Tuesday, 2020-10-27

*** mugsie has quit IRC01:00
*** mugsie has joined #airshipit01:04
airship-irc-bot<kk6740> @mattmceuen @ih616h , i am not sure about this, but is it possible to revert this commit https://review.opendev.org/#/c/750604 ? There are couple of issues with it: 1. Unit tests don’t work anymore with it 2. it uses exec system calls to invoke `gpg` command, which basically means that airshipctl now has hard external dependency on `gpg` binary, which we were trying to avoid i think, because instead of binary, airshipctl needs to02:20
airship-irc-botmanage it’s external dependencies. If we want this to work, I would recommend that this should be invoked as a plugin in docker using new way of invoking plugins, and MUST have it’s code base separated from `airshipctl` , because now even if u don’t need this plugin, you are still required to have `gpg`  binary installed for unit tests to work 3. i think we introduce vulnerability  when storing in /tmp secret encryption keys02:20
airship-irc-bot<kk6740> sorry, if that is being too radical, but i am really concerned about this commit. I know i should have reviewed it as well, to express my opinion beforehand02:24
airship-irc-bot<raliev> I agree with @kk6740, this patchset requires further work: `sops.go` and `sops_test.go`  were not even reviewed by anybody since patch consists of 1500 lines of code and it's hard to cover every changed/added file. In addition to Kostiantyn comments: 1. this patch uses ioutil-* and os-* functions to perform file manipulations which is not quite good since we trying to establish single standard to work with files using document FS and02:54
airship-irc-botavoid multiple approaches to have at least unit tests consistent  2. patch consists duplicated testdata02:54
*** rezroo has joined #airshipit04:33
*** evrardjp has quit IRC05:33
*** evrardjp has joined #airshipit05:33
*** rezroo has quit IRC06:10
*** rezroo has joined #airshipit06:12
*** rezroo has quit IRC09:14
*** muhaha has joined #airshipit09:31
*** uzumaki has joined #airshipit10:00
*** muhaha has quit IRC10:26
*** TheR1nzler has joined #airshipit11:23
*** TheR1nzler has quit IRC12:00
airship-irc-bot<mattmceuen> @kk6740 if unit tests don't work any more, how did it pass the lint-unit gate?13:03
airship-irc-bot<mattmceuen> you just mean, if gpg isn't installed, right?13:03
airship-irc-bot<kk6740> no13:03
airship-irc-bot<mattmceuen> want to discuss it in the design call going on now? :)13:04
airship-irc-bot<kk6740> it’s not stable i think13:04
*** SRao has joined #airshipit13:05
airship-irc-bot<pb269f> We certainly need to fix the gpg dependency13:09
airship-irc-bot<se6518> Please, review the last part of proxy config in airshipctl: https://review.opendev.org/75660115:14
*** uzumaki has quit IRC16:01
airship-irc-bot<dwalt> Hey team, what is responsible for building the replacement transformer and templater Docker containers in a local run?17:09
airship-irc-bot<sirajudeen.yasin> `make images` should build the plugin images17:10
airship-irc-bot<sirajudeen.yasin> @dwalt, does it answer your query ?17:14
airship-irc-bot<dwalt> That helps. Thanks @sirajudeen.yasin17:17
airship-irc-bot<uday> Hey @kk6740 and @raliev, just fyi, I am looking into the test failures caused by https://review.opendev.org/#/c/750604 and hopeful I will be able to debug and fix this in a bit, will keep you posted. I propose we comment out the offending test case while I work on addressing your feedback passed here and fixing the unit tests if I cannot do it in an other hour or two. Let me know what you think. @Hey17:18
*** uzumaki has joined #airshipit17:47
*** SRao has quit IRC17:55
airship-irc-bot<kk6740> @uday i dont know what would be the best way to go, removing test seems scary tbh18:04
airship-irc-bot<sb464f> I Team, is there any order of deploying cluster objects when all applied together, ex:- when a kubeadmcontrolplane is applied will `kcp` object be created first no matter the order that is present in kustomization..?18:07
airship-irc-bot<uday> ok, so let me take a stab at this to fix it for the next hour. If I cannot get to fixing it, we can look into other options, one of which can also be to revert if everyone agrees to it.18:10
*** awander has joined #airshipit18:12
airship-irc-bot<kk6740> they should work togather fine in one bundle19:27
airship-irc-bot<kk6740> we have discussed reverting, and it was decided that it is not an option, so i would rather find a solution. I think you are on the right track with versions19:29
airship-irc-bot<alexander.hughes> hey team, quick reminder: Airship PTG sessions begin tomorrow, and run through Thursday.  0800-1200 CST / 1300-1700 UTC both days in https://www.openstack.org/ptg/rooms/cactus  agenda here: https://etherpad.opendev.org/p/wallaby-ptg-airship19:43
airship-irc-bot<sb464f> @kk6740 looks like it is not, as metal3machine template is being applied after kcp. the scaling of kcp is happening immediately by the time new metal3machineteplate is applied which is causing the machine that is coming up using old metal3machinetemplate.20:27
airship-irc-bot<sb464f> it’s random but hitting that issue which cause deployment to fail sometimes.20:27
airship-irc-bot<kk6740> which version of cluster-api cp provider ?20:28
airship-irc-bot<sb464f> 0.3.720:28
airship-irc-bot<kk6740> can we have one more debug sessionfor this ?20:28
airship-irc-bot<kk6740> in 30 minutes thou20:28
airship-irc-bot<sb464f> Sure.20:28
airship-irc-bot<sb464f> let me know when you are available.20:29
airship-irc-bot<kk6740> @sb464f i’m ready21:04
*** uzumaki has quit IRC22:23

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!