| *** Qiming has quit IRC | 00:00 | |
| *** heyongli has joined #senlin | 00:46 | |
| *** ChrisSen has joined #senlin | 00:50 | |
| *** ChrisSen has quit IRC | 01:02 | |
| *** Qiming has joined #senlin | 01:05 | |
| *** Yanyanhu has joined #senlin | 01:16 | |
| openstackgerrit | Qiming Teng proposed stackforge/python-senlinclient: Remove private copy of clustering_service https://review.openstack.org/209718 | 01:20 |
|---|---|---|
| *** Yanyanhu has quit IRC | 01:20 | |
| Qiming | heyongli, long time no see | 01:21 |
| heyongli | recently work on CI | 01:21 |
| Qiming | yes, it's a mess | 01:21 |
| heyongli | broken about 1 month... | 01:21 |
| Qiming | saw some discussions on mailinglist | 01:21 |
| Qiming | it is finally up and working? | 01:22 |
| heyongli | yeah | 01:22 |
| Qiming | bravo! | 01:22 |
| heyongli | gfw | 01:22 |
| Qiming | sigh | 01:23 |
| *** ChrisSen has joined #senlin | 01:23 | |
| *** Tennyson has joined #senlin | 01:35 | |
| *** jdandrea has quit IRC | 01:39 | |
| openstackgerrit | Yanyan Hu proposed stackforge/senlin: [WIP]Start point of Senlin functional tests https://review.openstack.org/208355 | 01:52 |
| *** Yanyan has joined #senlin | 01:52 | |
| Qiming | Yanyan, | 01:52 |
| Qiming | 208355 | 01:53 |
| Yanyan | hi | 01:53 |
| Qiming | you may want to remove the [WIP] tag? | 01:53 |
| Yanyan | oh, right | 01:53 |
| openstackgerrit | Yanyan Hu proposed stackforge/senlin: Start point of Senlin functional tests https://review.openstack.org/208355 | 01:53 |
| *** mathspanda has joined #senlin | 02:03 | |
| Yanyan | hi, Qiming, I found in lower version(e.g. 0.5.0) openstack-sdk, list related function methods like trusts() doesn't accept query parameters | 02:04 |
| Yanyan | but we have used this parameter in one method of our keystone_v3 driver http://git.openstack.org/cgit/stackforge/senlin/tree/senlin/drivers/openstack/keystone_v3.py#n113 | 02:05 |
| Qiming | right | 02:05 |
| Qiming | we don't use lower versions | 02:05 |
| Yanyan | so we a new user send request to senlin at first time, sdk will complain TypeError: trusts() got an unexpected keyword argument 'trustor_user_id' | 02:05 |
| Yanyan | yes | 02:05 |
| Qiming | this has been discussed several times | 02:05 |
| Yanyan | but if we fresh install Senlin | 02:05 |
| Yanyan | default version of sdk seems 0.5.0 | 02:06 |
| Qiming | I believe lixinhui has been caught be the problem for quite some while | 02:06 |
| Qiming | yes | 02:06 |
| Yanyan | em | 02:06 |
| Qiming | since sdk is releasing its 0.6.0 or 1.0 pretty soon | 02:06 |
| Yanyan | so maybe we can give some hints there to make user know this issue? | 02:06 |
| Qiming | this is not an user issue | 02:07 |
| Qiming | we are not there yet | 02:07 |
| Qiming | the sdk case is special | 02:07 |
| Yanyan | after the sdk requirement updated, this problem will not be there | 02:07 |
| Qiming | right | 02:07 |
| Yanyan | ok, so we just keep this in mind and wait for requirement update | 02:08 |
| Qiming | for the moment, we will just manually update the sdk version, i.e. do a reinstall of sdk from git | 02:08 |
| Qiming | yes | 02:08 |
| Qiming | we won't go back to 5.0 | 02:08 |
| Yanyan | ok | 02:08 |
| Qiming | it is now history | 02:08 |
| Yanyan | hmm, usually, I use 0.6.0, just becuase I reinstalled my system | 02:08 |
| Yanyan | found this issue | 02:08 |
| Qiming | ok | 02:10 |
| openstackgerrit | Yanyan Hu proposed stackforge/python-senlinclient: Using list query arguments as a single dict input parameter https://review.openstack.org/201946 | 02:19 |
| ChrisSen | The creation of a cluster with a specified capacity greater than 0 failed with error: | 02:21 |
| ChrisSen | http://paste.openstack.org/show/410422/ | 02:21 |
| Qiming | ChrisSen, it seems that your image was not found | 02:22 |
| Qiming | it is a bug reported several days ago and fixed | 02:22 |
| openstackgerrit | Yanyan Hu proposed stackforge/python-senlinclient: Using list query arguments as a single dict input parameter https://review.openstack.org/201946 | 02:23 |
| Qiming | Yanyan, you believe 201946 is the right way? | 02:26 |
| Qiming | previously, we have been fixing it in a different way: https://review.openstack.org/#/c/205759/1/senlinclient/v1/client.py | 02:26 |
| Yanyan | hi, Qiming, seems so | 02:28 |
| Yanyan | I found in 0.5.2, the queries has been used as an individual inpurt parameter 'params' | 02:28 |
| Qiming | instead of doing this individually, we were trying to do it in the client module, true? | 02:29 |
| Yanyan | oh, right | 02:30 |
| Yanyan | my mistake | 02:30 |
| Yanyan | forget that patch ;) | 02:30 |
| Qiming | ChrisSen, the problem you reported has been fixed before, I think | 02:33 |
| Qiming | please refer to this: https://review.openstack.org/#/c/206336/ | 02:33 |
| Yanyan | hi, Qiming, I have prepared the patch to the gate, will you guys check it before I propose it? | 02:38 |
| Qiming | just propose it | 02:40 |
| Yanyan | ok | 02:40 |
| *** elynn has joined #senlin | 02:43 | |
| Yanyan | hi, Qiming, the patch is here https://review.openstack.org/#/c/209736/ | 02:45 |
| Yanyan | please help to review, thanks | 02:45 |
| Yanyan | BTW, how can I add dependency between two patches in different project | 02:46 |
| Yanyan | I think this patch should depend on the patch that prvoides senlin local support for functional test | 02:46 |
| Qiming | okay, approved local changes | 02:48 |
| Yanyan | ok, thanks | 02:48 |
| mathspanda | Hi, please look at this paste: http://paste.openstack.org/show/410563/ | 02:48 |
| Yanyan | just refered to the patches which add sdk and ceilometer's functional test to do this job | 02:49 |
| Qiming | mathspanda, right | 02:50 |
| Qiming | nova does have some expectations from the user-data | 02:50 |
| Qiming | maybe we should base64 encode the user-data property before handling it to nova | 02:51 |
| Qiming | mathspanda, can you have file a bug and propose a patch? | 02:51 |
| mathspanda | yes, i think so. | 02:51 |
| mathspanda | OK. | 02:52 |
| Qiming | umm. Nova sometimes doesn't speak English | 02:52 |
| openstackgerrit | Merged stackforge/senlin: Start point of Senlin functional tests https://review.openstack.org/208355 | 02:54 |
| *** Zhenqi has joined #senlin | 03:03 | |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Ceilometer alarm trigger support https://review.openstack.org/209494 | 03:11 |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Engine support for triggers https://review.openstack.org/209545 | 03:20 |
| openstackgerrit | LinPeiyu proposed stackforge/senlin: B64encode user_data content before handling it to nova https://review.openstack.org/209754 | 04:48 |
| ChrisSen | Qiming, the creation of a cluster of heat stacks with a capacity greater than 0 also failed with error. Is that the same issue or something else? | 05:05 |
| ChrisSen | http://paste.openstack.org/show/410894/ | 05:05 |
| Qiming | seems something different | 05:07 |
| Qiming | it looks like you dont have Heat installed and configured | 05:07 |
| openstackgerrit | LinPeiyu proposed stackforge/senlin: B64encode user_data content before handling it to nova https://review.openstack.org/209754 | 05:47 |
| openstackgerrit | xu-haiwei proposed stackforge/senlin: Add test for heat_v1 driver https://review.openstack.org/209775 | 06:28 |
| openstackgerrit | xu-haiwei proposed stackforge/senlin: Add test for heat_v1 driver https://review.openstack.org/209775 | 06:38 |
| *** ChrisSen has quit IRC | 07:02 | |
| *** elynn has quit IRC | 07:05 | |
| openstackgerrit | Yanyan Hu proposed stackforge/senlin: Enable SenlinTest profile type before starting functional test https://review.openstack.org/209804 | 07:59 |
| openstackgerrit | Yanyan Hu proposed stackforge/senlin: Enable os.senlin.test profile type for functional test https://review.openstack.org/209804 | 07:59 |
| openstackgerrit | Yanyan Hu proposed stackforge/senlin: Enable os.senlin.test profile type for functional test https://review.openstack.org/209804 | 08:01 |
| *** ChrisSen has joined #senlin | 08:08 | |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Polish the nova server profile unit test https://review.openstack.org/209807 | 08:11 |
| ChrisSen | I tried to delete a nova server node that was not assigned to a cluster. But the deletion didn't succeed and an error in senlin engine occurred: | 08:12 |
| ChrisSen | http://paste.openstack.org/show/411080/ | 08:12 |
| openstackgerrit | Merged stackforge/senlin: Ceilometer alarm trigger support https://review.openstack.org/209494 | 08:15 |
| *** heyongli has quit IRC | 08:19 | |
| *** heyongli has joined #senlin | 08:19 | |
| Qiming | ChrisSen, it looks like our recent revision to exception handling is breaking some cases here | 08:19 |
| xuhaiwei | Qiming, you mean the client side failed? | 08:22 |
| *** Tennyson has quit IRC | 08:22 | |
| Qiming | no | 08:22 |
| Qiming | ChrisSen, you may want to check this: http://git.openstack.org/cgit/stackforge/senlin/tree/senlin/drivers/openstack/nova_v2.py#n122 | 08:23 |
| Qiming | add some LOG.error() calls there and see if 1) 'delete_server' was successful 2) sleep logic worked or not | 08:24 |
| Qiming | and 3) whether line 126 was hit | 08:24 |
| Qiming | also ... when debugging, you will want to check the log from senlin-engine/senlin-api | 08:25 |
| Qiming | not just from the client side outputs | 08:25 |
| Qiming | sometimes, client side messages won't give you a lot hints | 08:25 |
| xuhaiwei | really strange, this should be a bug newly born | 08:27 |
| Qiming | right, xuhaiwei | 08:27 |
| Qiming | need to find out the root cause | 08:28 |
| Qiming | it could be that the server is already gone | 08:28 |
| Qiming | the server deletion took too long | 08:28 |
| Qiming | the config file doesn't have a reasonable timeout value | 08:28 |
| xuhaiwei | yes, seems so, server is deleted, bug node is still there | 08:28 |
| xuhaiwei | bug/but/ | 08:29 |
| Qiming | the delete-and-wait logic didn't work as expected | 08:29 |
| Qiming | that's why I'm asking ChrisSen to do some investigation there | 08:29 |
| xuhaiwei | Qiming, I got the reason | 08:35 |
| xuhaiwei | ChrisSen, are you also investigating? | 08:37 |
| Qiming | xuhaiwei, I don't think so | 08:47 |
| Qiming | if you have found the root cause, please propose a fix | 08:47 |
| xuhaiwei | ok | 08:47 |
| *** Zhenqi has quit IRC | 09:09 | |
| xuhaiwei | Qiming | 09:18 |
| xuhaiwei | want to talk about the bug with you | 09:18 |
| Qiming | okay | 09:18 |
| xuhaiwei | the reason is that we translated the bug from SDK to HTTP exceptions already, and we still want to catch SDK exception type | 09:20 |
| xuhaiwei | sorry, again | 09:20 |
| xuhaiwei | the reason is that we translated the exception from SDK to HTTP exceptions already, and we still want to catch SDK exception type | 09:20 |
| xuhaiwei | http://git.openstack.org/cgit/stackforge/senlin/tree/senlin/drivers/openstack/nova_v2.py#n127 | 09:21 |
| xuhaiwei | the sdk.exc.NotFoundException has been translated to HTTPNotFound by the decorator | 09:21 |
| Qiming | I see | 09:23 |
| xuhaiwei | an earthquake just happened here | 09:23 |
| Qiming | it is not a surprise for you, ;) | 09:23 |
| xuhaiwei | so we should just catch Exception instead of SDK type exception? | 09:24 |
| xuhaiwei | just about 3 or 4 degree I think | 09:24 |
| Yanyan | @@ | 09:25 |
| Qiming | how about this | 09:25 |
| Qiming | we remove the "wait" logic from the driver layer | 09:25 |
| Qiming | implement it in the profile layer | 09:25 |
| Qiming | then we keep the driver clean | 09:26 |
| xuhaiwei | that is one way, I think | 09:26 |
| Qiming | and we won't have be in the situation having a decorated method calling another decorated method | 09:26 |
| xuhaiwei | but I am afraid we will meet the similar problems again for some other drivers | 09:26 |
| Qiming | it is not that unsual to have a profile do wait_for_status() kind of thing | 09:27 |
| xuhaiwei | it seems we have made things more complicated by ourselves | 09:27 |
| Qiming | for example, we have been doing that in the stack profile | 09:27 |
| Qiming | it is just an architecture design decision | 09:27 |
| xuhaiwei | I will think about it | 09:28 |
| Qiming | deletion is special | 09:28 |
| Qiming | for other cases, we can leverage the sdk's support | 09:29 |
| Qiming | http://git.openstack.org/cgit/stackforge/python-openstacksdk/tree/openstack/resource.py#n939 | 09:29 |
| Qiming | xuhaiwei, I'm trying to propose a fix of this | 09:36 |
| xuhaiwei | ok | 09:36 |
| Qiming | please help review later | 09:36 |
| xuhaiwei | yap | 09:36 |
| Qiming | meanwhile, please help review this: https://review.openstack.org/209807 | 09:39 |
| Qiming | or else I have to construct a dependency chain | 09:39 |
| Qiming | the nova server profile test case has been revised | 09:39 |
| *** mathspanda has quit IRC | 09:41 | |
| xuhaiwei | ok | 09:49 |
| xuhaiwei | see you tomorrow guys | 10:22 |
| Qiming | we really need to push on reviews | 10:27 |
| Qiming | having dependencies is really a nightmare | 10:27 |
| Qiming | 6320b51 (HEAD, wait-server-delete) Revise server delete-wait logic | 10:27 |
| Qiming | 035aad8 Ceilometer alarm trigger support | 10:27 |
| Qiming | d059b98 (tune-nova-prof-test) Polish the nova server profile unit test | 10:27 |
| Qiming | Do you really want to submit the above commits? | 10:27 |
| Qiming | Type 'yes' to confirm, other to cancel: yes | 10:27 |
| Qiming | remote: Resolving deltas: 100% (19/19) | 10:27 |
| Qiming | remote: Processing changes: refs: 1, done | 10:27 |
| Qiming | To https://tengqm:ZcK4qjpWkuIY@review.openstack.org/stackforge/senlin.git | 10:28 |
| Qiming | ! [remote rejected] HEAD -> refs/publish/master/wait-server-delete (change 209494 closed) | 10:28 |
| Qiming | error: failed to push some refs to 'https://tengqm:ZcK4qjpWkuIY@review.openstack.org/stackforge/senlin.git' | 10:28 |
| Qiming | every time you push something .... | 10:28 |
| *** Yanyan has quit IRC | 10:28 | |
| *** Qiming has quit IRC | 10:49 | |
| *** Qiming has joined #senlin | 11:48 | |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Revise server delete-wait logic https://review.openstack.org/209899 | 11:51 |
| *** ChrisSen has quit IRC | 11:56 | |
| *** lkarm has joined #senlin | 12:04 | |
| *** jdandrea has joined #senlin | 12:30 | |
| openstackgerrit | Merged stackforge/senlin: Polish the nova server profile unit test https://review.openstack.org/209807 | 13:22 |
| *** jhroyal has joined #senlin | 13:36 | |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Fix coverage test https://review.openstack.org/209936 | 13:39 |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Added one unit test for webhook store https://review.openstack.org/209937 | 13:44 |
| *** jhroyal1 has joined #senlin | 13:59 | |
| *** jhroyal has quit IRC | 14:02 | |
| *** mathspanda has joined #senlin | 14:04 | |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Add unit test for different HM type https://review.openstack.org/209950 | 14:16 |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Fix alarm trigger validation https://review.openstack.org/209953 | 14:29 |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Fix deletion policy candidate selection https://review.openstack.org/209958 | 14:33 |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Unit test for base policy module https://review.openstack.org/209968 | 14:51 |
| *** jhroyal1 has left #senlin | 14:56 | |
| *** jhroyal1 has joined #senlin | 14:56 | |
| *** jhroyal1 is now known as jhroyal | 14:57 | |
| openstackgerrit | Qiming Teng proposed stackforge/senlin: Delete the AES based crypto module https://review.openstack.org/209974 | 15:00 |
| *** Qiming has quit IRC | 15:16 | |
| *** Qiming has joined #senlin | 15:35 | |
| *** Qiming has quit IRC | 15:48 | |
| *** mathspanda has quit IRC | 16:14 | |
| *** jdandrea has quit IRC | 16:38 | |
| *** jdandrea has joined #senlin | 16:39 | |
| *** jruano has joined #senlin | 19:41 | |
| *** jruano has quit IRC | 20:01 | |
| *** lkarm has quit IRC | 20:58 | |
| *** jhroyal has quit IRC | 21:33 | |
| openstackgerrit | Merged stackforge/senlin: Added one unit test for webhook store https://review.openstack.org/209937 | 23:18 |
| *** Qiming has joined #senlin | 23:36 | |
| openstackgerrit | Merged stackforge/senlin: Delete the AES based crypto module https://review.openstack.org/209974 | 23:44 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!