13:31:15 <chandankumar> #startmeeting RDO meeting - 2017-09-19
13:31:16 <openstack> Meeting started Tue Sep 19 13:31:15 2017 UTC and is due to finish in 60 minutes.  The chair is chandankumar. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:31:17 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:31:19 <openstack> The meeting name has been set to 'rdo_meeting___2017_09_19'
13:32:13 <chandankumar> apevec: jschlueter number80 jpena amoralej snecklifter jatanmalde aditya_r jruzicka rbowen office hour time
13:32:18 <jpena> o/
13:32:19 <chandankumar> #topic Roll Call
13:32:24 <chandankumar> #chair jpena
13:32:24 <jatanmalde> \o/
13:32:25 <openstack> Current chairs: chandankumar jpena
13:32:26 <amoralej> o/
13:32:28 <snecklifter> o/
13:32:34 <chandankumar> #chair jatanmalde amoralej snecklifter
13:32:34 <openstack> Current chairs: amoralej chandankumar jatanmalde jpena snecklifter
13:32:35 <number80> o/
13:32:40 <chandankumar> #chair number80
13:32:41 <openstack> Current chairs: amoralej chandankumar jatanmalde jpena number80 snecklifter
13:32:42 <aditya_r> \o
13:32:48 <chandankumar> #chair aditya_r
13:32:49 <openstack> Current chairs: aditya_r amoralej chandankumar jatanmalde jpena number80 snecklifter
13:32:50 <jatanmalde> hi guys.
13:33:08 <chandankumar> #Link https://review.rdoproject.org/etherpad/p/rdo-office-hour-easyfixes
13:33:21 <chandankumar> Today's Agenda:
13:33:45 <chandankumar> * Let's cleanup Open easyfix review https://review.rdoproject.org/r/#/q/status:open+branch:rpm-master+easyfix,175
13:34:04 <chandankumar> thanks to aditya_r for fixing the stuff, really appreciated
13:34:23 <aditya_r> chandankumar, welcome :)
13:34:32 <jatanmalde> Great work aditya_r
13:34:45 <chandankumar> * Cleanup python3 packages: https://trunk-primary.rdoproject.org/f25/report.html
13:34:58 <aditya_r> jatanmalde, thanks
13:35:04 <chandankumar> #info Easyfix open review queue >= 175
13:35:06 <snecklifter> chandankumar: number80 raised a point with the macro fixes that fedora doesn't have openstack-macros
13:35:50 <number80> well, I may work on a more durable solution as I will be at a pycon sprint
13:36:11 <number80> but I think we should not merge until then changes on clients
13:36:28 <rbowen> Thanks, chandankumar
13:36:44 <chandankumar> number80: i am adding -W on clients review keep it on hold
13:37:04 <chandankumar> number80: what about oslo libraries?
13:37:14 <number80> the same
13:38:01 <chandankumar> #info Keep -W for oslo + clients till openstack-macros is available for fedora
13:38:40 <chandankumar> jpena: amoralej snecklifter https://review.rdoproject.org/r/#/q/topic:%22easyfix%25234%22+status:open
13:38:45 <chandankumar> might be good for review
13:39:38 <jpena> aditya_r: there are many reviews where the description is "Moved the description to global variable redhat-openstack/easyfix#4". Could you fix them to add a blank line between the first and second lines?
13:39:57 <snecklifter> aditya_r: and a few merge conflicts to fix :)
13:40:24 <jatanmalde> guys, I have added myself to a review
13:40:29 <jatanmalde> I want to review a code
13:40:29 <aditya_r> jpena, snecklifter: Sure, will check.
13:40:41 <jatanmalde> how can I procede
13:40:43 <jschlueter> chandankumar: o/
13:40:52 <chandankumar> #chair jschlueter
13:40:53 <openstack> Current chairs: aditya_r amoralej chandankumar jatanmalde jpena jschlueter number80 snecklifter
13:41:09 <chandankumar> Ah Nice, So many people today :-)
13:42:25 <jatanmalde> How can review a patch ?
13:42:28 <snecklifter> https://review.rdoproject.org/r/#/q/topic:easyfix/23+status:open
13:42:37 <snecklifter> ^^^ would appreciate reviews on all these
13:43:00 <jatanmalde> snecklifter, sure I will
13:43:16 <jatanmalde> But need some help in how should I procede
13:43:52 <snecklifter> number80: %{?systemd_requires} <- this is already in fedora, no?
13:44:14 <number80> Yes, systemd macros are shipped from systemd
13:44:19 <jpena> snecklifter: yes, it requires "BuildRequires: systemd"
13:44:22 <number80> s/from/by/
13:44:59 <snecklifter> cool, so nothing holding them back from getting merged :)
13:45:07 <number80> nope
13:45:45 <rdogerrit> Javier Peña proposed openstack/vitrageclient-distgit rpm-master: Moved the description to global variable  https://review.rdoproject.org/r/9351
13:46:36 <rdogerrit> Javier Peña proposed openstack/oslo-concurrency-distgit rpm-master: Moved the description to global variable  https://review.rdoproject.org/r/9348
13:47:58 <jschlueter> chandankumar: https://review.rdoproject.org/r/#/c/9330/1 commit message formatting
13:48:38 <chandankumar> jschlueter: aditya_r is working on that i think
13:48:51 <jschlueter> chandankumar: ack
13:48:56 <chandankumar> jschlueter: -1 will work on it i think
13:49:25 <jschlueter> aditya_r: ack I added 2 -1's so far on commit message formatting
13:49:49 <jschlueter> aditya_r: have you seen standard how to write a good commit message for git?
13:50:15 <jschlueter> s/standard/any of the many available pages for/
13:50:52 <jschlueter> chandankumar: that's linked in the README for easyfixs or in the getting started docs right ?
13:51:16 <sfbender> Frederic Lepied proposed DLRN master: Support to build in Fedora Copr  https://softwarefactory-project.io/r/9717
13:51:50 <chandankumar> jschlueter: i think getting started docs would be goofd
13:51:54 <chandankumar> *good
13:52:16 <chandankumar> jschlueter: https://www.rdoproject.org/documentation/onboarding/
13:52:22 <chandankumar> it would be good to add here
13:52:50 <dmsimard> jpena: hi, re: xstatic
13:53:02 <rdogerrit> rdo-trunk created openstack/vitrage-distgit rpm-master: openstack-vitrage: failed to build b51339e  https://review.rdoproject.org/r/9678
13:53:09 <jschlueter> chandankumar: ack
13:53:12 <dmsimard> jpena: we probably have them in RDO right ? is the issue that they're not up to date enough ?
13:53:17 <chandankumar> pradk: https://review.rdoproject.org/r/#/c/9600/2
13:53:51 <jpena> dmsimard: we only have them on CBS, which is CentOS-only. For the Fedora Rawhide RDO Trunk builder, we need dependencies in Fedora
13:54:22 <number80> xstatic only in CBS?
13:54:27 <number80> which ones?
13:55:10 <dmsimard> number80: I submitted two xstatic packages a long time ago, I guess we never got around to sending them to fedora
13:55:21 <dmsimard> number80: https://bugzilla.redhat.com/show_bug.cgi?id=1364607 was one
13:55:22 <openstack> bugzilla.redhat.com bug 1364607 in Package Review "Review Request: python-XStatic-objectpath - ObjectPath JavaScript library packaged for setuptools (easy_install) / pip." [Unspecified,Assigned] - Assigned to karlthered
13:55:33 <aditya_r> jschlueter, the description is "Moved the description to global variable redhat-openstack/easyfix#4" that twice. That time I was in learning phase. I have created new one for that, please go through.
13:55:51 <dmsimard> number80: https://bugzilla.redhat.com/show_bug.cgi?id=1364603 was the other
13:55:51 <openstack> bugzilla.redhat.com bug 1364603 in Package Review "Review Request: python-XStatic-Angular-Schema-Form - Angular-Schema-Form JavaScript library packaged for setuptools (easy_install) / pip." [Unspecified,Assigned] - Assigned to karlthered
13:56:17 <pradk> chandankumar, do we need openstack-macros as a build requires if its doing pre clean?
13:56:21 <dmsimard> number80: ah, https://bugzilla.redhat.com/show_bug.cgi?id=1364620 too
13:56:22 <openstack> bugzilla.redhat.com bug 1364620 in Package Review "Review Request: python-XStatic-tv4 - tv4 - Tiny Validator (for v4 JSON Schema) JavaScript library packaged for setuptools (easy_install) / pip." [Unspecified,Assigned] - Assigned to nobody
13:56:55 <pradk> chandankumar, i'll test and merge today
13:57:23 <chandankumar> pradk: yes for the added macro
13:57:25 <number80> Damn
13:57:35 * number80 will do these reviews
13:57:39 <chandankumar> pradk: Thanks :-)
13:58:01 <aditya_r> jschlueter, Please skip that  description is "Moved the description to global variable redhat-openstack/easyfix#4" if possible.
13:58:09 <number80> so it means that we dropped the prereview suffix in NVR ...
13:58:19 <rdobot> [sensu] NEW: master.monitoring.rdoproject.org - check-delorean-master-current @ http://tinyurl.com/y8hxewm2 |#| Build failure on centos7-master/current: vitrage: http://trunk.rdoproject.org/centos7-master/report.html
13:58:35 <jschlueter> aditya_r: I'll let it up to you to get fixed up and then look again
13:59:48 <aditya_r> jschlueter, ok.
14:01:47 <rdogerrit> Jon Schlueter proposed openstack/zaqar-distgit rpm-master: Dropping python-sqlite3dbm Require  https://review.rdoproject.org/r/9676
14:02:59 <jruzicka> apevec, number80, jpena ad https://github.com/softwarefactory-project/rdopkg/issues/77
14:03:07 <dmsimard> jpena, number80, apevec, jschlueter, amoralej: are we good for an adhoc meeting to debrief PTG sometime today ?
14:03:36 <jruzicka> there are two possible Release bump strategies:
14:04:03 <jpena> dmsimard: my afternoon has several meetings... I have a 30min slot in 1h, or then in 2:30h
14:04:11 <amoralej> dmsimard, it's ok for me
14:04:17 <rdogerrit> Pradeep Kilambi created openstack/gnocchi-distgit newton-rdo: Rebase to 3.0.15  https://review.rdoproject.org/r/9679
14:04:22 <number80> dmsimard: I'm in the middle of calls
14:04:27 <rdogerrit> Merged openstack/oslo-reports-distgit rpm-master: Moved the description to global variable  https://review.rdoproject.org/r/9350
14:04:28 <dmsimard> jpena: yeah we're in the same meeting for sf dfg
14:04:32 <jruzicka> 1) bump last numeric part of Release (0.1%{?milestone} -> 0.2%{?milestone})
14:04:53 * number80 will be going to local pycon tomorrow so no calls allowed in the train
14:04:53 <rdogerrit> Merged openstack/os-collect-config-distgit rpm-master: convert individual requires to macro  https://review.rdoproject.org/r/9514
14:05:19 <jruzicka> 2) add extra numeric part (0.1234456.deadbeef -> 0.1234456.deadbeef.1)
14:06:04 <jruzicka> 2) is used by DLRN and to support both I need to do guessing
14:06:20 <number80> => 1
14:06:38 <jruzicka> we use 1) in RDO but DLRN spec require 2)
14:07:16 <jruzicka> supporting both in rdopkg requires dark magick like "if the numbers in release are really long, it's probably dlrn .spec"
14:08:01 <jruzicka> jpena, can't we make tese compatible? ^
14:08:39 <number80> jruzicka: or just have NVR bumping strategies
14:08:52 <number80> one by default and people can configure another one
14:09:09 <rdogerrit> Jatan Malde proposed openstack/oslo-reports-distgit rpm-master: Added the doc_guard to it.  https://review.rdoproject.org/r/9444
14:09:37 <jruzicka> number80, there still needs to be a deterministic way to tell them apart to avoid extra configuration file.
14:10:00 <jpena> jruzicka: in a meeting, will read back later
14:10:09 <jschlueter> chandankumar: https://review.rdoproject.org/r/#/c/9676/ here is a related one but a bit more difficult to detect easy fix
14:10:10 * number80 thinking about it
14:10:12 <jruzicka> jpena, ack, thanks
14:10:18 * number80 is in a call
14:10:30 <jschlueter> number80, flaper87: ^^ can you guys peek at https://review.rdoproject.org/r/#/c/9676/
14:10:56 <jruzicka> allright, everyone's busy ;) Think about it and I'll rise on RDO meeting, probably.
14:11:35 <number80> jruzicka: I'm growing old, I can't focus on more than one or two things at a time :)
14:12:09 <number80> I almost froze in the call, trying to get back in context :)
14:13:48 <number80> vkmc: you may be interested too if you have time ^
14:14:12 * number80 is not giving -W but please wait
14:14:22 <aditya_r> jschlueter, How to close any review by yourself?
14:14:33 <aditya_r> See you are trying to fix it https://review.rdoproject.org/r/#/c/9334/.
14:14:48 <aditya_r> But it is already get fixed. https://github.com/rdo-packages/mox3-distgit
14:15:54 <chandankumar> aditya_r: you can use abandon button to close the review , also please add a message to the abandon box
14:16:24 <aditya_r> chandankumar, ok.
14:16:27 <jschlueter> aditya_r: if you want to use the "abandon" button if you don't want to keep the patch around anymore, but it's good pattern to include a message as to why it's being abandoned ... what chandankumar said
14:16:51 <aditya_r> jschlueter, thanks.
14:19:13 <jschlueter> jruzicka: I ran into odd edge case where rdopkg fix duplicated the %{dist} flag when it bumped ... I need to get test case added to repeat issue
14:22:24 <sfbender> Frederic Lepied proposed DLRN master: Save src.rpm when build fails  https://softwarefactory-project.io/r/9046
14:23:28 <jruzicka> jschlueter, yeah I got patch for that already, related to https://github.com/softwarefactory-project/rdopkg/issues/77
14:23:45 <jruzicka> jschlueter, it's happending with %{dist} (with missing '?'), right?
14:24:01 <rdogerrit> Merged openstack/magnum-ui-distgit rpm-master: Replace rm -rf requirements.txt step with proper rpm macro  https://review.rdoproject.org/r/9630
14:25:52 <rdogerrit> Merged openstack/novajoin-distgit rpm-master: convert individual requires to macro  https://review.rdoproject.org/r/9524
14:29:30 <jschlueter> jruzicka: hmm actually 1.2.3-20170801.deadbeef%{?dist}.1 bumped to .2.3-20170801%{?dist}.1.%{?dist} IIRCE
14:31:09 <jschlueter> jruzicka: might be able to use datetime-munging to check if long set of numbers is dlrn based or not ...
14:31:16 <rdogerrit> Aditya Ramteke proposed openstack/neutron-vpnaas-distgit rpm-master: Replaced rm -rf requirements.txt step with proper rpm macro.  https://review.rdoproject.org/r/9626
14:31:19 <rdogerrit> Merged openstack/gnocchi-distgit rpm-master: Replaced rm -rf requirements.txt step with proper rpm macro.  https://review.rdoproject.org/r/9600
14:32:00 <jschlueter> jruzicka: but the issue is going from %{?dist}.1 to %{dist}.1.%{?dist} or something silly like that ... where the dist part is now duplicated
14:33:29 <rdogerrit> Merged openstack/manila-ui-distgit rpm-master: Replace rm -rf requirements.txt step with proper rpm macro.  https://review.rdoproject.org/r/9579
14:33:32 <rdogerrit> Merged openstack/neutron-fwaas-distgit rpm-master: Replace rm -rf requirements.txt step with proper rpm macro  https://review.rdoproject.org/r/9607
14:33:41 <jruzicka> jschlueter, ah when it's not at the end... right :)
14:33:55 <jschlueter> jruzicka: yep
14:35:26 <jruzicka> jschlueter, {%dist}.1 is silly thought, is it not? :)
14:36:52 <jschlueter> jruzicka: it's out in the wild ... but ok with it just not duplicating dist entry
14:37:22 <jschlueter> jruzicka: I've seen that flavor from tvignaud and IIRC some fedora/RHEL packages as well
14:38:12 <tvignaud> jschlueter: I'm find with only adding .1 at end :-)
14:38:25 <rdogerrit> Merged openstack/ironic-inspector-distgit rpm-master: Replaced rm -rf requirements.txt step with proper rpm macro.  https://review.rdoproject.org/r/9555
14:39:33 <jpena> jruzicka, jschlueter: about the rdopkg issue, I wonder why we want to modify a spec generated by DLRN
14:40:03 <jschlueter> jpena: for downstream workflow ...
14:40:25 <jschlueter> we pull dlrn generated spec and tarball as baseline and sometimes we have to fix it for one reason or another
14:40:50 <jschlueter> downstream patches, differences in packaging which we haven't gotten into RDO yet
14:40:57 <jpena> jschlueter: ok, I understand.
14:41:09 <jschlueter> just need a repeatable way to bump a NVR one way or another
14:41:15 <jschlueter> so it doesn't look funny
14:41:46 <jruzicka> jschlueter, I could simply unbreak "bump last numeric" but bumping the timestamp is a little bit evil, isn't it? :)
14:41:52 <amoralej> apevec, the ping problem may be related to the image itself
14:42:06 <amoralej> same version of dib in the same system works fine with a previous image
14:42:08 <rdogerrit> Merged openstack/ironic-ui-distgit rpm-master: Replace rm -rf requirements.txt step with proper rpm macro  https://review.rdoproject.org/r/9624
14:42:10 <jpena> jruzicka: DLRN always uses the same algorithm to generate the release: https://github.com/softwarefactory-project/DLRN/blob/master/scripts/common-functions#L44
14:42:11 <rdogerrit> Merged openstack/ironic-python-agent-distgit rpm-master: Replaced rm -rf requirements.txt step with proper rpm macro.  https://review.rdoproject.org/r/9563
14:42:13 <rdogerrit> Merged openstack/panko-distgit rpm-master: Replace rm -rf requirements.txt step with proper rpm macro  https://review.rdoproject.org/r/9627
14:42:32 <jpena> jruzicka: maybe we could just detect it with a regexp, and in that case apply the special case?
14:42:49 <panda> jpena: I lost the review that should solve the java exception on jenkins
14:43:02 <jschlueter> jruzicka: I'm ok with it ... as long as it drops the git sha1 from the NVR
14:43:19 <jschlueter> just want an easy to identify pattern to see that it's different
14:43:37 <jpena> panda: it was https://review.openstack.org/504416
14:43:47 <jruzicka> jpena, yeah sure, if it can be detected in a clear way.
14:43:49 <jschlueter> jruzicka, jpena: other way is to use the global dlrn flag to make decisions
14:43:53 <jruzicka> jpena, alternatively:
14:43:58 <jruzicka> RELEASE=0.$(date --utc "+%Y%m%d%H%M%S").$2.1
14:43:58 <jpena> panda: which is merged now, buuuuuut we have another issue
14:44:12 <jruzicka> adding extra .1 to bump would also work
14:44:21 <jruzicka> wait, no
14:44:24 <jschlueter> just don't want stacking .1.1.1.1
14:45:05 <panda> jpena: yay
14:45:09 <jruzicka> jschlueter, ofc, once there's .1 in place it can be bumped... but that's not the correct format
14:46:10 <jpena> panda: TL;DR: since upstream Gerrit was updated, some stuff is not (yet) working, and project creation is one of them. We have a new project in project-config, but there's no repo in git.openstack.org, which breaks image builds
14:46:20 <jschlueter> jruzicka: rdopkg can be opinionated for format supported and recommended but we might want to do a quick check for what is out in wild where it is used for patterns and make sure we create scenarios for likely best behavior on it
14:46:29 <rdogerrit> Christopher Brown proposed openstack/ironic-python-agent-distgit rpm-master: convert individual requires to macro  https://review.rdoproject.org/r/9520
14:46:30 <jpena> pabelanger: do you have an estimated timeframe to get new projects created in review.openstack.org?
14:46:54 <rdogerrit> Christopher Brown proposed openstack/ironic-inspector-distgit rpm-master: convert individual requires to macro  https://review.rdoproject.org/r/9527
14:46:57 <jpena> (for context, networking-lagopus is in project-config but there's no repo)
14:47:49 <dmsimard> panda, trown, weshay: btw (I know it brings us back) but tristanC ended up finding the root cause for the issue http://eavesdrop.openstack.org/irclogs/%23rdo/%23rdo.2017-08-04.log.html#t2017-08-04T16:10:08
14:47:52 <jpena> dmsimard: btw, meeting finished early, I have time fotr that new meeting, now
14:48:07 <dmsimard> jpena: we have dfg thing in 12 minutes, can wait until tomorrow
14:48:13 <jruzicka> jschlueter, correct.
14:48:14 <jpena> ok then
14:48:33 <jpena> dmsimard: wait, wasn't it at half-past?
14:48:45 <rdogerrit> Merged openstack/vmware-nsx-distgit rpm-master: Replace rm -rf requirements.txt step with proper rpm macro  https://review.rdoproject.org/r/9635
14:48:59 <rdogerrit> Merged openstack/networking-vsphere-distgit rpm-master: Replace rm -rf requirements.txt step with proper rpm macro.  https://review.rdoproject.org/r/9671
14:49:00 <dmsimard> panda, trown, weshay: it was indeed an ara issue and it is caused by an old enough version of Flask (2013) so it seems like an old version of flask was installed for some reason
14:49:16 <dmsimard> jpena: dang you're right
14:49:30 <dmsimard> if number80, apevec, amoralej ans jschlueter are good, we can fire something up
14:50:37 <rdogerrit> Merged openstack/heat-distgit rpm-master: convert individual requires to macro  https://review.rdoproject.org/r/9513
14:51:35 <adarazs> dmsimard: hi, did somebody bring it up to you that we have this jenkins error in our containers building jobs?
14:51:37 <adarazs> 10:01:35 FATAL: remote file operation failed: /home/jenkins/secretFiles/41ae2d48-dfc9-4964-a6d6-7b9bf000dad2 at hudson.remoting.Channel@25dc0229:upstream-centos-7-rdo-cloud-tripleo-5918: java.io.IOException: java.lang.reflect.InvocationTargetException
14:51:42 <adarazs> 10:01:35 java.io.IOException: remote file operation failed: /home/jenkins/secretFiles/41ae2d48-dfc9-4964-a6d6-7b9bf000dad2 at hudson.remoting.Channel@25dc0229:upstream-centos-7-rdo-cloud-tripleo-5918: java.io.IOException: java.lang.reflect.InvocationTargetException
14:51:45 <dmsimard> adarazs: yes
14:51:58 <dmsimard> adarazs: it's unrelated to the actual container builds
14:51:59 <adarazs> dmsimard: is there some bug on it I can follow?
14:52:17 <dmsimard> adarazs: it's a mix of centos 7.4 issue and from the upstream image
14:52:17 <adarazs> dmsimard: I assumed so, probably some jenkins configuration issue, but we didn't change anything in the JJBs I think
14:52:46 <adarazs> dmsimard: is there any bug or something about it?
14:53:03 <jpena> adarazs: does it say anything about java.nio?
14:53:13 <dmsimard> adarazs: I don't believe there is, jpena and I have been hunting the issue down
14:53:14 <jpena> if so, it's https://bugs.launchpad.net/tripleo/+bug/1717959
14:53:15 <openstack> Launchpad bug 1717959 in tripleo "RDO upstream-centos-7 nodepool image is out of date" [Critical,Triaged] - Assigned to Alan Pevec (apevec)
14:53:21 <dmsimard> oh well there you go
14:53:38 <adarazs> thanks :)
14:53:42 <dmsimard> jpena: looks like the devstack change is merged, I'll go and look at the image builds
14:54:08 <jpena> dmsimard: yes, and now we have another issue with a project that is present in project-config, but has no repo in git.openstack.org
14:54:24 <dmsimard> ugh
14:54:38 <jpena> yep, perfect storm
14:54:48 <dmsimard> jpena: syncing project-config at this point is dangerous, there's been a lot of changes due to zuul v3
14:55:11 <dmsimard> and even if we get the image to build, it doesn't mean it will end up working
14:55:28 <jpena> dmsimard: is there a way for us to pin to an earlier commit?
14:55:50 <dmsimard> jpena: later or earlier ? because earlier was breaking us (hence why I stopped syncing)
14:56:22 <jpena> at least earlier than the current master, where this networking-lagopus project was defined
14:57:02 <dmsimard> jpena: do you have the commit ? trying to find something right now
14:57:26 <jpena> dmsimard: https://github.com/openstack-infra/project-config/commit/0335417d0d1b6b69cc95da0134b15f7e67878f3b
14:57:42 <dmsimard> yikes, that's like yesterday
14:57:55 <dmsimard> not confident at all
14:57:57 <dmsimard> let me see
14:58:49 <jpena> dmsimard: you said you had pinned project-config to an earlier commit, right? Then, why is diskimage-builder using the latest master?
14:59:06 <dmsimard> jpena: only project-config is pinned
14:59:21 <dmsimard> jpena: well, it's technically not pinned, we're just not syncing it anymore
14:59:38 <jpena> dmsimard: that's what I meant. So someting else is pulling the latest master
14:59:46 <jpena> I mean, some element
14:59:51 <dmsimard> yeah
14:59:57 <jpena> ok, makes sense
15:00:05 <dmsimard> it pulls devstack master I would guess
15:00:12 <trown> dmsimard: jpena there is nothing we can do to workaround https://git.openstack.org/cgit/openstack/networking-lagopus not existing?
15:00:27 <trown> adarazs: ya that is https://bugs.launchpad.net/tripleo/+bug/1717959
15:00:28 <openstack> Launchpad bug 1717959 in tripleo "RDO upstream-centos-7 nodepool image is out of date" [Critical,Triaged] - Assigned to Alan Pevec (apevec)
15:00:41 <dmsimard> trown: looking at how to work around this right now, yes
15:00:46 <adarazs> :/
15:00:50 <trown> adarazs: whoops was scrolled up... see you already saw
15:01:07 <adarazs> trown: thanks for the confirmation though :)
15:01:19 <apevec> wth is lagopus :)
15:01:43 <rdogerrit> Merged openstack/networking-cisco-distgit rpm-master: Replaced rm -rf requirements.txt step with proper rpm macro.  https://review.rdoproject.org/r/9551
15:01:45 <rdogerrit> Merged openstack/trove-distgit rpm-master: Replaced rm -rf requirements.txt step with proper rpm macro.  https://review.rdoproject.org/r/9569
15:01:50 * apevec reassigns that LP to avoid pings ...
15:02:00 <trown> lol
15:02:12 <trown> it is always the goofy ones that break us
15:02:18 <trown> pandas and lagopuses
15:02:31 <jpena> dmsimard: maybe we could enclose line 123 in /usr/share/diskimage-builder/elements/source-repositories/extra-data.d/98-source-repositories in a "set +e ... set -e", but that could hide other issues
15:02:38 <dmsimard> jpena: I just pinged -infra about lagopous, it sounds like a bug on their end that the review merged and the things weren't created
15:02:56 * trown thinks lagopus is some cryptozoological creature and the next breakage will be caused by python-chupacabra
15:03:02 <dmsimard> jpena: they're currently troubleshooting an issue with post jobs not triggering after the gerrit upgrade so it's likely related
15:03:05 <rdogerrit> Merged openstack/watcher-distgit rpm-master: Replaced rm -rf requirements.txt step with proper rpm macro.  https://review.rdoproject.org/r/9567
15:03:05 <rdogerrit> Merged openstack/murano-agent-distgit rpm-master: Replaced rm -rf requirements.txt step with proper rpm macro.  https://review.rdoproject.org/r/9570
15:04:31 <amoralej> apevec, we need to backport  https://review.openstack.org/#/c/500212 to ocata and newton but there is no stable branch for dib 1.x
15:05:03 <amoralej> we can carry the patch in distgit or bump to 2.x in newton/ocata
15:05:24 <amoralej> i don't like doing that big change post-ga
15:05:50 <rdogerrit> Merged openstack/mistral-distgit rpm-master: Replaced rm -rf requirements.txt step with proper rpm macro  https://review.rdoproject.org/r/9580
15:08:23 <rdogerrit> Christopher Brown proposed openstack/watcher-distgit rpm-master: convert individual requires to macro  https://review.rdoproject.org/r/9510
15:08:35 <rdogerrit> Christopher Brown proposed openstack/networking-cisco-distgit rpm-master: convert individual requires to macro  https://review.rdoproject.org/r/9532
15:08:47 <jruzicka> jpena, is there a better way to detect DLRN spec then 0.LOOOONG_NUMBER.LONG_STRING?
15:10:03 <jruzicka> sematic releases are incredibly unlikely to contain numbers with more than 4 digits, like 1.1337.1, but it's still guessing
15:10:22 <jpena> jruzicka: well, it's more like 0.14digits.7hexdigits . Other than that, I can't imagine a better way
15:10:22 <pabelanger> jpena: still a while out, fixing gerrit upgrades still
15:11:23 <jschlueter> jruzicka: yea problem with checking LONG_STRING is that we get a few random cases where 7hexdigits = 1234567 or something like it
15:11:53 <dmsimard> jpena: looks like it might be a little while still for the repository fix upstream, it's not just gerrit
15:12:19 <jpena> dmsimard: ok, so... local hack in diskimage-builder?
15:12:44 <dmsimard> jpena: I'm not sure I understand why we end up with that project
15:12:44 <jruzicka> jpena, that's specific enough... if the 0.14digits.7hexdigits is guaranteed, I'll check for that.
15:12:54 <dmsimard> jpena: if we're pinned.. maybe devstack ?
15:13:18 <jpena> dmsimard: let me trace it back. I looked at it this morning, but didn't write it down
15:14:05 <jpena> jruzicka: yes, it's always the same because of the date format and short git we're using
15:14:23 <dmsimard> jpena: lol https://github.com/openstack-infra/project-config/tree/master/nodepool/elements/openstack-repos
15:14:31 <dmsimard> Url to a yaml file contains custom list of repos. The custom yaml file has the same structure as the default file: 'https://git.openstack.org/cgit/openstack-infra/project-config/plain/gerrit/projects.yaml'
15:14:44 <dmsimard> So even though we're pinned, it's using the master version of projects.yaml
15:14:57 <dmsimard> sounds like we have a good enough workaround there, sending a patch
15:16:55 <dmsimard> jpena: using the URL https://git.openstack.org/cgit/openstack-infra/project-config/plain/gerrit/projects.yaml?id=d2efe2e71ed07e17b8f0922934665dfb3a6fdbd0 should work
15:17:05 <dmsimard> that commit is the same one we have in our pin
15:17:21 <dmsimard> sending a patch
15:18:09 <jruzicka> jpena, defaults for short git were changing in git, see https://github.com/softwarefactory-project/rdopkg/commit/9a0708b0d34596125e52f89dc446f2d7a5fcedcb
15:18:37 <jruzicka> it's not guaranteed to be 7 chars now.
15:20:15 <remix_tj> Hi, i'm running tripleo-quickstart against stable/newton image. I'm hitting this bug in overcloud deployment: https://bugzilla.redhat.com/show_bug.cgi?id=1446511
15:20:16 <openstack> bugzilla.redhat.com bug 1446511 in puppet-aodh "Latest OSP10 build fails with: Error: Could not find resource 'Class[Aodh::Db]' for relationship on 'Service[httpd]' on node overcloud-controller-1.localdomain" [Urgent,Closed: errata] - Assigned to pkilambi
15:21:07 <remix_tj> i've seen that in cbs stable build there is puppet-aodh version 9.5.0-1, that i find also in the overcloud image
15:21:42 <rdogerrit> Christopher Brown proposed openstack/mistral-distgit rpm-master: remove tempest test workaround  https://review.rdoproject.org/r/9466
15:21:53 <rdogerrit> David Moreau Simard created config master: Pin the location of the project list used by upstream nodepool element  https://review.rdoproject.org/r/9680
15:21:59 <dmsimard> jpena, trown, adarazs ^
15:22:00 <rdogerrit> Christopher Brown proposed openstack/murano-agent-distgit rpm-master: convert individual requires to macro  https://review.rdoproject.org/r/9504
15:22:14 <rdogerrit> Christopher Brown proposed openstack/trove-distgit rpm-master: convert individual requires to macro  https://review.rdoproject.org/r/9506
15:22:18 <remix_tj> how can i check if that package has the fix for the bug i'm reporting?
15:22:29 <remix_tj> s/reporting/experiencing/ ?
15:22:49 <remix_tj> i think the package in cbs hasn't that fix, but i think i should check
15:22:53 <jpena> jruzicka: ok, so we have to patch DLRN to follow the same convention as rdopkg.
15:23:09 <rdogerrit> Aditya Ramteke proposed openstack/keystoneclient-distgit rpm-master:  Moved package description to global variable and replace with it.  https://review.rdoproject.org/r/9331
15:23:17 <dmsimard> remix_tj: looks like the fix is https://github.com/openstack/puppet-aodh/commit/a73d235fd9720699eaaa007d037d098fddcd801d
15:23:33 <dmsimard> remix_tj: the easiest way to know if you have that fix is to look at the actual code to see if the bits are there
15:23:51 <dmsimard> iirc modules are installed in /usr/share/openstack-puppet/modules
15:24:00 <adarazs> dmsimard: if you say so that this will work, I'll believe you. :)
15:24:10 <remix_tj> dmsimard: yes, it's the fix i was looking. Thank you. I'll check if is also in a tagged version
15:24:18 <remix_tj> and why is not available in newton cbs
15:24:29 <dmsimard> adarazs: I can explain why it should work so you don't have to blindly trust me :P
15:25:06 <dmsimard> adarazs: while our version of project-config is pinned, this nodepool element always retrieved it's project list from master: https://github.com/openstack-infra/project-config/tree/master/nodepool/elements/openstack-repos
15:25:08 <dmsimard> see the readme
15:25:52 <dmsimard> remix_tj: sure -- if there was a tag for the newton series after that patch landed, it should probably be packaged by now
15:26:13 <dmsimard> november 2016.. if it's not there, we've surely missed something
15:26:25 <chandankumar> 4 mins left
15:26:43 <chandankumar> #info, rpm-macros related reviews got cleanedup
15:27:48 <dmsimard> adarazs, trown, weshay: this project-config and image dance will likely get increasingly painful and the divergence between the upstream image and the one in review.rdo is increasing.. wish we could had discussed this at the ptg
15:27:56 <rdogerrit> Aditya Ramteke proposed openstack/shaker-distgit rpm-master: Moved package description to global variable and replace with it.  https://review.rdoproject.org/r/9333
15:28:31 <weshay> dmsimard, can we setup a gate job on it?
15:28:39 <dmsimard> weshay: a gate job on what ?
15:28:53 <jatanmalde> aditya_r, https://review.rdoproject.org/r/9333 needs rebase probably.
15:28:58 <weshay> the image
15:29:20 <weshay> if you are concerned about the divergence and reliability
15:30:03 <dmsimard> weshay: we can't gate anything, it's not something under our control at all
15:30:09 <sfbender> Javier Peña created DLRN master: Ensure the short hash is always 7 chars long in build_rpm.sh  https://softwarefactory-project.io/r/9721
15:30:10 <dmsimard> weshay: we're talking about devstack and project-config
15:30:18 <dmsimard> weshay: and the image builds are periodic
15:30:48 <dmsimard> weshay: migration to zuul v3 will hurt for sure
15:31:00 <weshay> we need a mid-cycle :)
15:31:12 <dmsimard> mid cycle is in like 3 months
15:31:15 <dmsimard> lol
15:31:50 <chandankumar> #endmeeting