Meshram proposed openstack/magnum: Objects from Bay : Retrieve k8s API client https://review.openstack.org/220344 | 03:25 |
| madhuri | I restarted the etcd but no luck | 03:25 |
| *** devkulkarni has joined #openstack-containers | 03:25 | |
| *** shuquan has quit IRC | 03:25 | |
| hongbin | madhuri: sudo systemctl status etcd | 03:26 |
| hongbin | what is the status before and after restart | 03:26 |
| openstackgerrit | Vilobh Meshram proposed openstack/magnum: WIP : Objects from Bay - Replication Controller https://review.openstack.org/213368 | 03:27 |
| openstackgerrit | Vilobh Meshram proposed openstack/magnum: Objects from Bay : Retrieve k8s API client https://review.openstack.org/220344 | 03:27 |
| *** shuquan has joined #openstack-containers | 03:28 | |
| madhuri | hongbin: it failed to start | 03:28 |
| madhuri | http://paste.openstack.org/show/473682/ | 03:28 |
| hongbin | do your vm have external internet access? | 03:29 |
| hongbin | the log indicates it cannot connect to public discovery service | 03:30 |
| madhuri | Ok this is proxy issue | 03:30 |
| madhuri | Let me check | 03:30 |
| *** fawadkhaliq has joined #openstack-containers | 03:32 | |
| *** dimtruck is now known as zz_dimtruck | 03:32 | |
| hongbin | madhuri: have to leave now | 03:36 |
| madhuri | hongbin: Sure. Thank you | 03:37 |
| madhuri | hongbin: Good night | 03:37 |
| hongbin | madhuri: seems you find the root cause, it should be solved soon | 03:37 |
| hongbin | madhuri: thx. see you later | 03:37 |
| *** hongbin has quit IRC | 03:38 | |
| *** devkulkarni has quit IRC | 03:44 | |
| *** shuquan has quit IRC | 03:45 | |
| *** Marga__ has quit IRC | 03:46 | |
| *** Marga_ has joined #openstack-containers | 03:46 | |
| *** dims_ has joined #openstack-containers | 03:47 | |
| *** tbh has joined #openstack-containers | 03:47 | |
| *** kebray has quit IRC | 03:49 | |
| *** Marga_ has quit IRC | 03:50 | |
| *** devkulkarni has joined #openstack-containers | 03:52 | |
| *** shuquan has joined #openstack-containers | 03:53 | |
| *** dims_ has quit IRC | 03:54 | |
| *** jerrygb has quit IRC | 03:56 | |
| *** shuquan has quit IRC | 03:58 | |
| *** Marga_ has joined #openstack-containers | 04:02 | |
| *** eghobo has joined #openstack-containers | 04:05 | |
| *** jwcroppe has joined #openstack-containers | 04:14 | |
| *** achanda has quit IRC | 04:17 | |
| *** Tango has joined #openstack-containers | 04:21 | |
| *** harshs has quit IRC | 04:26 | |
| *** fawadk has joined #openstack-containers | 04:33 | |
| *** fawadkhaliq has quit IRC | 04:35 | |
| *** Marga_ has quit IRC | 04:36 | |
| *** Marga_ has joined #openstack-containers | 04:37 | |
| *** deepthi has joined #openstack-containers | 04:40 | |
| *** Tango|2 has joined #openstack-containers | 04:46 | |
| *** achanda has joined #openstack-containers | 04:48 | |
| *** Tango has quit IRC | 04:48 | |
| *** devkulkarni has quit IRC | 04:51 | |
| *** harshs has joined #openstack-containers | 04:59 | |
| *** mathspanda has joined #openstack-containers | 05:02 | |
| *** ganeshna has quit IRC | 05:07 | |
| *** ganeshna has joined #openstack-containers | 05:07 | |
| *** vimal has joined #openstack-containers | 05:27 | |
| *** harshs_ has joined #openstack-containers | 05:28 | |
| *** harshs has quit IRC | 05:30 | |
| *** harshs_ is now known as harshs | 05:30 | |
| *** ganeshna has quit IRC | 05:39 | |
| *** ganeshna has joined #openstack-containers | 05:40 | |
| *** harshs_ has joined #openstack-containers | 05:46 | |
| *** harshs has quit IRC | 05:49 | |
| *** harshs_ is now known as harshs | 05:49 | |
| eliqiao | hongbin: around? | 05:50 |
| eliqiao | anyone know where's magnum's gate git repository? | 05:50 |
| *** dims_ has joined #openstack-containers | 05:50 | |
| eliqiao | hi dims_ | 05:51 |
| *** tbh has quit IRC | 05:52 | |
| *** tbh has joined #openstack-containers | 05:53 | |
| *** harshs has quit IRC | 05:53 | |
| *** dims_ has quit IRC | 05:55 | |
| eliqiao | ping cores! | 05:57 |
| *** chandankumar has joined #openstack-containers | 05:58 | |
| *** jerrygb has joined #openstack-containers | 06:08 | |
| *** jerrygb has quit IRC | 06:12 | |
| *** fawadk has quit IRC | 06:14 | |
| *** fawadkhaliq has joined #openstack-containers | 06:17 | |
| *** chandankumar has quit IRC | 06:18 | |
| *** jwcroppe has quit IRC | 06:23 | |
| *** jwcroppe has joined #openstack-containers | 06:23 | |
| *** jwcroppe has quit IRC | 06:28 | |
| *** vilobhmm11 has quit IRC | 06:31 | |
| *** fawadk has joined #openstack-containers | 06:32 | |
| *** ganeshna has quit IRC | 06:32 | |
| *** fawadkhaliq has quit IRC | 06:33 | |
| *** ganeshna has joined #openstack-containers | 06:34 | |
| *** suro-patz has joined #openstack-containers | 06:35 | |
| *** j___ has joined #openstack-containers | 06:41 | |
| *** fawadkhaliq has joined #openstack-containers | 06:51 | |
| *** fawadk has quit IRC | 06:52 | |
| *** chandankumar has joined #openstack-containers | 06:53 | |
| *** alex_klimov has joined #openstack-containers | 06:55 | |
| *** eghobo has quit IRC | 06:55 | |
| *** alex_klimov has quit IRC | 06:56 | |
| *** alex_klimov has joined #openstack-containers | 06:56 | |
| *** fawadk has joined #openstack-containers | 06:58 | |
| *** fawadkhaliq has quit IRC | 07:00 | |
| *** BertrandN has joined #openstack-containers | 07:02 | |
| *** Tango|2 has quit IRC | 07:08 | |
| *** fawadkhaliq has joined #openstack-containers | 07:09 | |
| *** vimal has quit IRC | 07:09 | |
| *** fawadk has quit IRC | 07:10 | |
| *** achanda has quit IRC | 07:14 | |
| *** vimal has joined #openstack-containers | 07:16 | |
| *** BertrandN has quit IRC | 07:20 | |
| *** ganeshna has quit IRC | 07:23 | |
| *** jwcroppe has joined #openstack-containers | 07:23 | |
| *** suro-patz has quit IRC | 07:27 | |
| *** BertrandN has joined #openstack-containers | 07:37 | |
| *** fawadkhaliq has quit IRC | 07:49 | |
| *** fawadkhaliq has joined #openstack-containers | 07:50 | |
| *** dims_ has joined #openstack-containers | 07:52 | |
| *** dims_ has quit IRC | 07:59 | |
| eliqiao | cores, please look at https://review.openstack.org/226634 , that is blocking gate. | 08:05 |
| *** ganeshna has joined #openstack-containers | 08:06 | |
| *** jwcroppe has quit IRC | 08:19 | |
| *** ganeshna has quit IRC | 08:21 | |
| *** ganeshna has joined #openstack-containers | 08:23 | |
| eliqiao | ping tcammann1 | 08:30 |
| eliqiao | ping tcammann_ | 08:30 |
| *** junhongl has joined #openstack-containers | 08:32 | |
| *** alex_klimov has quit IRC | 08:36 | |
| *** alex_klimov has joined #openstack-containers | 08:36 | |
| *** jerrygb has joined #openstack-containers | 09:08 | |
| *** dims_ has joined #openstack-containers | 09:11 | |
| *** jerrygb has quit IRC | 09:13 | |
| *** jwcroppe has joined #openstack-containers | 09:16 | |
| *** madhuri has quit IRC | 09:36 | |
| *** ganeshna_ has joined #openstack-containers | 09:42 | |
| *** ganeshna has quit IRC | 09:43 | |
| *** i316653 has joined #openstack-containers | 09:44 | |
| i316653 | Hi | 09:44 |
| *** dims_ has quit IRC | 09:44 | |
| *** jwcroppe_ has joined #openstack-containers | 09:47 | |
| i316653 | How does Magnum find hosts? If there is the error "Message: No valid host was found. There are not enough hosts available." How would I debug this? | 09:48 |
| *** jwcroppe has quit IRC | 09:49 | |
| tcammann_ | hello eliqiao | 09:57 |
| tcammann_ | i316653: It uses Nova | 09:57 |
| tcammann_ | eliqiao: I'll take a look at the patch asap | 09:58 |
| *** lan has joined #openstack-containers | 10:00 | |
| *** Kennan_Vacation has quit IRC | 10:01 | |
| *** Kennan_Vacation has joined #openstack-containers | 10:02 | |
| *** dims_ has joined #openstack-containers | 10:09 | |
| *** alex_klimov has quit IRC | 10:09 | |
| *** alex_klimov has joined #openstack-containers | 10:24 | |
| *** ganeshna_ has quit IRC | 10:26 | |
| i316653 | tcammann_: Where does the fedora-21-atomic-3 image then run via Nova? ...and who would I get the error not enough hosts available? | 10:26 |
| *** i316653 is now known as meehanman | 10:29 | |
| *** jwcroppe_ has quit IRC | 10:30 | |
| *** jwcroppe has joined #openstack-containers | 10:31 | |
| *** ganeshna has joined #openstack-containers | 10:33 | |
| *** jwcroppe has quit IRC | 10:35 | |
| *** BhavyaM has joined #openstack-containers | 10:38 | |
| *** alex_klimov has quit IRC | 10:39 | |
| *** alex_klimov has joined #openstack-containers | 10:40 | |
| *** alex_klimov has quit IRC | 10:40 | |
| *** alex_klimov has joined #openstack-containers | 10:40 | |
| *** ganeshna has quit IRC | 10:40 | |
| *** ganeshna has joined #openstack-containers | 10:42 | |
| *** yuanying_ has joined #openstack-containers | 11:04 | |
| openstackgerrit | Tom Cammann proposed openstack/magnum: Unwedge the gate by only enabling barbican once https://review.openstack.org/226699 | 11:10 |
| tcammann_ | meehanman: No valid hosts comes from nova scheduler, I would check those logs | 11:11 |
| tcammann_ | You will need a understanding of Nova to debug this issue, I suggest reading some of the docs on Nova's architecture | 11:11 |
| tcammann_ | eliqiao: It's gonna take a while for infra to merge that patch, so I've put up one in Magnum which should do the same | 11:12 |
| *** lan has quit IRC | 11:13 | |
| meehanman | Thanks @tcammann_ | 11:16 |
| *** meehanman has quit IRC | 11:16 | |
| *** mathspanda has quit IRC | 11:18 | |
| *** jwcroppe has joined #openstack-containers | 11:21 | |
| *** meehanman has joined #openstack-containers | 11:21 | |
| *** yuanying_ has quit IRC | 11:23 | |
| openstackgerrit | Tom Cammann proposed openstack/magnum: Use --max-complexity flake8 rule https://review.openstack.org/224184 | 11:24 |
| openstackgerrit | Tom Cammann proposed openstack/magnum: Fix E251 unnecessarily ignored pep8 rule https://review.openstack.org/224198 | 11:24 |
| openstackgerrit | Tom Cammann proposed openstack/magnum: Fix H405 and E131 ignored pep8 rules https://review.openstack.org/224183 | 11:24 |
| *** shakamunyi has quit IRC | 11:25 | |
| *** jerrygb has joined #openstack-containers | 11:35 | |
| *** jerrygb has quit IRC | 11:39 | |
| *** devkulkarni has joined #openstack-containers | 11:42 | |
| *** ganeshna has quit IRC | 11:44 | |
| *** jerrygb has joined #openstack-containers | 11:48 | |
| *** achanda has joined #openstack-containers | 12:03 | |
| *** achanda has quit IRC | 12:18 | |
| *** fawadkhaliq has quit IRC | 12:21 | |
| *** fawadkhaliq has joined #openstack-containers | 12:22 | |
| *** devkulkarni has quit IRC | 12:22 | |
| *** jwcroppe_ has joined #openstack-containers | 12:22 | |
| *** zhenguo has quit IRC | 12:24 | |
| tcammann_ | https://review.openstack.org/226699 - Needed to unwedge the gate! | 12:24 |
| *** jwcroppe has quit IRC | 12:25 | |
| *** jwcroppe_ is now known as jwcroppe | 12:29 | |
| *** vimal has quit IRC | 12:32 | |
| *** vimal has joined #openstack-containers | 12:34 | |
| *** vimal has quit IRC | 12:39 | |
| *** rpothier has joined #openstack-containers | 12:45 | |
| *** vimal has joined #openstack-containers | 12:48 | |
| *** vimal is now known as vimal|busy | 12:51 | |
| *** vimal|busy is now known as vimal | 12:51 | |
| *** rlrossit has joined #openstack-containers | 13:02 | |
| *** deepthi has quit IRC | 13:07 | |
| *** julim has joined #openstack-containers | 13:11 | |
| *** erkules_ is now known as erkules | 13:17 | |
| *** erkules has joined #openstack-containers | 13:17 | |
| *** dims_ has quit IRC | 13:21 | |
| *** dims_ has joined #openstack-containers | 13:21 | |
| *** jwcroppe has quit IRC | 13:26 | |
| *** Tango|2 has joined #openstack-containers | 13:30 | |
| *** bradjones has quit IRC | 13:35 | |
| *** chandankumar has quit IRC | 13:39 | |
| *** jwcroppe has joined #openstack-containers | 13:49 | |
| *** ganeshna has joined #openstack-containers | 13:50 | |
| *** tbh has quit IRC | 13:56 | |
| *** ganeshna has quit IRC | 13:57 | |
| *** dane_leblanc has joined #openstack-containers | 14:01 | |
| *** jerrygb has quit IRC | 14:03 | |
| *** jerrygb has joined #openstack-containers | 14:07 | |
| *** fawadkhaliq has quit IRC | 14:15 | |
| *** chandankumar has joined #openstack-containers | 14:18 | |
| *** harshs has joined #openstack-containers | 14:18 | |
| *** devkulkarni has joined #openstack-containers | 14:19 | |
| *** zz_dimtruck is now known as dimtruck | 14:28 | |
| *** Drago has joined #openstack-containers | 14:38 | |
| *** Drago has quit IRC | 14:39 | |
| *** Drago has joined #openstack-containers | 14:40 | |
| *** j___ has quit IRC | 14:47 | |
| *** harshs_ has joined #openstack-containers | 14:48 | |
| *** harshs has quit IRC | 14:50 | |
| *** harshs_ is now known as harshs | 14:50 | |
| openstackgerrit | Hongbin Lu proposed openstack/magnum: Implement bay monitoring and notifications https://review.openstack.org/215335 | 14:52 |
| *** j___ has joined #openstack-containers | 15:16 | |
| *** hongbin has joined #openstack-containers | 15:22 | |
| *** j___ has quit IRC | 15:25 | |
| *** kebray has joined #openstack-containers | 15:26 | |
| *** harshs has quit IRC | 15:26 | |
| *** j___ has joined #openstack-containers | 15:27 | |
| *** adrian_otto has joined #openstack-containers | 15:28 | |
| *** kebray has quit IRC | 15:30 | |
| *** alejandrito has joined #openstack-containers | 15:33 | |
| *** puranamr has joined #openstack-containers | 15:35 | |
| *** kebray has joined #openstack-containers | 15:45 | |
| *** kebray has quit IRC | 15:46 | |
| *** puranamr has quit IRC | 15:47 | |
| *** kebray has joined #openstack-containers | 15:50 | |
| *** puranamr has joined #openstack-containers | 15:50 | |
| *** ganeshna has joined #openstack-containers | 15:52 | |
| *** ganeshna has quit IRC | 15:52 | |
| *** harshs has joined #openstack-containers | 15:52 | |
| *** chandankumar has quit IRC | 15:52 | |
| *** ganeshna has joined #openstack-containers | 15:56 | |
| *** sdake has joined #openstack-containers | 15:57 | |
| *** harshs has quit IRC | 15:57 | |
| *** harshs has joined #openstack-containers | 15:58 | |
| *** harshs has quit IRC | 15:58 | |
| *** BhavyaM has quit IRC | 15:59 | |
| openstackgerrit | Merged openstack/magnum: Unwedge the gate by only enabling barbican once https://review.openstack.org/226699 | 15:59 |
| *** alex_klimov has quit IRC | 16:02 | |
| *** sthillma has joined #openstack-containers | 16:03 | |
| *** sthillma_ has joined #openstack-containers | 16:04 | |
| *** sdake has quit IRC | 16:04 | |
| *** sdake_ has joined #openstack-containers | 16:06 | |
| *** sthillma has quit IRC | 16:08 | |
| *** sthillma_ is now known as sthillma | 16:08 | |
| *** vimal has quit IRC | 16:10 | |
| *** sdake has joined #openstack-containers | 16:10 | |
| *** kebray has quit IRC | 16:11 | |
| *** kebray has joined #openstack-containers | 16:12 | |
| *** sdake_ has quit IRC | 16:14 | |
| *** dane_leblanc has quit IRC | 16:15 | |
| *** dane_leblanc has joined #openstack-containers | 16:15 | |
| *** BertrandN has quit IRC | 16:16 | |
| *** amit213 has quit IRC | 16:17 | |
| *** amit213 has joined #openstack-containers | 16:17 | |
| *** manjeets has joined #openstack-containers | 16:19 | |
| manjeets | adrian_otto and hongbin: please check documentation update I rebased the patch https://review.openstack.org/#/c/221504/ | 16:20 |
| openstackgerrit | Hongbin Lu proposed openstack/magnum: Implement bay monitoring and notifications https://review.openstack.org/215335 | 16:20 |
| hongbin | manjeets: will review that later | 16:21 |
| manjeets | ok | 16:22 |
| adrian_otto | thanks manjeets | 16:23 |
| tcammann_ | adrian_otto: +1 on the patch backlog cleaning :) | 16:25 |
| *** jwcroppe has quit IRC | 16:25 | |
| adrian_otto | tcammann1 or tcammann_ are you around? | 16:26 |
| tcammann_ | Yes, I just messaged you... ^^ | 16:26 |
| openstackgerrit | Merged openstack/magnum: Remove coding:utf-8 https://review.openstack.org/211057 | 16:26 |
| adrian_otto | aha, I did find one I wanted to ask you about… an important one in my view: https://review.openstack.org/211771 | 16:27 |
| adrian_otto | Any chance you could post a revision on that patch today? | 16:27 |
| adrian_otto | or should we find someone to adopt that work? | 16:27 |
| adrian_otto | tcammann1: yes, I think that policy tweak was a good one. I like applying more common sense. | 16:28 |
| tcammann_ | Yes, I did need some poking. My home PC and laptop are bust at the moment is going to be my goto excuse. But I'm still in the office :) | 16:29 |
| tcammann_ | Need a +2/A on the gate fix to get anything passing CI | 16:30 |
| adrian_otto | tcammann1: I just wanted an indication from you i I should abandon that one, and look for a foster parent, or jsut leave it for now and expect a revision soon… maybe you could post a comment in the comment stream letting us know you are still working on it. | 16:30 |
| adrian_otto | the gate fix merged earlier, didn't it? | 16:30 |
| adrian_otto | tcammann_: are you talking about the one in project-config? | 16:31 |
| openstackgerrit | Merged openstack/magnum: Docs update for new fedora atomic 1.0.4 https://review.openstack.org/221504 | 16:32 |
| tcammann_ | adrian_otto: No, you +2'd it earlier but it is the equivilent of the project-config one | 16:34 |
| *** kebray has quit IRC | 16:35 | |
| tcammann_ | oh you are right! | 16:35 |
| tcammann_ | it did land | 16:35 |
| adrian_otto | ok, so we should be in good shape when those rechecks I ordered complete | 16:35 |
| *** kebray has joined #openstack-containers | 16:40 | |
| *** j___ has quit IRC | 16:41 | |
| *** daneyon has joined #openstack-containers | 16:43 | |
| *** daneyon_ has quit IRC | 16:43 | |
| *** eghobo has joined #openstack-containers | 16:47 | |
| *** banix has joined #openstack-containers | 16:47 | |
| *** puranamr has quit IRC | 16:47 | |
| *** meehanman has quit IRC | 16:50 | |
| *** suro-patz has joined #openstack-containers | 16:51 | |
| *** manjeets has quit IRC | 16:53 | |
| *** unicell has joined #openstack-containers | 16:54 | |
| *** jwcroppe has joined #openstack-containers | 16:55 | |
| adrian_otto | Tango|2: ping | 16:57 |
| *** harshs has joined #openstack-containers | 16:58 | |
| *** Tango|2 has quit IRC | 16:59 | |
| adrian_otto | yuanying: ping | 16:59 |
| *** mfalatic has joined #openstack-containers | 17:02 | |
| suro-patz | tcammann_: tcammann1: ping | 17:02 |
| *** mfalatic has quit IRC | 17:03 | |
| *** dane_leblanc has quit IRC | 17:04 | |
| *** mfalatic has joined #openstack-containers | 17:04 | |
| tcammann_ | sup! | 17:06 |
| tcammann_ | suro-patz: | 17:06 |
| suro-patz | regarding https://review.openstack.org/#/c/220702/, I hear the consensus to rename healthcheck | 17:07 |
| suro-patz | I am planning to rename it to magnum_service | 17:07 |
| suro-patz | tcammmann_:^^ | 17:07 |
| suro-patz | before I start making the wholesome renaming, I thought I will bounce the idea with you once | 17:08 |
| tcammann_ | Why not service? | 17:08 |
| suro-patz | I had weighed that option | 17:08 |
| tcammann_ | We are probably getting rid of the existing service table with removable of objects from the db | 17:08 |
| tcammann_ | *removal | 17:09 |
| suro-patz | that will create mich more wider touch surface for this patch | 17:09 |
| suro-patz | much | 17:09 |
| openstackgerrit | Adrian Otto proposed openstack/magnum: Add TLS support in Magnum. https://review.openstack.org/194905 | 17:09 |
| tcammann_ | I agree, but consistency with other services is useful. | 17:09 |
| suro-patz | and from code reading point of view I felt magnum already had the notion of service which people are used to | 17:10 |
| *** achanda has joined #openstack-containers | 17:10 | |
| tcammann_ | Like I mention in the patch, we have already renamed the api. Why not commit fully and rename the db | 17:10 |
| suro-patz | We have renamed the CLI - user facing part | 17:11 |
| *** dane_leblanc has joined #openstack-containers | 17:13 | |
| suro-patz | tcammann_: I am guessing you still prefer me to change to 'service' than 'magnum-service' | 17:14 |
| suro-patz | my worry is that will create a very large touch surface, given we are almost at the end of the cycle | 17:15 |
| *** suro-patz has quit IRC | 17:24 | |
| *** suro-patz has joined #openstack-containers | 17:25 | |
| eghobo | tcammann_: curent services table, is it for Kub services? | 17:25 |
| suro-patz | eghobo: correct | 17:25 |
| suro-patz | adrian_otto: any comment on the above? | 17:26 |
| eghobo | aha, it's better rename it and move healthcheck to services ;) | 17:26 |
| suro-patz | eghobo: I am proposing making WIP:healthcheck to magnum_service and keep the kub-services as services to reduce code touch points | 17:27 |
| suro-patz | given the release cycle timeline | 17:28 |
| suro-patz | but if community feels that renaming is absolutely needed, I will commit for doing the same. eghobo: tcammann_: | 17:28 |
| eghobo | I see, my personal opinion rename and use right name from beginning. but magnum_services will work as intermedia solution too | 17:30 |
| adrian_otto | my suggestion on this subject is to table it for the M release. | 17:31 |
| adrian_otto | this is not going to make a profound difference for users | 17:31 |
| suro-patz | but the rest of the implementation was all done! | 17:31 |
| adrian_otto | and there are a bunch of things in flight that will that could use additional eyes and hands | 17:31 |
| eghobo | adrian_otto: make sense | 17:31 |
| adrian_otto | suro-patz: I'm not saying it's a bad idea, I'm simply concerned about the timing of it. | 17:32 |
| suro-patz | we can create a tech debt for the renaming to address in M | 17:32 |
| adrian_otto | yes, that would be appropriate | 17:32 |
| suro-patz | cool! | 17:32 |
| adrian_otto | WHat I really want is for this topic to get totally finished: | 17:33 |
| adrian_otto | https://review.openstack.org/#/q/status:open+project:openstack/magnum+branch:master+topic:bp/secure-kubernetes,n,z | 17:33 |
| suro-patz | adrian_otto: I can offer some help, as the other BP is mostly wrapped up for | 17:33 |
| adrian_otto | so if you have some free cycles to contribute, please see about adopting one or more of these patches and driving it home | 17:33 |
| *** devkulkarni has quit IRC | 17:35 | |
| adrian_otto | let me know if any guidance is needed for how to make your own revisions to an existing patch | 17:35 |
| suro-patz | will get back, if I need any assistance on these | 17:35 |
| suro-patz | do you have any particular patch in mind out of these 7? | 17:36 |
| suro-patz | I will pick up, as I plough through them | 17:36 |
| adrian_otto | suro-patz: https://review.openstack.org/207324 is one that got stuck, and needs a good workaround | 17:39 |
| suro-patz | adrian_otto: ack | 17:41 |
| adrian_otto | tx suro-patz! | 17:43 |
| openstackgerrit | Hongbin Lu proposed openstack/magnum: Rename image property os-distro to os_distro https://review.openstack.org/226893 | 17:43 |
| *** vilobhmm11 has joined #openstack-containers | 17:43 | |
| *** rlrossit has quit IRC | 17:48 | |
| *** sthillma has quit IRC | 17:50 | |
| *** Tango|2 has joined #openstack-containers | 17:52 | |
| *** devkulkarni has joined #openstack-containers | 17:53 | |
| *** rlrossit has joined #openstack-containers | 17:53 | |
| *** ganeshna has quit IRC | 18:07 | |
| *** jwcroppe has quit IRC | 18:08 | |
| *** annegentle has joined #openstack-containers | 18:13 | |
| *** eghobo_ has joined #openstack-containers | 18:19 | |
| *** eghobo has quit IRC | 18:22 | |
| openstackgerrit | Tom Cammann proposed openstack/magnum: [WIP] Generate heat templates from master template https://review.openstack.org/211771 | 18:22 |
| tcammann_ | adrian_otto: put up another patch on that heat template one. I'll try and update the otehr one soon | 18:23 |
| adrian_otto | tcammann_: thanks!! | 18:25 |
| *** achanda has quit IRC | 18:26 | |
| adrian_otto | it's pretty late to be at work Thanks for staying to get this in. | 18:26 |
| *** dims__ has joined #openstack-containers | 18:28 | |
| *** achanda has joined #openstack-containers | 18:30 | |
| *** dims_ has quit IRC | 18:31 | |
| openstackgerrit | Hongbin Lu proposed openstack/magnum: Use option --os-distro to set image property https://review.openstack.org/226893 | 18:34 |
| vilobhmm11 | adrian_otto : regarding https://review.openstack.org/#/c/217908/ IMHO we should make similar changes for other k8s objects as well (including filing seperate tech-debt bugs to check if we might need it in future); to be consistent | 18:35 |
| adrian_otto | vilobhmm11: ok, I'll read that one now | 18:40 |
| vilobhmm11 | adrian_otto : thanks! | 18:40 |
| adrian_otto | vilobhmm11: ok, so this is on the topic of removing unused code, filing bugs for anticipated improvements to use the code at a later time, when we will pull it back in, right? | 18:41 |
| vilobhmm11 | adrian_otto : yes | 18:42 |
| *** eghobo_ has quit IRC | 18:42 | |
| vilobhmm11 | based on your comment here https://review.openstack.org/#/c/217908/ IMHO we should do it for all k8s objects (Service/rc/pods) | 18:42 |
| adrian_otto | Yes, we can certainly do more of that. I'd caution us from putting too much energy into that at the end of a release cycle. | 18:42 |
| adrian_otto | but I don't object to it | 18:43 |
| adrian_otto | I do agree that using a measured, and consistent approach to this is the right approach | 18:44 |
| vilobhmm11 | adrian_otto : the patches have already been proposed …they won't consume much of time | 18:45 |
| vilobhmm11 | just need to rebase them | 18:45 |
| vilobhmm11 | I agree with you plan | 18:45 |
| adrian_otto | great! | 18:46 |
| vilobhmm11 | adrian_otto : thanks! | 18:47 |
| *** kebray has quit IRC | 18:52 | |
| *** alex_klimov has joined #openstack-containers | 19:03 | |
| openstackgerrit | Surojit Pathak proposed openstack/magnum: Backend support for magnum service https://review.openstack.org/220702 | 19:05 |
| *** Marga_ has quit IRC | 19:06 | |
| *** Marga_ has joined #openstack-containers | 19:08 | |
| *** Marga_ has quit IRC | 19:13 | |
| openstackgerrit | Surojit Pathak proposed openstack/magnum: Adding API support for magnum service https://review.openstack.org/226078 | 19:22 |
| *** dims__ has quit IRC | 19:24 | |
| *** dims_ has joined #openstack-containers | 19:25 | |
| *** annegentle has quit IRC | 19:32 | |
| *** annegentle has joined #openstack-containers | 19:32 | |
| *** Marga_ has joined #openstack-containers | 19:37 | |
| *** achanda has quit IRC | 19:41 | |
| *** kebray has joined #openstack-containers | 19:46 | |
| *** Marga_ has quit IRC | 19:50 | |
| *** rlrossit1 has joined #openstack-containers | 19:52 | |
| *** rlrossit has quit IRC | 19:52 | |
| *** rlrossit1 has quit IRC | 19:57 | |
| *** rlrossit has joined #openstack-containers | 20:01 | |
| *** fawadkhaliq has joined #openstack-containers | 20:02 | |
| *** fawadkhaliq has quit IRC | 20:05 | |
| *** jwcroppe has joined #openstack-containers | 20:08 | |
| *** gangil has joined #openstack-containers | 20:11 | |
| *** sdake_ has joined #openstack-containers | 20:22 | |
| openstackgerrit | Hongbin Lu proposed openstack/magnum: Use api version 1 to set image property https://review.openstack.org/226893 | 20:23 |
| *** sdake has quit IRC | 20:25 | |
| *** sthillma has joined #openstack-containers | 20:38 | |
| *** achanda has joined #openstack-containers | 20:42 | |
| suro-patz | hongbin: aware of this failure http://logs.openstack.org/78/226078/3/check/gate-functional-dsvm-magnum/f424944/console.html? | 20:46 |
| *** achanda has quit IRC | 20:47 | |
| hongbin | suro-patz: yes, trying to fix that | 20:47 |
| suro-patz | hongbin: cool! | 20:48 |
| *** achanda has joined #openstack-containers | 20:49 | |
| *** dane_leblanc_ has joined #openstack-containers | 20:53 | |
| *** achanda has quit IRC | 20:55 | |
| *** achanda has joined #openstack-containers | 20:55 | |
| *** dane_leblanc has quit IRC | 20:56 | |
| *** puranamr has joined #openstack-containers | 20:57 | |
| *** harshs has quit IRC | 21:01 | |
| *** harshs has joined #openstack-containers | 21:04 | |
| *** rlrossit has left #openstack-containers | 21:10 | |
| openstackgerrit | James E. Blair proposed openstack/magnum: Remove ERROR_ON_CLONE references https://review.openstack.org/227000 | 21:11 |
| *** julim has quit IRC | 21:12 | |
| *** puranamr has quit IRC | 21:12 | |
| *** Marga_ has joined #openstack-containers | 21:21 | |
| daneyon | hongbin adrian_otto sdake tcammann_ As discussed during our weekly meeting, my flannel driver implementation for swarm replaces swarm's public discovery with a local etcd discovery. Note: etcd still uses public discovery. In this case, their is no need for several configuration directives such as swarm_discovery_url_format, public_swarm_discovery, public_swarm_discovery_url | 21:32 |
| daneyon | flannel for swarm patch: https://review.openstack.org/#/c/224367/3 | 21:32 |
| *** achanda has quit IRC | 21:32 | |
| adrian_otto | thanks daneyon. I'll have a look in a moment. | 21:32 |
| *** eghobo has joined #openstack-containers | 21:32 | |
| *** rpothier has quit IRC | 21:33 | |
| daneyon | We could defaualt public_swarm_discovery to false, which would use etcd. However, if set to true, then flannel would not work. | 21:33 |
| *** kebray has quit IRC | 21:34 | |
| *** achanda has joined #openstack-containers | 21:36 | |
| *** dane_leblanc_ has quit IRC | 21:37 | |
| daneyon | unless the group has a strong opinion about keeping swarm public discovery, I'm inclined to remove all the swarm public discovery stuff and use etcd. I don't see any network drivers that require swarm's public discovery mechanism and I don;t know why a user/operator would want to use the public discovery if we have a local etcd-based discovery mechanism. Since etcd still uses public discovery, I have filed a bp to address disc | 21:38 |
| daneyon | overy more holistically https://blueprints.launchpad.net/magnum/+spec/bay-type-discovery-options I just want to make sure the core's are on board so I don't have to go back and make major changes to the patch. | 21:38 |
| adrian_otto | the reason it's there is that we did not have etcd included in every bay | 21:46 |
| adrian_otto | daneyon: so rather than forcing everyone to set up etcd on their own as a prerequisite to use magnum, we used the public registry as a fallback, and allow you to specify an address of your own if you have one. | 21:49 |
| daneyon | adrian_otto the heat template sets up etcd as part of instantiating the bay. This is just how we do it in the k8s bay type. | 21:50 |
| daneyon | swarm will use etcd for backing discovery of the swarm cluster instead of using Docker Hub | 21:51 |
| daneyon | let me think for a sec | 21:51 |
| *** harshs has quit IRC | 21:51 | |
| daneyon | adrian_otto what is the value of allowing a user to specify their own public discovery if we have one (etcd) baked into the swarm bay type? | 21:52 |
| daneyon | I think it leaves the door open to issues. | 21:52 |
| daneyon | I am trying to understand why a user would want to manually specify and change configs to use a public discovery over one that is part of the swarm bay deployment. | 21:53 |
| adrian_otto | I might have one discovery service for my cloud tat's on my network, and all of my bays use that | 21:53 |
| adrian_otto | that way I don't need to run N copies of etcd for that purpose | 21:54 |
| *** harshs has joined #openstack-containers | 21:54 | |
| adrian_otto | lower system utilization overall by leveraging a more widely shared resource | 21:54 |
| openstackgerrit | Ton Ngo proposed openstack/magnum: Migrate to Kubernetes Release 1 (docs) https://review.openstack.org/207300 | 21:54 |
| eghobo | can team review https://review.openstack.org/#/c/226893/, it will fix gate | 21:54 |
| adrian_otto | eghobo: done | 21:55 |
| daneyon | then maybe discovery should be more pluggable inside magnum. | 21:55 |
| adrian_otto | you should be able to specify your own discovery URL today | 21:56 |
| hongbin | adrian_otto: daneyon How about move the etcd to magnum, making magnum-as-a-etcd | 21:56 |
| adrian_otto | did that code somehow not land? | 21:56 |
| eghobo | thx, sorry I didn't realize that you have discussion | 21:56 |
| adrian_otto | hongbin: Do you mean run etcd on a magnum bay by default, and allow all bays to access that for discovery? | 21:57 |
| *** achanda has quit IRC | 21:57 | |
| daneyon | hongbin can you explain a bit more? | 21:57 |
| hongbin | run a etcd process in magnum control plane, like magnum-etcd | 21:57 |
| hongbin | It is in parallel with magnum-api and magnum-conductor | 21:58 |
| hongbin | It solves the efficiency issue | 21:58 |
| adrian_otto | yes, that makes sense | 21:58 |
| eghobo | not sure, we need discovery just for token | 21:59 |
| *** sdake_ has quit IRC | 21:59 | |
| eghobo | why we cannot generate it and pass to heat | 21:59 |
| daneyon | adrian_otto the swarm public discovery is only used for bootstrapping a swarm cluster: https://github.com/docker/swarm/tree/master/discovery | 21:59 |
| adrian_otto | does etcd have multi-tenancy support, or is it globally writable? | 21:59 |
| daneyon | etcd is not multi-tenant and is not secure | 21:59 |
| adrian_otto | right, so that justifies enabling it per bay as an optional service | 22:00 |
| hongbin | k. then the multi-tenancy is a problem | 22:00 |
| daneyon | so swarm's public discovery is not meant to be used for application discovery services | 22:00 |
| *** achanda has joined #openstack-containers | 22:00 | |
| adrian_otto | such that we can use a networking boundary to limit access to etcd | 22:01 |
| daneyon | is their a need to have multiple bays use the same public discovery token? I don;t think so, since each bay is meant to be a clustering system. | 22:01 |
| *** klindgren has quit IRC | 22:01 | |
| *** muralia has quit IRC | 22:02 | |
| *** klindgren has joined #openstack-containers | 22:02 | |
| *** muralia has joined #openstack-containers | 22:02 | |
| adrian_otto | not technically, no, but Imagine that you have 50,000 bays or more | 22:03 |
| adrian_otto | now you have triple that: 150,000 etcd processes running on your infrastrucuture | 22:04 |
| daneyon | adrian_otto hongbin eghobo I can see having options for etcd discovery, public (how it's done today) and a local option (default to public). I am still struggling on why we even need Swarm's public discovery. Keep in mind, Swarm's public disocery mechanism is only used for bootstrapping a swarm cluster and not used for app discovery purposes. | 22:04 |
| adrian_otto | hogging up your memory when all you really want to know is the membership info for each swarm cluster | 22:04 |
| adrian_otto | and you could accomplish that with a single etcd per tenant, rather than one per bay… and in a private cloud scenario, possibly one per cloud. | 22:05 |
| daneyon | adrian_otto that's a good point. I think in that case it would be beneficial for OS to have a k/v store as a service so M can just consume that service instead of creating etcd clusters per bay. | 22:06 |
| adrian_otto | swarm just keeps a list of the nodes in the discovery service | 22:07 |
| daneyon | correct | 22:07 |
| adrian_otto | so that it does not need tom implement shared state | 22:07 |
| eghobo | daneyon: does it mean we need etcd only for | 22:07 |
| eghobo | $ swarm create | 22:07 |
| eghobo | 6856663cdefdec325839a4b7e1de38e8 # <- this is your unique <cluster_id> | 22:07 |
| adrian_otto | that just guarantees that no two swarm clusters end up with the same hash string | 22:08 |
| daneyon | eghobo etcd could be used for than just bootstrapping the swarm cluster. it can be used for shared config service, app discovery, etc.. | 22:08 |
| daneyon | whereas swarm pub disc is only used for bootstrapping a swarm cluster | 22:08 |
| daneyon | eghobo: https://github.com/openstack/magnum/blob/master/magnum/templates/docker-swarm/fragments/write-swarm-master-service.yaml#L19 | 22:09 |
| adrian_otto | yes, it may be useful to have an etcd in your bay for app level discovery regardless of the bay type | 22:09 |
| openstackgerrit | Ton Ngo proposed openstack/magnum: Migrate to Kubernetes Release 1 (docs) https://review.openstack.org/207300 | 22:09 |
| daneyon | used for $DISCOVRY_URL | 22:09 |
| eghobo | daneyon: honestly I don't difference between kub and swarm | 22:10 |
| adrian_otto | I just think that feature should be controlled by a setting in the baymodel that defaults to off. | 22:10 |
| daneyon | adrian_otto we need etcd in all bays that will use flannel for networking. | 22:10 |
| daneyon | it would be helpful if flannel supported diff k/v backends, but it does not. | 22:11 |
| adrian_otto | ok, what does flannel use it for? | 22:11 |
| daneyon | shared config | 22:11 |
| adrian_otto | ok | 22:11 |
| eghobo | daneyon: but we can run etcd at swarm masters, can't we? | 22:11 |
| adrian_otto | yes. | 22:11 |
| daneyon | eghobo yes, that's how i implemented it | 22:12 |
| adrian_otto | we should do it on a separate container on the swarm master node(s) of the bay | 22:12 |
| hongbin | The problem is we could have multiple masters in future | 22:12 |
| hongbin | maybe etcd in a dedicated vm | 22:12 |
| adrian_otto | can etcd support TLS? | 22:12 |
| *** devkulkarni has quit IRC | 22:13 | |
| daneyon | eghobo hongbin adrian_otto https://docs.docker.com/swarm/multi-manager-setup/ | 22:13 |
| vilobhmm11 | https://coreos.com/etcd/docs/0.4.7/etcd-security/ "etcd supports SSL/TLS as well as authentication through client certificates, both for clients to server as well as peer (server to server / cluster) communication." | 22:13 |
| daneyon | ^ keep in mind that we do not provide ha/scale for swarm bay types | 22:13 |
| vilobhmm11 | adrian_otto : ^^ | 22:13 |
| eghobo | hongbin but we can run etcd at each master and join then to cluster | 22:13 |
| daneyon | as the doc states, etcd, consul or zk is required for swarm master ha | 22:13 |
| hongbin | eghobo: we need another etcd to bootstrap the etcd cluster in master | 22:14 |
| daneyon | further reason to implement a real k/v store for swarm | 22:14 |
| *** annegent_ has joined #openstack-containers | 22:14 | |
| eghobo | loop ;), but it's just for token | 22:14 |
| daneyon | and if we need a real kv store for swarm ha, do we really want to support a separate Docker Hub discovery just for bootstrapping a swarm cluster? | 22:15 |
| adrian_otto | magnum could have a generate_uuid function that would generate unique seed strings | 22:15 |
| daneyon | I think, at least for the near term, keep it simple. | 22:15 |
| daneyon | less knobs for users to mess with | 22:15 |
| *** openstackgerrit has quit IRC | 22:16 | |
| *** openstackgerrit has joined #openstack-containers | 22:16 | |
| hongbin | daneyon: there is already a project to implement k/v store for openstack https://wiki.openstack.org/wiki/MagnetoDB | 22:17 |
| hongbin | And this project is going to die | 22:17 |
| hongbin | since no activity | 22:18 |
| *** annegentle has quit IRC | 22:18 | |
| hongbin | eghobo: that is how k8s HA works currently | 22:22 |
| hongbin | eghobo: sorry, mistake | 22:23 |
| daneyon | yeah, doesn't look good. | 22:23 |
| daneyon | and doesn't have support for etcd or consul | 22:24 |
| eghobo | daneyon: sorry, what doesn't support etcd? | 22:25 |
| daneyon | hongbin we could use static list or dns for cluster bootstrapping too: https://github.com/coreos/etcd/blob/master/Documentation/clustering.md | 22:25 |
| daneyon | eghobo the Magneto project | 22:26 |
| hongbin | I am trying to use static before, but cannot figure out how to do it in a heat template | 22:26 |
| eghobo | yep, it's Cassandra I don't see any values at all | 22:27 |
| *** banix has quit IRC | 22:27 | |
| eghobo | hongbin, daneyon: can we postpohone decision for couple of days, I need to do some reading | 22:34 |
| daneyon | hongbin maybe we could create a resource that would take the ip_addr output from each master iteration and create the ETCD_INITIAL_CLUSTER and ETCD_INITIAL_CLUSTER_STATE=new env's then starts up the etcd service and all other services | 22:34 |
| hongbin | eghobo: np from me | 22:35 |
| daneyon | it sounds like maybe i just need to keep all the swarm public discovery stuff and just change the default of public_swarm_discovery from true to false. | 22:35 |
| daneyon | false will use etcd for bootstrapping the swarm cluster | 22:36 |
| daneyon | keep in mind that setting this to true will cause breakage | 22:36 |
| *** alejandrito has quit IRC | 22:36 | |
| daneyon | hongbin adrian_otto ^ | 22:36 |
| hongbin | daneyon: I think you original proposal is completely switching from public_swarm_discovery to etcd discovery | 22:38 |
| *** kebray has joined #openstack-containers | 22:38 | |
| daneyon | hongbin correct... remove swarm public discovery and just use etcd | 22:39 |
| hongbin | I think I like this better than a breakage config (public_swarm_discovery) | 22:39 |
| hongbin | although I haven't review your patch yet. I might change my mind after | 22:40 |
| hongbin | :) | 22:40 |
| daneyon | hongbin we could support multiple discovery mechanisms per bay by changing discovery_url to etcd_discovery_url and swarm_discovery_url | 22:41 |
| hongbin | daneyon: sounds good | 22:41 |
| daneyon | however, I think that's overkill | 22:41 |
| daneyon | imo we need to pick one | 22:42 |
| adrian_otto | hummm | 22:42 |
| daneyon | otherwise, we overcomplicate the situation | 22:42 |
| hongbin | I am OK with that | 22:42 |
| eghobo | daneyon +2 | 22:42 |
| adrian_otto | having two settings for this seems excessive | 22:42 |
| *** dims__ has joined #openstack-containers | 22:43 | |
| daneyon | again, i like using etcd since swarm's public discovery mechanism is only for bootstrapping the swarm cluster and not for cluster ha, app discovery services, etc.. | 22:43 |
| daneyon | swarm will use etcd for bootstrapping. We will need etcd to add swarm master ha/scale. | 22:43 |
| daneyon | ^ either etcd, consul or zk. | 22:43 |
| *** dims_ has quit IRC | 22:46 | |
| *** annegent_ has quit IRC | 22:49 | |
| daneyon | adrian_otto agreed, but if we run 2 discovery services in a bay, each requires its own unique discovery url. | 22:49 |
| adrian_otto | oh, my you lost me, sorry. Why would we run 2 discovery services in the same bay? I thought the idea was to simply one one per bay, and use it for both purposes? | 22:51 |
| *** alex_klimov has quit IRC | 22:54 | |
| *** jerrygb has quit IRC | 22:55 | |
| *** jwcroppe has quit IRC | 22:56 | |
| *** jwcroppe has joined #openstack-containers | 22:57 | |
| daneyon | adrian_otto If we wanted to continue having Swarm use a public discovery mechanism and support flannel. This is because flannel requires etcd. So 1 discovery service, Docker Hub, for Swarm bootstrapping and 1 for bootstrapping etcd. | 22:59 |
| *** hongbin has quit IRC | 22:59 | |
| daneyon | adrian_otto since etcd, consul or zk is required for swarm master ha, I think it's preferred to simply settle on using etcd for bootstrapping the swarm cluster instead of using Docker Hub. | 23:01 |
| *** jwcroppe has quit IRC | 23:01 | |
| eghobo | daneyon: +1 | 23:03 |
| daneyon | eghobo thx for the feedback | 23:03 |
| *** annegentle has joined #openstack-containers | 23:04 | |
| adrian_otto | agreed | 23:05 |
| *** annegentle has quit IRC | 23:05 | |
| *** eghobo has quit IRC | 23:05 | |
| openstackgerrit | Merged openstack/magnum: Use api version 1 to set image property https://review.openstack.org/226893 | 23:12 |
| *** adrian_otto has quit IRC | 23:19 | |
| *** kebray has quit IRC | 23:25 | |
| *** devkulkarni has joined #openstack-containers | 23:27 | |
| *** achanda has quit IRC | 23:33 | |
| *** ganeshna has joined #openstack-containers | 23:33 | |
| openstackgerrit | Motohiro/Yuanying Otsuka proposed openstack/magnum: Make bay.api_address contains protocol https://review.openstack.org/223025 | 23:34 |
| *** jwcroppe has joined #openstack-containers | 23:37 | |
| *** dimtruck is now known as zz_dimtruck | 23:42 | |
| *** Tango|2 has quit IRC | 23:43 | |
| *** jwcroppe has quit IRC | 23:45 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!