13:06:05 <hhorak> #startmeeting SCLo SIG sync-up (2017-06-13)
13:06:05 <centbot> Meeting started Tue Jun 13 13:06:05 2017 UTC.  The chair is hhorak. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:06:05 <centbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:06:10 <hhorak> #meetingname sclo-sig-syncup
13:06:10 <centbot> The meeting name has been set to 'sclo-sig-syncup'
13:06:16 <hhorak> better late than never
13:06:18 <hhorak> hello :)
13:06:27 <jstanek> hello
13:06:31 <pvalena> hhorak: hi
13:11:31 <hhorak> so, one question for wide audience, has anybody saw error like this one?
13:11:32 <hhorak> Installing : centos-release-scl-rh-2-2.el7.centos.noarch                  1/2
13:11:33 <hhorak> Installing : centos-release-scl-2-2.el7.centos.noarch                     2/2
13:11:33 <hhorak> Rpmdb checksum is invalid: dCDPT(pkg checksums): centos-release-scl-rh.noarch 0:2-2.el7.centos - u
13:15:06 <hhorak> let's suppose it was an issue happening only on one computer, but if anybody saw it, let me know.. I expect it was not related to the sclo package..
13:15:08 <pvalena> hhorak: have you tried # rpm --rebuilddb ?
13:15:35 <hhorak> pvalena: it did not happen on my computer
13:15:46 <hhorak> it was just brought to my attention yesterday
13:20:55 <hhorak> Anyway, we have one another report in our centos bugs tracker: https://bugs.centos.org/view.php?id=13404
13:21:19 <hhorak> I've just directed the reporter to https://bugzilla.redhat.com/enter_bug.cgi?product=softwarecollections.org...
13:22:31 <hhorak> we don't have it written yet, but it would make sense to direct reporters who have issues with particular SCLs to bugzilla.redhat.com, since it can be then cloned to RHSCL product easier and also it has better user experience IMHO :)
13:22:57 <hhorak> does this ^ sound reasonable to others?
13:23:05 <RemiFedora> +1
13:24:42 <pvalena> hhorak: can we make the bugs/reports auto-forwarded (or -cloned) somehow?
13:24:53 <pvalena> hhorak: otherwise it sounds fine
13:26:03 <hhorak> pvalena: I don't think it makes sense to bother with it..
13:26:43 <hhorak> we can re-think it when there would be hundreds of reports, but I don't expect it will happen
13:26:52 <pvalena> hhorak: ok. It just seems a bit silly to just direct people somewhere else when they already made an effort to report a bug
13:27:00 <pvalena> (IMHO^)
13:32:00 <hhorak> pvalena: if we would report the bugs for them to bugzilla.redhat.com, we would need to also report any further updates to bugzilla report to bugs.centos.org.. if they report the bug to bugzilla themselves, they will be notified automatically about updates..
13:32:25 <hhorak> it simply does not make sense to me to connect those two reporting systems
13:32:47 <hhorak> I think we should just make this distinction clear on https://wiki.centos.org/SpecialInterestGroup/SCLo..
13:32:53 <pvalena> hhorak: is it posiible then to redirect people before they report a bug?
13:33:00 <pvalena> hhorak: yes :)
13:33:10 <hhorak> :)
13:39:25 <hhorak> And I have one another thing.. I still see failing tests for nodejs6 -- https://ci.centos.org/view/SCLo-pkg-mirror/job/SCLo-pkg-rh-nodejs6-rh-C6-buildlogs-x86_64/10/console
13:39:47 <hhorak> jstanek: did you take a look at what might be wrong there? ^
13:40:01 <hhorak> because http://cbs.centos.org/koji/packageinfo?packageID=5559 looks like properly tagged..
13:40:36 <hhorak> and it provides rh-nodejs6-npm(lodash.uniq) = 4.5.0
13:43:45 <hhorak> I have a theory, that there is again something wrong with ExclusiveArch tag, because the spec includes: ExclusiveArch: %{nodejs_arches} noarch ... and we saw issues like this before, because if some macro is not expanded properly for Arch-related tags, such package does not get through `mash`...
13:48:37 <hhorak> alphacc or kbsingh: is it possible for SIG members to see the logs from cbs -> buildlogs (IIRC the `mash` call)?
13:48:59 <hhorak> anyway, this ^ theory might be easily proven or dis-proven by building a new package without the ExclusiveArch thing..
13:49:47 <zsvetlik> hhorak: https://bugzilla.redhat.com/show_bug.cgi?id=1457763
13:50:32 <jstanek> hhorak: no, not yet… right now it seems you have more information about it than me
13:51:26 <hhorak> jstanek: that's just a guess, but I remember such troubles with nodejs before..
13:52:01 <jstanek> hhorak: ok, I will take a look at that tomorrow
13:52:19 <kbsingh> hhorak: the mash stuff runs on the cbs side, i will let alphacc comment on that
13:52:31 <hhorak> kbsingh: thx
13:52:34 <kbsingh> the push from cbs to buildlogs is just a rsync
13:52:45 <hhorak> ok
13:53:09 <Arrfab> hhorak: mash doesn't do the cbs => buildlogs, but it prepares the repo
13:53:28 <Arrfab> so after mash it's under https://cbs.centos.org/repos/
13:57:26 <hhorak> Arrfab: I think it still might be related, basically I'm missing a package rh-nodejs6-nodejs-lodash.uniq in https://cbs.centos.org/repos/sclo6-rh-nodejs6-rh-testing/x86_64/os/Packages/, even it is tagged properly (http://cbs.centos.org/koji/buildinfo?buildID=16448)
13:57:48 <hhorak> it would be great to have a way to find out why.. without bothering you :)
13:58:34 <Arrfab> hhorak: if you need a new arch, mash has to be modified to run from scratch and regen the whole repo
13:58:38 <Arrfab> let me have a look
13:59:30 <hhorak> Arrfab: my suspicious is that mash does not understand the arch, because the macro might be not expanded in the ExclusiveArch: %{nodejs_arches}
14:00:24 <hhorak> but I'm not able to prove it easily...
14:01:06 <Arrfab> hhorak: in mash, these are the arches for that tag : arches = ppc64le i386 x86_64 aarch64
14:04:40 <hhorak> Arrfab: noarch is there automatically I suppose
14:05:36 <srn_prgmr> is there a meeting today?
14:05:47 <Arrfab> hhorak: not in the mash log
14:06:00 <hhorak> Arrfab: I'm just trying to find out why the package rh-nodejs6-nodejs-lodash.uniq is not in the cbs repo..
14:06:30 <Arrfab> hhorak: wait a minute : May 23 04:52 sclo6-nodejs010-rh-testing.buildlist
14:06:46 <Arrfab> it seems that's the list that it's using
14:07:01 <Arrfab> can remove it and on next mash run it should restart from scratch
14:07:59 <sbonazzo|afk> srn_prgmr: we should have virt sig meeting gwd no updates on my side today
14:08:23 <hhorak> Arrfab: ok, let's see..
14:08:25 <hhorak> thx
14:08:28 <hhorak> anyway, we need to close the SCLo meeting I guess.... zsvetlik: don't worry, I did not ignore your issue, will take a look...
14:08:31 <hhorak> #endmeeting