08:00:03 #startmeeting third-party 08:00:04 Meeting started Tue Dec 16 08:00:03 2014 UTC and is due to finish in 60 minutes. The chair is anteaya. Information about MeetBot at http://wiki.debian.org/MeetBot. 08:00:05 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 08:00:08 The meeting name has been set to 'third_party' 08:00:15 hello trinaths 08:00:23 hi anteaya 08:00:37 Howdy 08:00:42 hey jhesketh 08:01:12 anyone else here for the third-party meeting? 08:01:33 heyongli jyuso are you about? 08:01:47 whos here for the ambassador meeting? 08:01:54 anteaya: hello 08:02:00 anteaya: Hi 08:02:02 hi, anteaya , hi 08:02:13 hello all 08:02:18 hi all 08:02:20 sarob I guess we need to move the ambassador meeting to #openstack-meeting-3 08:02:25 hi,everyone 08:03:03 so as we are just getting our group started and getting to know each other, right now we don't have an agenda 08:03:24 does anyone have a place which they would like to begin? 08:03:58 I have some doubts in CI. 08:03:58 anteaya: yes, just want to get attention the a review by omrim for FAQ 08:04:05 great 08:04:13 who wants to go first? 08:04:30 trinaths: go ahead 08:04:41 nuritv: thank you. 08:04:43 Murad: there you are 08:04:49 yes I am 08:04:52 Hi everybody 08:05:06 hi Murad 08:06:07 trinaths: okay so whenever you are ready 08:06:45 I installed CI wil jaypipes documentation. In the CI Jenkins node, there are several git repos. and the same repos are existing with in the job nodes. Zuul-merger updates with the commit. how is that updated commit trasfered to job node? 08:07:13 okay good question 08:07:28 typo error: transfered 08:07:28 does anyone have a response to trinaths' question? 08:08:07 trinaths: the zuul-merger reports back a ref to zuul that is then passed to the job 08:08:35 so zuul then requests a worker to do a job with these params: http://ci.openstack.org/zuul/launchers.html#zuul-parameters 08:08:55 the Jenkins gearman plugin exposes those params as environment variables 08:09:08 the gerrit-git-prep.sh script looks for those params and sets up a working directory for the test 08:09:38 jhesketh: okay. 08:10:09 then why a git checkout is required in jenkins node? 08:10:20 i mean the master node 08:10:45 trinaths: hmm, it should only be the slave node.. but I don't run Jenkins so I'm not sure 08:10:55 my best guess is that it's not needed and there from a legacy setup 08:11:34 jhesketh: okay. I will refer again. 08:11:46 sorry I haven't got a specific answer there 08:12:32 jhesketh: I missed to answer you 08:13:11 trinaths: do you need anything more here? 08:13:52 or is it nuritv's turn? 08:14:03 anteaya: no. I have a que [2] When ever a job is trigger, devstack-gate updates all the projects, like, nova, neutron, cinder etc.. 08:14:40 this takes more time for the actual build to hapen and the test to run. How to simplify this? any suggestions 08:14:58 trinaths: shall we let nuritv have a turn? 08:15:12 anteaya: yes. Please go ahead 08:15:19 thank you 08:15:30 nuritv: let's hear from you now please 08:15:47 anteaya: thanks. i have several issues i would like to raise :) 08:16:14 * anteaya listens 08:16:18 first of all, i would like to get everyone attention to the FAQ section by omrim 08:16:35 there is still alot to do, but any more reviews will be helpful 08:16:43 https://review.openstack.org/#/c/141817/3 08:16:55 #link https://review.openstack.org/#/c/141817/3 08:17:16 thank you to omrim for offering this patch 08:17:51 pleia2 gave a pretty through review 08:18:03 so, if anyone has other FAQ that are worth mentioning - please post a comment 08:18:17 anteaya: yes, we all keep learning :) 08:18:33 :D 08:18:56 anteaya: can i continue to the next topic? 08:19:08 oh yes 08:20:05 thanks. so, i am concern about the big change happening on Neutron (spliiting the plugins), and wonder how it is going to affect CI and what we should do in advance 08:20:21 that is a good question 08:20:40 I don't think there is anyone here from Neutron right now 08:20:51 so I will go ahead and respond to that one 08:21:13 I don't know exactly how the plugin split will affect CI systems 08:21:36 however I was at the Neutron mid-cycle when the advanced services were split out 08:22:19 and the immediate effect of that split was that all CIs that tested advanced services were suggested to turn themselves off until the new advanced services were running again 08:22:53 since one of the consequences of the split was that right afterwards the testing on the split out advanced services would not work 08:23:03 "new advanced services" 08:23:16 so tests needed to be modified within openstack to get the infra jobs running again 08:23:40 and to be honest I have now lost track and I don't know if the tests on advance services are working yet 08:23:55 they were expected to be down until sometime this week 08:24:20 so it is possible that during the split itself the plugins might not be able to be tested 08:24:46 which should not be a problem since new code to the plugins should be frozen for the duration of the split 08:24:52 I heard tempests test will also be splitted. 08:24:58 as for how to prepare 08:25:20 I would suggest ensuring you read the meeting logs of any meetings that discuss the split 08:25:40 #link http://eavesdrop.openstack.org/ 08:25:56 this is where all openstack channel and meeting logs are accessed 08:26:00 help yourself 08:26:19 also, I would find Edgar Magma in #openstack-neutron 08:26:19 anteaya: my concern is more about ml2 md that will be out of tree. this would need us,as vendors, change code, git branch, etc. 08:26:38 it very well might require all of that 08:26:54 have you been able to attend the ml2 meetings at all? 08:27:08 anteaya: not yet. will start this week 08:27:25 okay, I would definitely suggest that 08:27:37 have you had a chance to look at the ml2 meeting logs? 08:27:59 anteaya: not yet. will start this week also :) 08:28:02 okay 08:28:08 so start there 08:28:21 also talk to Edgar 08:28:35 he answers a lot of third party CI questions for Neutron 08:28:47 he is emagma in irc 08:29:15 he might have some specifics 08:29:28 does that help? 08:29:36 yes! . will do. i guess it now just to keep track on what will be the decision. 08:29:47 okay great 08:29:51 and that brings me to my last Q 08:30:29 when ever our CI is down. is it better to turn it off and not vote at all, or jut mark it down at the Wiki. 08:30:37 both 08:30:53 anteaya:ok. thanks. 08:30:54 if your CI is not functioning as expected turn it off 08:31:18 there is nothing worse for your CI that to be non-functioning yet commenting on patches 08:31:30 as you train developers to ignore your system 08:31:49 since you are not monitoring when you are posting quality information or noise 08:31:58 so they will just assume you always post noise 08:32:04 however 08:32:04 nuritv: This the latest log of the last ML2 meeting http://eavesdrop.openstack.org/meetings/networking_ml2/2014/networking_ml2.2014-12-03-16.03.log.html held on Dec-3rd 08:32:16 if you monitor your system and its quality 08:32:27 and only post comments when your system is actually working 08:32:41 then developers will soon learn that your system can be trusted 08:33:03 and updating the wikipage helps people to find accurate status information on your system 08:33:23 got it! thanks 08:33:28 so again, it shows you are paying attention to how the output from your system is percieved by those using it 08:33:29 trinaths: thanks 08:33:32 thank you 08:34:05 nuritv: anything else right now? 08:34:37 anteaya: no. thanks 08:34:45 great 08:34:56 does anyone else have anything they wish to discuss? 08:35:04 I always found this in zuul log: zuul.Gearman: Looking for lost builds 08:35:13 hello Guest24100 08:35:20 Hi 08:35:24 are you able to introduce yourself? 08:35:29 I don't believe I know you 08:35:43 You can call me bob 08:35:52 okay bob, welcome 08:36:05 So what does that mean? 08:36:09 so you have a question about gearman 08:36:16 my question [2] 08:36:21 yes 08:36:23 does anyone have a response to bob's question? 08:36:24 zuul.Gearman: Looking for lost builds 08:36:44 Guest24100: I too have this log in my merger-debug.log 08:37:08 bob : have you tried to stop zuu and zuul-merger service and then rm -rf /var/log/zuul/* then restart the service and checked 08:37:25 hello nikil22, glad you made it 08:37:48 Yes, i have tried that 08:38:02 if i restart it, it will be ok 08:38:28 But I don't want always restart it 08:39:15 bob : oh ok , yes can't restart all time . I just got few errors similar and i just did above step and it worked 08:40:00 Thanks, I have not yet do the "rm -rf /var/log/zuul/*" 08:40:10 And i will try it 08:40:18 bob: thats just to clear logs but i guess it wont help 08:40:34 :-) 08:40:41 And I have questions related to devstack-gate. when does the function pre_test_hook works? Basically In CI system i want to enable the driver in neutron.conf and then start the service 08:41:04 let's ensure bob is statisfied before we move on 08:41:18 bob are we okay to move on, or do you have more? 08:41:37 No more, thanks! 08:41:42 great, thank you 08:41:50 :-) 08:41:53 nikil22: okay so your turn I believe 08:42:16 you have a question about the pre_test_hook funtion in devstack-gate 08:42:20 is that correct? 08:42:55 yes 08:43:09 does anyone have a reply to nikil22's question? 08:43:28 i guess that is the function to enable drivers and modify other configuration files related to specify plugin insted of default one 08:44:10 or perform any other commands to set up a specific environment 08:44:15 before the tests run 08:45:03 so it sounds like the pre_test_hook is a good place to begin your experiments in what you want to accomplish 08:45:26 did anyone else have anything more to add in response to this question? 08:46:14 nikil22: are you statisfied or do you need more? 08:47:10 * anteaya waits for nikil22 to reply 08:47:13 will the pre_test_hook will run before starting the porcess or after starting the neutron-server process. Because if we change the settings then we need to restart the services right 08:48:13 nikil22: I don't know exactly 08:48:36 but I would guess that the pre_test_hook runs before the services start 08:48:38 ok thanks i will do expriments then :-) 08:48:49 that sounds like a good way forward 08:48:58 nikil22: all done for this question? 08:49:45 * anteaya waits for nikil22 to reply 08:50:20 nikil22: can we move on? 08:50:31 yes 08:50:34 thank you 08:50:35 thanks 08:50:44 10 minutes remaining 08:50:55 who has something else they would like to discuss? 08:51:20 if noone else does, I have a question 08:51:31 Murad: hello 08:51:48 Hello 08:51:49 bob : have you ever tried to restart your jenkins-master machine and tried . so that it would restart the jenkins server also and all other services 08:52:00 Murad: hii 08:52:06 Hi 08:52:07 last week we discussed how to get your server to serve files in the browser without being prompted to download them 08:52:26 yes 08:52:27 Murad: do you know how to access the logs of the third-party meetings? 08:52:49 since I believe I saw you asking this exact same question in the -infra channel this week 08:52:51 yes I have a link for the previous meeting 08:52:56 and I was curious as to why that was 08:53:03 since we discussed it last week 08:53:18 and I wondered if perhaps you don't know where to find the meeting logs to read them 08:53:21 That was because I was using wrong way to compress files 08:53:30 I have to use "gzip" 08:53:30 oh, can you explain? 08:53:45 why do you have to use "gzip"? 08:53:51 I was using "tar" to produce gz files 08:53:56 what is making you do that? 08:54:18 in order to compress log files 08:54:37 well you use the tar command with the flag to output gziped files 08:54:50 to produce gzipped files 08:55:01 yes but that didn't work until I used "gzip" 08:55:25 okay so what was the question to -infra (which I am redirecting to this meeting) 08:56:20 when I was compressing files with TAR, it only worked on firefox. So I though that I might using wrong apache configurations 08:56:29 and that was my question 08:56:59 after investigation, I used gzip and it worked 08:57:24 okay so you discoverd the answer to your own question 08:57:31 yes 08:57:36 great 08:57:39 thank you 08:57:46 thanks 08:57:49 does anyone have anything else? 08:58:28 well time to wrap up then 08:58:29 anteaya: Our CI log server is online now.If you have free time,could you help to point out the ssl-cert issue of our logs:) 08:59:00 jyuso: you can pm me the ip address so I can view the logs 08:59:08 anteaya: OK 08:59:12 thank you 08:59:25 so thank you everyone for attending this week 08:59:30 see you next week 08:59:34 #endmeeting