| *** caphrim007 has quit IRC | 00:03 | |
| *** caphrim007 has joined #openstack-jjb | 00:04 | |
| *** caphrim007 has quit IRC | 00:05 | |
| *** caphrim007 has joined #openstack-jjb | 00:06 | |
| *** caphrim007 has quit IRC | 00:11 | |
| *** caphrim007 has joined #openstack-jjb | 01:20 | |
| *** caphrim007 has quit IRC | 01:39 | |
| openstackgerrit | Kien Ha proposed openstack-infra/jenkins-job-builder master: Add depth option to git shallow-clone https://review.openstack.org/425042 | 02:08 |
|---|---|---|
| openstackgerrit | Kien Ha proposed openstack-infra/jenkins-job-builder master: Add publishers related to GitLab Plugin https://review.openstack.org/518618 | 03:04 |
| openstackgerrit | Merged openstack-infra/jenkins-job-builder master: Add depth option to git shallow-clone https://review.openstack.org/425042 | 03:45 |
| *** caphrim007 has joined #openstack-jjb | 04:59 | |
| *** caphrim007 has quit IRC | 05:03 | |
| *** caphrim007 has joined #openstack-jjb | 05:17 | |
| *** hashar has joined #openstack-jjb | 07:51 | |
| *** electrofelix has joined #openstack-jjb | 09:03 | |
| beiske | zxiiro: Is everything good for release today? | 13:22 |
| zxiiro | beiske: i just woke up. I'll do a release later today | 13:36 |
| beiske | Thank you! | 13:36 |
| zxiiro | ssbarnea: electrofelix any patches we want merged before I cut a release? | 13:37 |
| ssbarnea | zxiiro: have you tested the runscript fix? | 13:37 |
| zxiiro | ssbarnea: yes. I need to fix your suggestion to not use replace though. I'll patch that up in a bit | 13:38 |
| ssbarnea | do it, i will also test it on two of my servers and lets release after that. | 13:39 |
| electrofelix | zxiiro: I think the main two patches I had around fixing the exceptions in python-jenkins have already landed | 13:44 |
| zxiiro | just remembered I have a meeting now. I'll do the patch after the meeting | 14:00 |
| *** caphrim007 has quit IRC | 14:04 | |
| *** caphrim007_ has joined #openstack-jjb | 14:07 | |
| *** caphrim007_ has quit IRC | 14:11 | |
| *** caphrim007 has joined #openstack-jjb | 14:13 | |
| *** caphrim007 has quit IRC | 14:17 | |
| *** tanhengyeow has joined #openstack-jjb | 14:53 | |
| openstackgerrit | Tan Heng Yeow proposed openstack-infra/jenkins-job-builder master: Improve support for EnvInject plugin https://review.openstack.org/570664 | 15:12 |
| openstackgerrit | Thanh Ha proposed openstack/python-jenkins master: Fix run_scripts() API https://review.openstack.org/574302 | 15:38 |
| zxiiro | ssbarnea: sorry for the delay. I just updated the patch. | 15:38 |
| zxiiro | ssbarnea: I'm gonna release jenkins-job-builder. you want to do python-jenkins? | 15:53 |
| openstackgerrit | Merged openstack-infra/jenkins-job-builder master: Add support for GraphView column to views https://review.openstack.org/574670 | 15:53 |
| ssbarnea | nope, got busy with something here | 15:54 |
| zxiiro | ssbarnea: ok i can do it | 15:55 |
| zxiiro | looks like there's only 2 patches from electrofelix in the release so should be pretty safe | 15:57 |
| ssbarnea | zxiiro: i still get exception while trying to run script | 15:58 |
| zxiiro | huh very odd... | 15:59 |
| ssbarnea | https://gist.github.com/ssbarnea/54fdae90acbb43064b61c31bfe5480b8 | 16:01 |
| ssbarnea | you do have access to the same machine | 16:01 |
| *** caphrim007 has joined #openstack-jjb | 16:02 | |
| zxiiro | ssbarnea: yeah | 16:02 |
| zxiiro | ssbarnea: I tried against that Jenkins server last week | 16:02 |
| ssbarnea | and to be sure, your CR is downloaded in python-jenkins repo, but this is obvious as you would not get the exception otherwise. | 16:03 |
| ssbarnea | sad that I don't get anything visible in jenkins ocnsole. | 16:03 |
| zxiiro | ssbarnea: yep. I've even confirmed via the magic string too so I know it's running that code | 16:03 |
| zxiiro | ssbarnea: are you using api token or password to authenticate? | 16:04 |
| zxiiro | i was using api token | 16:04 |
| ssbarnea | yep, inside .netrc | 16:04 |
| zxiiro | ssbarnea: hmm maybe try without netrc just to see? | 16:04 |
| ssbarnea | other requests are working normally. is there something special about run script? as opposed to the other http requests? | 16:05 |
| zxiiro | ssbarnea: Not sure. I don't know this code base very well | 16:05 |
| *** caphrim007 has quit IRC | 16:06 | |
| *** caphrim007 has joined #openstack-jjb | 16:06 | |
| zxiiro | beiske: JJB 2.0.10 is released in case you're not following the mailing list | 16:06 |
| *** caphrim007 has quit IRC | 16:06 | |
| *** caphrim007 has joined #openstack-jjb | 16:06 | |
| ssbarnea | zxiiro: found the issue! maybe is partially my fault. | 16:08 |
| ssbarnea | if you check my example is using HTTP, but server is configured to redirect http to https. | 16:09 |
| ssbarnea | well, this redirection works well for most requests, but not for the runScript one. | 16:09 |
| ssbarnea | i enabled debugging logging and this is now visible. | 16:09 |
| ssbarnea | i updated the gist with debug info, if I switch to https it works. | 16:10 |
| ssbarnea | still, it is supposed to work with both, so we need to fix it. | 16:11 |
| electrofelix | ssbarnea: I believe that we haven't configured requests to follow redirects | 16:11 |
| ssbarnea | based on requests docs, allow_redirects defaults to True -http://docs.python-requests.org/en/master/api/ | 16:12 |
| electrofelix | oh, I thought it was set to false? | 16:13 |
| electrofelix | ah, I see, it's disabled for HEAD requests | 16:14 |
| zxiiro | ssbarnea: ah makes sense, cool. I was starting to think it was voodoo magic | 16:14 |
| ssbarnea | well, i have haproxy in front of it | 16:15 |
| ssbarnea | usually i am using nginx, but a week ago I decided to transform my pfsense into a SSL load-balancer | 16:15 |
| ssbarnea | is this on my haproxy? as i do not see anything related to this on requests. | 16:18 |
| *** r-mibu has joined #openstack-jjb | 16:21 | |
| openstackgerrit | Tan Heng Yeow proposed openstack-infra/jenkins-job-builder master: Improve support for nodelabelparameter https://review.openstack.org/570659 | 16:22 |
| ssbarnea | tanhengyeow: out of curiosity, what is the purpose of label parameter? I only use node one where I put node name or label conditions like "pyslave && !master" | 16:26 |
| openstackgerrit | Tan Heng Yeow proposed openstack-infra/jenkins-job-builder master: Improve support for EnvInject plugin https://review.openstack.org/570664 | 16:30 |
| ssbarnea | zxiiro: i still dont know what is causing the failure with HTTP case and i would prefer to make it work. i doubt is related to my loadbalancer. | 16:45 |
| tanhengyeow | ssbarnea: From my understanding it is used to restrict where the job can be executed. I'm not using it though, saw that an user posted for this request @ https://storyboard.openstack.org/#!/story/2001287 and thought it might be a good opportunity to learn about implementing support for plugins :) | 16:45 |
| openstackgerrit | Tan Heng Yeow proposed openstack-infra/jenkins-job-builder master: Improve support for nodelabelparameter https://review.openstack.org/570659 | 16:51 |
| zxiiro | ssbarnea: ah yeah, we should look into fixing it in another patch | 17:01 |
| ssbarnea | it affects only this command, so it should be related. | 17:12 |
| zxiiro | ssbarnea: yeah but it's an issue unrelated to this patch imo. | 17:13 |
| zxiiro | ssbarnea: do we know that it's caused by requests? or has this issue always been an issue? | 17:13 |
| ssbarnea | i am going to some tests now | 17:24 |
| ssbarnea | first, the issue reproduces with other servers which are using different frontends. | 17:27 |
| ssbarnea | probably we could avoid it by detecting the url and replacing it with correct one, maybe even displaying a warning. | 17:28 |
| ssbarnea | this would speedup the library a LOT because at this moment all requests are redirects, so double the traffic. | 17:28 |
| zxiiro | that would be good | 17:36 |
| electrofelix | ssbarnea: wonder if we should log an issue with python-requests that for Session objects it should obey permanent redirects for the rest of the sessons | 17:43 |
| ssbarnea | good question, but i doubt it will be accepted as some people are redirecting based on requeted url, so it may break their "redirector" implementations. | 17:44 |
| ssbarnea | one problem that I have is that by default instantiating the class does not perform any requests, so I am unable to know that current url will be redirected. | 17:45 |
| ssbarnea | i find the issue quite important because i am sure we will get bug reports about people encountering this issue, so i want to avoid this from happening. | 17:45 |
| ssbarnea | i would prefer a clear failure than having "some" request failing silently, it would be a time waster for everyone | 17:46 |
| ssbarnea | is not personal here, as I can always fix the url, is just about "UX" (python-jenkins user..) | 17:47 |
| electrofelix | in two minds, it's nice to be able to instantiate a class without it starting performing requests until you're ready for it to start talking to the remote, but maybe there is a call that we should expose and have JJB perform to log about checking the remote server access? | 17:50 |
| electrofelix | as an example of usage | 17:50 |
| electrofelix | on the redirect that might not support updating requests to do this automatically, but maybe they would add an option to allow it to be enabled | 17:54 |
| ssbarnea | apparently that part of HTTP spec, i found in several places that a redirect POST request must be transformed into a GET. | 17:57 |
| ssbarnea | this means that we must be sure we don't get redirected before doing this. | 17:57 |
| ssbarnea | i think it has to do with security | 17:58 |
| electrofelix | ah, yeah, curl has a --post301 --post302 & --post303 to override that behaviour | 17:58 |
| ssbarnea | would you find a problem if we include an implicit request when we instantiate the class? | 17:58 |
| ssbarnea | if we try to connect to the server as soon we create the object, we should be able to detect the redirection and if the url parameter, also throwing a warning. | 17:59 |
| electrofelix | I'd certainly advise a conservative approach to it, if it can be established without using a request that requires authentication that would be preferable | 18:05 |
| *** electrofelix has quit IRC | 18:05 | |
| openstackgerrit | Tan Heng Yeow proposed openstack-infra/jenkins-job-builder master: Improve support for nodelabelparameter https://review.openstack.org/570659 | 18:15 |
| openstackgerrit | Sorin Sbarnea proposed openstack/python-jenkins master: detect and respect http redirects https://review.openstack.org/575817 | 18:28 |
| ssbarnea | zxiiro: ^^ this needs to be tested so we don't break some authentication types. | 18:30 |
| ssbarnea | but at least in my case it works really nice, doing what i wanted: correcting the url and telling consumer about that. | 18:31 |
| ssbarnea | and more important POST no longer fails silently. | 18:31 |
| *** tanhengyeow has quit IRC | 18:55 | |
| zxiiro | ssbarnea: cool. looking soon | 19:10 |
| zxiiro | ssbarnea: seems to work for me | 19:44 |
| zxiiro | ssbarnea: yeah, and works when I switch to your s-less http url too | 19:45 |
| *** hashar has quit IRC | 22:10 | |
| *** caphrim007_ has joined #openstack-jjb | 23:17 | |
| *** caphrim007 has quit IRC | 23:20 | |
| *** caphrim007_ has quit IRC | 23:20 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!