| *** dbpiv has joined #ara | 02:29 | |
| *** dbpiv has quit IRC | 03:57 | |
| *** jparrill has joined #ara | 05:58 | |
| *** herald85 has joined #ara | 06:43 | |
| *** krion has quit IRC | 07:01 | |
| *** krion has joined #ara | 07:03 | |
| *** micisuta has joined #ara | 07:53 | |
| *** idir___ has joined #ara | 08:04 | |
| *** herald85 has quit IRC | 08:55 | |
| *** idir___ has quit IRC | 11:36 | |
| *** idir___ has joined #ara | 12:04 | |
| idir___ | hi, i have this warning { Timeout value connect was 30, but it must be an int, float or None. } when iu'm trying to define the api timeout in ansible.cfg or as a env variable | 12:05 |
|---|---|---|
| *** dbpiv has joined #ara | 12:34 | |
| idir___ | dmsimard : The problem was the proxy ! | 13:07 |
| dmsimard | idir___: the proxy ? | 13:08 |
| cyberpear | it's always the proxy! | 13:09 |
| idir___ | yes i have a proxy localhost:port wihxh timed out after 30 | 13:09 |
| dmsimard | oh, okay, so it works now | 13:09 |
| dmsimard | ? | 13:09 |
| cyberpear | for good measure, I always `export no_proxy=corp.example.com,0,1,2,[...],254,255` to ensure nothing goes to the proxy that shouldn't, including bare IPs | 13:10 |
| dmsimard | for the timeout, it might be because the value is not typed as an integer here https://github.com/ansible-community/ara/blob/9065642a67013b97b829dd82df90f80f98187fc7/ara/plugins/callback/ara_default.py#L90-L97 | 13:10 |
| dmsimard | it's an easy fix if you'd like to send a patch :D | 13:10 |
| idir___ | i don't think it's can be a patch because it not a bug | 13:11 |
| dmsimard | need to add " type: integer " to the timeout documentation block ^ | 13:11 |
| idir___ | i don't think it's bug but, i found that, we have to extend the time out the first time when executing a playbook | 13:13 |
| idir___ | i tried this, but it doest work because of type | 13:13 |
| idir___ | i used no proxy to fix this with domaine name | 13:14 |
| idir___ | but there something else | 13:14 |
| dmsimard | idir___: if you're going over the 30 seconds timeout, there is something very wrong | 13:15 |
| idir___ | yes i think | 13:15 |
| dmsimard | but the "timeout must be int" bug can be fixed by adding type: integer in that area I linked | 13:15 |
| idir___ | now i received to reporting bur it's not a complet reporting and the warning is | 13:16 |
| idir___ | " [WARNING]: Failure using method (v2_playbook_on_play_start) in callback plugin (<ansible.plugins.callback./opt/jenkins- slave/mdsa403/workspace/ara_reporting/ansible/lib/python3.6/site- packages/ara/plugins/callback/ara_default.CallbackModule object at 0x7f13796b4780>): 'ascii' codec can't decode byte 0xe2 in position 11: ordinal not in range(128)" | 13:16 |
| dmsimard | yes, that's an encoding issue -- I would like to be able to reproduce that issue so we can fix it | 13:17 |
| idir___ | and this one too, when trying to getfacts | 13:17 |
| idir___ | [WARNING]: Failure using method (v2_playbook_on_task_start) in callback plugin (<ansible.plugins.callback./opt/jenkins- slave/mdsa403/workspace/ara_reporting/ansible/lib/python3.6/site- packages/ara/plugins/callback/ara_default.CallbackModule object at 0x7f13796b4780>): 'NoneType' object is not subscriptable | 13:17 |
| dmsimard | there will be a lot of cascading failures if something fails early on | 13:17 |
| dmsimard | because of relationships (playbook -> play -> task -> result <- host) so if you can't create the playbook, everything else will fail | 13:18 |
| dmsimard | we need to investigate the first failure | 13:18 |
| dmsimard | we need better exception handling too, but yeah | 13:18 |
| idir___ | okey | 13:18 |
| idir___ | i will investigate the bug and tell you if i found something | 13:20 |
| dmsimard | encoding errors are typically due to non-ascii characters somewhere (éè€, etc.) | 13:21 |
| idir___ | yes, but where lol | 13:22 |
| dmsimard | in file paths ? inside playbook / role files ? | 13:23 |
| dmsimard | I also have this "where" question, btw, I'd like to fix the issue :p | 13:23 |
| idir___ | hhh i'm joking bro | 13:24 |
| idir___ | the char is this one "â" | 13:25 |
| *** gvincent has quit IRC | 13:29 | |
| idir___ | this char is not used, may be in a comments but, we never use it elsewhere | 13:31 |
| dmsimard | out of curiosity | 13:32 |
| dmsimard | can you try a basic playbook without and with the character to see if it's really that ? | 13:32 |
| idir___ | it complicated because i'm using jenkins getting playbooks from git | 13:40 |
| idir___ | entreprise work! | 13:40 |
| idir___ | but i'll something | 13:40 |
| idir___ | try | 13:40 |
| pabelanger | dmsimard: my hunch is https://logs.zuul.ansible.com/89/57789/8d9f8e0547417362c0241ab039e360035b778478/third-party-check-silent/ansible-test-network-integration-ios-python27/b35b883/job-output.html#l3301 is the reason ara bombs on encoding. I should have database in next hour. | 13:48 |
| *** gvincent has joined #ara | 13:49 | |
| pabelanger | dmsimard: https://logs.zuul.ansible.com/89/57789/8d9f8e0547417362c0241ab039e360035b778478/third-party-check-silent/ansible-test-network-integration-ios-python27/8ab0a7b/controller/ara-report/ansible.sqlite | 14:04 |
| pabelanger | that should have bad data | 14:04 |
| pabelanger | I'll look more into it shortly | 14:04 |
| pabelanger | dmsimard: it also is likely limted to python27 | 14:05 |
| pabelanger | limited* | 14:05 |
| openstackgerrit | David Moreau Simard proposed recordsansible/ara master: Make sure the API client timeout is an integer and not a string https://review.opendev.org/666846 | 14:15 |
| openstackgerrit | David Moreau Simard proposed recordsansible/ara master: ara-api role: Set proper path for the gunicorn pid file https://review.opendev.org/666848 | 14:16 |
| dmsimard | idir___: fix for the timeout bug: https://review.opendev.org/666846 | 14:17 |
| dmsimard | pabelanger: so this one then ? https://github.com/ansible/ansible/blob/6b162142a7e5d911f72f23a21abc0bfd4b946355/test/integration/targets/ios_file/tests/cli/net_put.yaml#L53 | 14:21 |
| pabelanger | dmsimard: yah, think so | 14:25 |
| idir___ | dmsimard, you think thatn this can solve the problem of the proxy? | 14:28 |
| dmsimard | idir___: no, that has nothing to do with the proxy | 14:29 |
| idir___ | ok, does the fix available on pypi?; | 14:29 |
| dmsimard | idir___: it will be available on pypi when the patch merges and is released | 14:43 |
| idir___ | ok, i tried another playbook and i have same error with ascii | 14:46 |
| *** jungleslow has quit IRC | 14:48 | |
| dmsimard | idir___: and without the special character it works ? | 14:50 |
| idir___ | the problem is : there isn't special characters | 14:52 |
| idir___ | in all my files : playbooks - roles- inventories - group_vars - host_vars | 14:52 |
| dmsimard | what about the filesystem path in which these files are located ? | 14:52 |
| dmsimard | is there like /home/école/something/playbook.yml | 14:53 |
| idir___ | the files are clone from git | 14:53 |
| idir___ | cloned | 14:53 |
| idir___ | directely in a jenkins slave | 14:54 |
| dmsimard | idir___: so what's the full path to the playbook file for example ? | 14:55 |
| idir___ | qdsq | 14:57 |
| idir___ | "/opt/jenkins-slave/mdsa403/workspace/ara_reporting/deploy_sonar.yml" | 14:57 |
| openstackgerrit | David Moreau Simard proposed recordsansible/ara master: Relocate api_root_dir outside of /tmp for integration tests https://review.opendev.org/666859 | 15:05 |
| openstackgerrit | David Moreau Simard proposed recordsansible/ara master: Relocate api_root_dir outside of /tmp for integration tests https://review.opendev.org/666859 | 15:06 |
| dmsimard | pabelanger: looking now | 15:14 |
| dmsimard | pabelanger: hum, that database looks broken to me | 15:15 |
| dmsimard | I think the webserver is presenting it as a string instead of a binary file | 15:16 |
| dmsimard | needs to be binary | 15:17 |
| idir___ | dmsimard : do you think that it can be a ansible bugs | 15:19 |
| openstackgerrit | Merged recordsansible/ara master: ara-api role: Set proper path for the gunicorn pid file https://review.opendev.org/666848 | 15:19 |
| dmsimard | idir___: hard to tell without a reliable way to reproduce the issue | 15:19 |
| dmsimard | for example we have tests for those non-ascii characters such as https://github.com/ansible-community/ara/blob/9065642a67013b97b829dd82df90f80f98187fc7/tests/integration/smoke.yaml#L36-L66 | 15:20 |
| dmsimard | but it seems you are hitting an issue not covered by tests | 15:20 |
| dmsimard | so I want to know what exactly is going on, add a test to reproduce the issue and then fix it | 15:21 |
| openstackgerrit | David Moreau Simard proposed recordsansible/ara master: Relocate api_root_dir outside of /tmp for integration tests https://review.opendev.org/666859 | 15:21 |
| openstackgerrit | David Moreau Simard proposed recordsansible/ara master: Add distributed_sqlite db backend and WSGI app https://review.opendev.org/665323 | 15:22 |
| idir___ | when i execute from the host directly whitout using jenkins it's good | 15:23 |
| dmsimard | so you need to find out what is causing that | 15:23 |
| dmsimard | need to compare the environment where the issue reproduces and the environment where it doesn't reproduce | 15:24 |
| dmsimard | for example, is /etc/locale.conf the same ? I dunno | 15:24 |
| dmsimard | for example I have LANG="en_CA.UTF-8" | 15:25 |
| idir___ | in dev hosts where i execute manualy it ok | 15:27 |
| idir___ | i'll check ppr hosts, i don't have access | 15:28 |
| idir___ | rightnow | 15:28 |
| pabelanger | dmsimard: oh, possible. There might be something wrong with htmlify logs and swift | 15:28 |
| openstackgerrit | Merged recordsansible/ara master: Make sure the API client timeout is an integer and not a string https://review.opendev.org/666846 | 15:30 |
| openstackgerrit | Merged recordsansible/ara master: Add toggle to disable sql migrations for the offline client https://review.opendev.org/666723 | 15:30 |
| *** ktibi has joined #ara | 15:41 | |
| *** idir___ has quit IRC | 16:13 | |
| *** micisuta has quit IRC | 16:21 | |
| *** ktibi has quit IRC | 16:22 | |
| openstackgerrit | Merged recordsansible/ara master: Relocate api_root_dir outside of /tmp for integration tests https://review.opendev.org/666859 | 16:34 |
| *** gvincent has quit IRC | 17:20 | |
| openstackgerrit | David Moreau Simard proposed recordsansible/ara master: Run actual integration tests for the ara-api role jobs https://review.opendev.org/666894 | 17:35 |
| *** etienne has joined #ara | 18:22 | |
| *** micisuta has joined #ara | 18:25 | |
| *** micisuta1 has joined #ara | 18:28 | |
| *** micisuta has quit IRC | 18:30 | |
| *** micisuta1 is now known as micisuta | 18:30 | |
| *** dbpiv has quit IRC | 20:49 | |
| *** openstackgerrit has quit IRC | 21:33 | |
| *** micisuta has quit IRC | 22:03 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!