IRC logs for #cip for Wednesday, 2019-01-09

szlinhelmut: I've pinged Kazu.01:58
*** Vincent_HH_Wu has joined #cip02:26
*** Vincent_HH_Wu has left #cip04:17
*** VincentHHWu has joined #cip04:18
*** rajm has joined #cip07:59
szlinpatersonc: ping08:07
patersoncszlin: pong08:08
szlinpatersonc: I would like to introduce you to my colleague - VincentHHWu, who is responsible for kernel testing.08:12
patersoncHello VincentHHWu!08:13
VincentHHWuhi patersonc! Nice to meet you!08:14
szlinVincentHHWu: feel free to discuss kernel testing stuff with patersonc :D08:18
patersoncYes please :)08:18
patersoncDo you need access to the CIP LAVA lab?08:19
VincentHHWuWe are going to build up a corporate-level kernel testing infra. And I just switched to this job recently.08:20
patersoncGreat08:21
VincentHHWuso any info. related to kernel testing and CI/CD will be helpful for me.08:21
patersoncLet me know if you need any help. We've just gone through this process.08:21
patersoncThis may be helpful: https://gitlab.com/cip-project/cip-testing/lava-docker/blob/cip-lava-docker/documentation/setup_cip_lava_lab.md08:22
VincentHHWuThanks. Some cross-department discussions are expected, and maybe we will need some experiences on collaboration or system migration.08:31
VincentHHWuWe will let you know our progress.  Let's see what we can do together. :)08:36
patersonc:)08:36
*** toscalix has joined #cip08:52
*** rajm has quit IRC09:34
*** rajm has joined #cip11:24
wagipatersonc: slowly getting the work lab working again. right now playing with iwg20m14:42
wagiand obviously the kernel I produce does not boot15:04
patersonc:D15:28
szlinpatersonc:  <- expert15:29
patersoncLies15:29
wagishmobile_defconfig should be good for iwg20m, right?15:36
wagistupid me15:38
wagiI am using the v4.4 tree15:38
wagiinstead of the v4.4-cip tree15:39
patersoncNot v4.4-cip? ;)15:39
patersonchehe15:39
patersoncshmobile_defconfig is correct15:39
wagiso I have no the iwg20m board back for playing around15:39
wagiit took so long because of our very competent IT deparment15:40
wagiI was about to beserk15:40
patersonc:)15:40
patersoncI know that feeling!15:40
wagiinsane the amout of time spend to get a simple x86 box setup15:41
patersoncYou should have used an R-Pi or something similar ;)15:41
wagithe final step (which is obvious) involved installing a complete windows 10 image, because one of the install steps set the correct bios password I needed to disable secure boot15:41
wagiSure, but then I had no fun with lava15:42
patersoncAmazing15:42
wagiand if the first level support were any competent they could have told me that the password will be set when the official windows client is installed15:44
wagibut no, far too easy15:44
wagianyway, I have two x86 boxes now because of that15:44
wagiI might just use the second one for x86-64 testing15:45
wagiWhat drives me insane, is that my time as engineer is spend doing BS tasks. It seems engineering time is dirt cheep15:46
wagiSo enough with steaming. :)15:46
wagipatersonc: hurray, kernel boots15:52
waginow let's see if I get it also booting with lava15:52
*** rynofinn has joined #cip15:56
*** mungaip has joined #cip16:00
patersoncwagi: Great!16:00
patersoncHello rynofinn!16:01
rynofinnhello!16:01
patersoncIt looks like you're having more fun with the LAVA master :)16:01
rynofinnhehe indeed16:02
rynofinnnow it looks like the restore can't find the list of users to skip16:03
patersoncAnnoying16:03
patersoncDid you cherry-pick the same patch as yesterday?16:03
rynofinnyes, I stashed that and reapplied after the rebase16:04
patersoncOkay16:04
rynofinnbut the error is slightly different now.  Before it was fooled by the username format and didn't skip some users16:04
rynofinnnow it is not skipping anyone16:04
wagipatersonc: it kernel boots with lava, just some dns/nfs issues to resolv but that shouldn't be a real problem. (famous last words)16:05
rynofinnI've tried adding some debug statements to the setup.sh and they don't appear, so maybe setup.sh isn't getting called?16:05
patersoncrynofinn: mungaip is looking into it16:06
patersoncPerhaps16:06
mungaiprynofinn: Hello!16:06
rynofinnhello!16:06
patersoncwagi: Great. You may need to specify the nfs version.16:07
wagipatersonc: yep, that was missing16:07
mungaipI'm going to try to understand the difference between the patch you sent yesterday and the original for the users section16:08
wagipatersonc: user space is running16:08
patersoncwagi: What repository are you using for the lab?16:09
wagifor lava? I am using debian backports16:11
patersoncGood to know16:11
wagiand the smoke test passes16:12
wagiso I have a working setup now16:12
wagime goes home16:12
patersoncnn16:12
mungaiprynofinn: I'm not sure why the if statement is expecting an exit code of 0- when I run the previous command manually, I get an exit code of 116:29
rynofinnI also think the regex is bad as there won't always be a trailing space after usernames16:30
rynofinnso the last [[:space:]] probably should not be in the regex16:30
mungaipWorth experimenting at some point16:31
mungaipryanofinn: I was wrong but as you said the last [[:space:]] should not be in the regex16:45
mungaipI've just manually tried it on our internal master container with a user, and the user isn't found if the [[:space:]] is added at the end of the username16:48
mungaipI'm off now, so let me know if that fix works :)16:49
rynofinnI've removed the [[:space:]] from the end and it still tries to add the admin user16:53
rynofinnoh, I made the change to setup.sh but it didn't get copied into the container for use.  I need to do a full redeploy to test16:56
rynofinnugh, when I try to do a full redeploy, I get "unknown keyword groups" from lavalab-gen.py now17:07
rynofinnthat is totally new17:07
rynofinnI'm now convinced that when I make changes to lava-master/scripts/setup.sh they have no effect because setup.sh is now getting pulled from some other place into the container17:15
*** mungaip has quit IRC17:30
*** toscalix has quit IRC18:37
*** rajm has quit IRC19:40
patersoncHello rynofinn20:06
rynofinnhello20:06
patersoncHow's it going?20:07
rynofinnI managed to get the service up without data using the old code20:07
rynofinnnow I'm trying to get the restore to work again20:07
patersoncWith the kernelci/lava-docker?20:07
rynofinnright20:08
rynofinnso I'm still trying to rollback20:08
patersoncHow many changes do you have on top? Or is it pretty much vanilla, just with a different boards.yaml?20:08
rynofinnnot sure what changes you mean20:09
patersoncAs in, are you just cloning kernelci/lava-docker and updating boards.yaml? Or are there other patches you are also applying?20:10
rynofinnI made a tarball of /srv/lava-docker right before the rebase20:10
rynofinnand I'm using that20:10
rynofinnso in theory it should start right back up20:10
rynofinnbecause there are no changes20:10
patersoncOkay20:10
rynofinnbut rolling back has not worked20:11
rynofinnok, I got rollback to work20:22
patersonc:)20:22
rynofinnhad to restore to the point of Jan 420:22
patersoncThat's okay20:23
rynofinnI can give the rebase another shot tomorrow20:23
patersoncThanks for getting it up again20:24
rynofinn:)20:25
patersoncIt would be really good to be able to switch to using https://gitlab.com/cip-project/cip-testing/lava-docker/tree/cip-lava-docker20:25
patersoncWe've pre-added device healthcheck etc, so we'll avoid things like "Configuration Error: missing or invalid template.20:26
patersonc"20:26
patersonce.g. https://lava.ciplatform.org/scheduler/device/r8a7743-iwg20d-q7-0120:26
*** rajm has joined #cip22:44
*** rajm has quit IRC22:56

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!