Friday, 2016-11-04

rdogerritrdo-trunk proposed rdoinfo: Bump rdoinfo tags to latest upper-constraints
*** apetrich has quit IRC00:04
*** apetrich has joined #rdo00:04
*** v1k0d3n has quit IRC00:10
*** v1k0d3n has joined #rdo00:10
*** rpioso has quit IRC00:12
*** fultonj has quit IRC00:12
*** v1k0d3n has quit IRC00:14
*** fultonj has joined #rdo00:14
*** v1k0d3n has joined #rdo00:14
*** panda is now known as panda|zZ00:18
*** v1k0d3n has quit IRC00:22
*** seanatcisco has quit IRC00:27
*** links has joined #rdo00:27
*** Jeffrey4l has quit IRC00:30
*** zhugaoxiao has quit IRC00:30
rdogerritrdo-trunk created openstack/cinder-distgit: openstack-cinder: failed to build 6aceda6
*** zhugaoxiao has joined #rdo00:30
*** sdake has quit IRC00:30
*** Jeffrey4l has joined #rdo00:31
*** fandi__ has quit IRC00:34
*** limao has joined #rdo00:36
*** Pavo_ has joined #rdo00:39
*** Pavo has quit IRC00:41
*** Pavo_ is now known as Pavo00:41
*** karimb has quit IRC00:45
*** karimb has joined #rdo00:46
*** karimb has quit IRC00:48
*** percevalbot has quit IRC00:58
*** jkilpatr has quit IRC01:07
*** eaguilar has quit IRC01:10
*** Son_Goku has quit IRC01:13
*** eaguilar has joined #rdo01:19
*** jerrygb has joined #rdo01:26
*** jerrygb_ has quit IRC01:27
*** jerrygb_ has joined #rdo01:32
*** jerrygb has quit IRC01:35
*** jerrygb_ has quit IRC01:39
*** jerrygb has joined #rdo01:41
*** steveg_afk has quit IRC01:43
*** jerrygb has quit IRC01:49
*** jerrygb has joined #rdo01:50
*** mlammon has left #rdo01:55
*** chlong has joined #rdo02:06
*** jerrygb has quit IRC02:12
*** Son_Goku has joined #rdo02:13
*** jerrygb has joined #rdo02:13
*** zshi_afk is now known as zshi02:13
*** fzdarsky__ has joined #rdo02:14
*** Son_Goku has quit IRC02:14
*** rlandy has quit IRC02:14
*** Son_Goku has joined #rdo02:15
*** Son_Goku has quit IRC02:16
*** fzdarsky_ has quit IRC02:17
*** Son_Goku has joined #rdo02:17
*** rhallisey has quit IRC02:17
*** paragan has joined #rdo02:19
*** v1k0d3n has joined #rdo02:19
*** Son_Goku has quit IRC02:19
*** v1k0d3n has quit IRC02:26
*** jerrygb has quit IRC02:29
*** v1k0d3n has joined #rdo02:29
*** jerrygb has joined #rdo02:30
*** v1k0d3n has quit IRC02:34
*** hrw has quit IRC02:35
*** hrw has joined #rdo02:37
*** jerrygb_ has joined #rdo02:42
*** jerrygb has quit IRC02:44
*** crossbuilder has joined #rdo02:47
*** crossbuilder_ has quit IRC02:47
*** jerrygb has joined #rdo02:47
*** jerrygb_ has quit IRC02:50
*** dmacpher has joined #rdo02:54
*** jhershbe has joined #rdo03:00
*** gtrxcb has joined #rdo03:09
*** paragan has quit IRC03:31
*** apetrich has quit IRC03:35
*** apetrich has joined #rdo03:36
*** eaguilar has quit IRC03:45
*** sudipto has joined #rdo03:46
*** sudipto_ has joined #rdo03:47
*** imcleod has joined #rdo03:47
*** numans has joined #rdo03:51
*** imcleod has quit IRC04:00
*** imcleod has joined #rdo04:03
*** jhershbe has quit IRC04:13
*** seanatcisco has joined #rdo04:20
*** seanatcisco has quit IRC04:22
*** jerrygb has quit IRC04:24
*** abregman has quit IRC04:25
*** rdas has joined #rdo04:27
*** spr has quit IRC04:27
*** cliles has quit IRC04:41
*** ohochman has left #rdo04:48
*** satya4ever has joined #rdo04:52
*** livelace has joined #rdo04:57
*** satya4ever has quit IRC04:58
*** satya4ever has joined #rdo04:59
*** imcleod has quit IRC04:59
*** gkadam has joined #rdo05:02
*** gkadam is now known as Guest6077605:03
*** Guest60776 is now known as gkadam05:03
*** mlammon has joined #rdo05:05
*** mlammon is now known as mlammon|afk05:06
*** sudipto_ has quit IRC05:10
*** sudipto has quit IRC05:10
*** abehl has joined #rdo05:13
*** itamarjp has quit IRC05:24
*** anilvenkata has joined #rdo05:48
*** Son_Goku has joined #rdo05:49
*** saneax has joined #rdo05:52
*** vimal has joined #rdo05:53
*** rcernin has joined #rdo05:55
*** sudipto_ has joined #rdo05:56
*** sudipto has joined #rdo05:56
*** kaminohana has quit IRC06:14
*** florianf has joined #rdo06:38
*** richm has quit IRC06:41
*** florianf has quit IRC06:42
*** satya4ever is now known as satya4ever|lunch06:45
*** satya4ever|lunch is now known as satya4ever06:45
*** florianf has joined #rdo06:49
*** gfidente has quit IRC06:52
*** nyechiel has joined #rdo06:55
*** gildub_ has quit IRC07:00
*** as0bu has quit IRC07:00
*** mrunge has quit IRC07:03
*** tesseract has joined #rdo07:04
*** tesseract is now known as Guest1319407:04
*** rasca has joined #rdo07:05
*** mrunge has joined #rdo07:07
*** sdake has joined #rdo07:22
*** jprovazn has joined #rdo07:22
*** sdake has quit IRC07:28
*** sdake has joined #rdo07:31
*** pcaruana has joined #rdo07:33
*** chlong has quit IRC07:33
*** smeyer has joined #rdo07:40
*** mcornea has joined #rdo07:46
*** zoli|gone is now known as zoli|trng-afk07:48
*** dmacpher has quit IRC07:52
*** chandankumar has joined #rdo07:52
*** ccamacho has quit IRC07:53
chandankumardmellado, number80 jpena amoralej \o/07:55
*** iranzo has joined #rdo07:56
*** iranzo has quit IRC07:56
*** iranzo has joined #rdo07:56
*** gtrxcb has quit IRC07:58
*** florianf has quit IRC07:58
*** ohamada has joined #rdo07:59
*** shardy has joined #rdo08:00
*** jhershbe has joined #rdo08:04
*** florianf has joined #rdo08:04
*** ccamacho has joined #rdo08:05
*** livelace has quit IRC08:07
*** jaosorior has joined #rdo08:07
*** sdake has quit IRC08:07
*** d0ugal has joined #rdo08:08
*** fragatina has joined #rdo08:09
*** fragatina has quit IRC08:10
*** fragatina has joined #rdo08:10
*** danpawlik has joined #rdo08:18
*** athomas has joined #rdo08:19
*** d0ugal has quit IRC08:28
*** amoralej|off is now known as amoralej08:29
amoralejgood morning08:29
*** jpena|off is now known as jpena08:35
jpenagood morning08:36
*** karimb has joined #rdo08:37
*** xek has joined #rdo08:38
danpawlikgood morning :)08:39
*** jhershbe has quit IRC08:44
*** hewbrocca_afk is now known as hewbrocca08:48
*** ihrachys has joined #rdo08:48
*** paragan has joined #rdo08:50
*** paramite has joined #rdo08:53
*** d0ugal_ has joined #rdo08:58
*** gfidente has joined #rdo08:58
*** gfidente has quit IRC08:58
*** gfidente has joined #rdo08:58
*** jpich has joined #rdo08:59
*** strigazi_AFK is now known as strigazi09:04
*** gszasz has joined #rdo09:06
*** jlibosva1 has joined #rdo09:09
*** psanchez has quit IRC09:10
*** lucas-afk is now known as lucasagomes09:12
rdogerritJavier Peña proposed rdoinfo: Implement update-puppet-uc
*** nyechiel has quit IRC09:13
*** rdas has quit IRC09:13
*** psanchez has joined #rdo09:13
*** toanju has joined #rdo09:16
*** toanju has quit IRC09:16
*** toanju has joined #rdo09:17
*** pinglord has joined #rdo09:23
*** andrein has joined #rdo09:25
*** itamarjp has joined #rdo09:29
*** anilvenkata has quit IRC09:31
*** dtantsur|afk is now known as dtantsur09:32
*** fale has left #rdo09:32
*** milan has joined #rdo09:44
*** percevalbot has joined #rdo09:46
*** pinglord has quit IRC09:46
rdogerritMerged rdoinfo: Bump rdoinfo tags to latest upper-constraints
*** apevec has joined #rdo09:48
*** d0ugal_ is now known as d0ugal09:51
*** d0ugal has joined #rdo09:52
*** rhallisey has joined #rdo09:53
*** panda|zZ is now known as panda09:55
*** zshi is now known as zshi_afk10:02
*** anilvenkata has joined #rdo10:03
*** thrash|g0ne is now known as thrash10:06
rdogerritJavier Peña proposed openstack/oslo-messaging-distgit: Add python-tenacity as a requirement
*** limao has quit IRC10:11
*** athomas has quit IRC10:15
*** athomas has joined #rdo10:21
*** jlibosva1 is now known as jlibosva10:24
*** sambetts|afk is now known as sambetts10:25
*** ealcaniz has joined #rdo10:26
mrungenumber80, or anyone else, I would need help in pushing vitrageclient a bit more. Unfortunately, the review got stuck, since the submitter went AWOL10:34
mrunge bug 1379786 in Package Review "Review Request: python-vitrageclient - Python bindings to the Vitrage API" [Unspecified,Assigned] - Assigned to mrunge10:34
mrungeI really should not review my changes ;-)10:34
miscRCA as a service, wow10:35
mrungemisc, the rest is already there and reviewed10:36
jpenamrunge: I can review the package. Being a client, should this one belong to Fedora?10:36
mrungeif you want to try it, I have a copr10:36
*** paragan has quit IRC10:36
mrungejpena, yes, I think it should be a fedora review10:36
miscI could do a review too, but if jpena is on it, I can go back on reading my 2 weeks of mail backlog and see which fire do still burn10:36
mrungemisc, jpena I won't mind10:37
mrungeto be strict, I should submit a new review then in fedora10:37
mrungeand close the current one, no?10:37
jpenamrunge: mmm... yep, I think so10:37
jpenawe could change the product to Fedora, but I think it was an issue if you need to do changes to a review submitted by someone else10:38
mrungewill do then10:38
*** kbyrne has quit IRC10:40
*** abehl has quit IRC10:40
rdogerritchkumar246 created openstack/ceilometer-distgit: Seperate entrypoint for tempest plugin
*** hrybacki has quit IRC10:42
*** kbyrne has joined #rdo10:42
mrungejpena, then it's bug 1391892 in Package Review "Review Request: python-vitrageclient - Python client for Virage REST API" [Medium,New] - Assigned to nobody10:43
*** kun_huang has quit IRC10:43
*** setkeh has quit IRC10:43
*** afazekas has quit IRC10:44
*** OlaNys has quit IRC10:44
*** rivorozec has quit IRC10:44
*** zhenguo has quit IRC10:44
*** coredump has quit IRC10:44
*** Duck has quit IRC10:44
*** tdasilva has quit IRC10:44
jpenamrunge: ack, on my way to review it10:44
*** OlaNys has joined #rdo10:44
mrungeawesome, thank you!10:44
*** hrybacki has joined #rdo10:46
*** kun_huang has joined #rdo10:46
*** coredump has joined #rdo10:46
*** zhenguo has joined #rdo10:47
*** openstackgerrit has quit IRC10:47
*** openstackgerrit has joined #rdo10:48
*** ChanServ sets mode: +v openstackgerrit10:48
*** setkeh has joined #rdo10:48
*** setkeh has joined #rdo10:48
*** afazekas has joined #rdo10:49
rdogerritAlfredo Moralejo created openstack/oslo-rootwrap-distgit: Update to 5.1.1
*** afazekas has quit IRC10:54
*** afazekas has joined #rdo10:55
*** Duck has joined #rdo10:56
jpenamrunge: sent some comments in the review10:57
*** steveg_afk has joined #rdo10:57
rdogerritJon Schlueter proposed rdoinfo: Add stable branch for python-networking-odl for newton
mrungejpena, thank you, will address your comments ASAP10:58
rdogerritchkumar246 created openstack/gnocchi-distgit: Seperate entrypoint for tempest plugin
*** pbandark has joined #rdo11:00
amoralejvkmc, about locale11:03
vkmcamoralej, we create subpackages for them, right?11:03
amoralejit depends11:03
vkmcand remove locale/ from the codebase11:03
vkmcdepends on?11:04
amoralejwhat we do is11:04
amoralej1. compile .po into .mo files11:04
amoralej2. move .mo to the right location (under /usr/share/locale )11:05
amoralej3. remove .po files11:05
amoralejand then we have two cases11:05
amoralejfor libraries that support py3 packaging, we create a subpackage -lang11:05
rdogerritJon Schlueter created openstack/muranoclient-distgit: remove stale sources file
*** andrein has quit IRC11:06
vkmcall right11:06
amoralejfor packages without py3 packaging we left it into the -common package11:06
jpenanumber80: could you tag in -testing?11:06
jpenanumber80: it's required by python-astroid11:07
apevecis that new dep?11:07
amoralejvkmc, do you have any particular issue with it?11:07
apevecjpena, it has been in -candidate since January ?11:07
*** jkilpatr has joined #rdo11:07
vkmcamoralej, not really with packaging itself...11:08
vkmcit's about this patch set I submitted for Manila UI11:08
rdogerritMerged openstack/oslo-messaging-distgit: Add python-tenacity as a requirement
jpenaapevec: it was built the same day as astroid, so I guess it was just forgotten when tagging11:08
vkmcI'm performing some checks over locale/ because Django will fail to start if it doesn't find locale/11:09
amoralejvkmc, for dashboard packages it's slighly different11:09
vkmcand I got a core reviewer asking me to remove that... because he claims it's needless since we already know they are there11:10
apevecjpena, so how was astroid working until now?11:10
vkmcwhich made me think of RDO packages and the way we handle locales11:10
*** richm has joined #rdo11:10
vkmcdoes that make sense?11:10
jpenaapevec: I think we've never used it. I know it's a dependency for freezerclient (seen in the upstream rpm-packaging spec)11:10
apevecah ok11:11
apevecjpena, tagged11:11
amoralejfor dashboard packages we don't need to move compiled messages into a different directory because of how django looks for .mo files11:11
amoralejso we don't remove the locale directory11:12
amoralejlet me check one11:12
amoralejvkmc, you can see in rpm -qlp|grep locale11:14
vkmcamoralej, all right, thanks11:14
amoralejfrom a rdo perspective it's safe to get rid of the check11:14
amoralejalthough it wouldn't hurt...11:14
amoralejbtw, this reminds me we need to add i18n into manila-ui package in RDO11:15
vkmcamoralej, coolio, let me know if I can give a hand with that11:15
amoralejare you taking care of manila stuff packaging in RDO?11:16
vkmcamoralej, not really, but I'm an engineer for Manila so I can help with it11:16
vkmcI used to do it for a project I was working before Manila11:17
amoralejit's not big deal, i can manage it unless you are interested in getting some experience with packaging at RDO, :)11:18
vkmcamoralej, k, thanks for that :)11:18
amoralejupstream project engineers are allways welcome in RDO-land11:19
*** alexismonville has joined #rdo11:19
rdogerritrdo-trunk created openstack/oslo-config-distgit: python-oslo-config: failed to build 30af7de
* amoralej checking ^11:20
rdogerritrdo-trunk created openstack/openstacksdk-distgit: python-openstacksdk: failed to build c646788
jpenamrunge: one last thing on the vitrageclient review. When testing the package, it fails because it tried to import oslo_log. I see it's fixed on master (, but we'd need it in the spec until there is a new release. Do you mind adding it?11:21
rdogerritrdo-trunk created openstack/glance-distgit: openstack-glance: failed to build d1939eb
*** pblaho has joined #rdo11:22
mrungejpena, will do11:22
mrungejpena, I can't tell why the builds still worked for me11:22
rdogerritJon Schlueter proposed rdoinfo: Add stable branch for python-networking-odl for newton
*** alexismonville has quit IRC11:24
jpenaanyone looking at the rdo-FTBFS errors in master? It looks like the mock dir was stuck11:24
rdogerritJon Schlueter proposed rdoinfo: Add stable branch for python-networking-odl for newton
*** jprovazn has quit IRC11:26
mrungejpena, so, you're expecting checks being executed during build, right?11:28
jpenamrunge: well, they're optional. That said, if we run them it's better11:28
mrungeok, will look into that right now11:29
rdogerritJon Schlueter proposed puppet/puppet-stdlib-distgit: Update to 4.13.1 (f2492ee916c1c8e0345514045432c4a049674029)
*** mburned_out is now known as mburned11:31
ihrachysapevec: btw remember I once said that does not work for safari? I had a duty to validate it's not a single machine issue. well, jlibosva told me lately he has the same issues.11:32
apevecjlibosva, you too, macos user!11:33
apevecjpena, what was stuck?11:34
jlibosvaapevec: :) I am! Though I use chrome and it doesn't work for me in Safari either11:34
jpenaapevec, one of the tmpfs directories created by mock failed to unmount11:34
*** dprince has joined #rdo11:36
rdogerritchkumar246 created openstack/neutron-distgit: Seperate entrypoint for tempest plugin
apevecjpena, but which ftbfs is it? I see only one which you tracked down?
apevecah amoralej already abandoned those caused by stuckage(sp?)11:39
jpenathat's it11:39
amoralejcurrent dlrn run is working fine11:39
*** tdasilva has joined #rdo11:40
*** d0ugal has quit IRC11:40
ihrachysjlibosva: keep me honest @
jlibosvaihrachys: so was it officially decided we'll split?11:41
apevecihrachys, this workaround until tempest plugin is separate project11:41
ihrachysjlibosva: I think the general feeling on that session was yeah, we better do, since everyone does11:41
apevecjlibosva, there wasn't much opposition in dmellado's session at summit11:41
apevecihrachys, so we want to do it in rdo for newton11:41
ihrachysjlibosva: we should probably have discussion in upstream specific to neutron11:42
chandankumarihrachys,  regarding seperating tempest plugins from component is targetted for Pike. it is just a workaround for ocata release11:42
apevecand yes, we'll help upstream but that's targeting Pike iiuc11:42
* dmellado comes back for a second11:42
dmelladoyes, we'll propose that for Pike11:42
dmelladoand jlibosva I'll have you drink one of those cocktails if you oppose it11:43
ihrachysapevec: I don't see goals this way. if we plan to do it tomorrow, we can start pursuing it now11:43
apevecso we can carry workaround in rdo for newton and ocata11:43
apevecthere can be parallel goals11:43
apevecthis is just to make sense with our subpackaging which we already did in rdo11:43
*** sreichar has quit IRC11:44
ihrachysdoes everyone land those patches?11:44
apevecother projects in rdo? yes, we started with example in cinder11:45
*** d0ugal has joined #rdo11:45
jlibosvaihrachys: I tend to agree with apevec, for Newton we won't split repos so WA might be a way to go11:45
ihrachysjlibosva: wa == workaround?11:45
dmelladojlibosva: ihrachys do you have any concerns re those workarounds?11:45
apevecchandankumar, btw, cinder epi patch needs rebase11:45
jlibosvadmellado: I need to de-cypher that11:46
*** hexo_ has joined #rdo11:46
*** ealcaniz has quit IRC11:46
ihrachysdmellado: more magic in packaging? :) fake eggs not reflecting upstream? :)11:46
*** sdake has joined #rdo11:46
ihrachysI mean, nothing huge, just uneasy11:46
dmelladoihrachys: that's only for the time being, and it'll be dropped in P11:46
*** eaguilar has joined #rdo11:47
ihrachysdmellado: have we considered keeping a macro for that in common place?11:48
ihrachysseems rather template-able11:48
dmelladoihrachys: that could be done11:48
apevecwe'll add macro11:48
apevecjust wanted first iteration with direct patch to see it working11:49
dmelladojlibosva: I was told about your cocktail by ltomasbo11:49
apevecnow, that's an unexpected change of topic :)11:49
dmelladoapevec: I was replying jlibosva from before xD11:50
dmelladobut yeah, coming back from the macro, we could put it up in some repo11:50
apevecor is that jedi distraction trick? :)11:50
dmelladobut I was wondering where11:50
apevecdmellado, we have place11:50
apeveccurrently rdo-rpm-macros, but working on upstreaming it to rpm-packaging/openstack-macros11:50
dmelladoapevec: ack11:50
apevecright and it will end here
ihrachysapevec: dmellado: ok, I removed -2, but put -1 anyhow for some minor stuff just to keep you mad11:51
dmelladoihrachys: heh11:51
*** seanatcisco has joined #rdo11:51
*** zoli|trng-afk is now known as zoli11:52
*** zoli is now known as zoli|lunch11:52
apevecihrachys, thanks :)11:52
*** gkadam has quit IRC11:53
ihrachysapevec: love being mad? :)11:53
chandankumarihrachys, thanks updating!11:54
apevecno, it allows me to exercise jedi-anger-controll skillz :)11:54
dmelladoapevec: don't go to the dark side xD11:55
dmelladoapevec: chandankumar ihrachys in any case let me know if you need any clarifications on this11:55
dmelladoI'll be mostly offline today due to personal issues but I'll be checking emails11:56
*** morazi has quit IRC11:57
*** dpeacock has joined #rdo11:57
ihrachysdmellado: ack. it could probably help folks to see some write-up on the plan, so that you could link to it on any questions and in commit message.11:59
ihrachysthen maybe I would not bug you that much with -2s11:59
*** jayg|g0n3 is now known as jayg12:03
*** fultonj_ has joined #rdo12:06
rdogerritchkumar246 proposed openstack/cinder-distgit: Seperate entrypoint for tempest plugin
*** chlong has joined #rdo12:13
*** xek has quit IRC12:17
rdogerritchkumar246 proposed openstack/neutron-distgit: Seperate entrypoint for tempest plugin
ihrachysapevec: dmellado: jlibosva: +2d ^12:23
ihrachyslet it burn12:23
*** jerrygb has joined #rdo12:23
*** sudipto_ has quit IRC12:25
*** sudipto has quit IRC12:25
*** imcleod has joined #rdo12:28
*** jerrygb has quit IRC12:31
*** eaguilar has quit IRC12:32
*** maeca1 has joined #rdo12:33
*** ohochman has joined #rdo12:35
*** hexo_ is now known as wrongosaurus12:35
rdogerritIhar Hrachyshka created rdopkg: Fixed -c argument parsing
ihrachysjruzicka: ^12:36
*** rlandy has joined #rdo12:36
rdogerritchkumar246 created openstack/neutron-fwaas-distgit: Seperate entrypoint for tempest plugin
*** eaguilar has joined #rdo12:41
*** jerrygb has joined #rdo12:42
rdogerritchkumar246 proposed openstack/aodh-distgit: Seperate entrypoint for tempest plugin
rdogerritchkumar246 proposed openstack/ceilometer-distgit: Seperate entrypoint for tempest plugin
*** ohochman has left #rdo12:45
*** chlong has quit IRC12:46
*** rodrigods has quit IRC12:46
*** rodrigods has joined #rdo12:46
rdogerritchkumar246 proposed openstack/gnocchi-distgit: Seperate entrypoint for tempest plugin
*** alexismonville has joined #rdo12:47
rdogerritchkumar246 proposed openstack/keystone-distgit: Seperate entrypoint for tempest plugin
*** lucasagomes is now known as lucas-hungry12:52
*** percevalbot has quit IRC12:53
dmsimardhello #rdo12:54
*** percevalbot has joined #rdo12:54
dmsimardamoralej: weirdo jobs looking much more green this morning, thanks for pointing out the log issue :(12:54
*** maeca1 has quit IRC12:55
*** morazi has joined #rdo12:55
amoralejyeap dmsimard, we've even promoted newton12:55
dmsimardstill seeing sporadic timeouts though12:55
dmsimardespecially p-o-i 002 ?12:55
dmsimardI'll follow .. it's not normal that jobs are timing out at 90 minutes12:56
miscdmsimard: hey, I am back from PTO, and did see you have added me to review stuff, do you still need that (and is that urgent)12:57
dmsimardmisc: yup, I still need that -- not urgent. By early next week would be nice ?12:58
dmsimardARA also helped me figure out a callback issue in 2.2 :)12:58
*** spr has joined #rdo12:58
*** jcoufal has joined #rdo12:58
miscdmsimard: I would do that, yeah, I am just reading the backlog and stuff12:59
ihrachysapevec: wassup with ?12:59 bug 1379646 in Package Review "New package: python-networking-bagpipe" [Unspecified,New] - Assigned to apevec12:59
ihrachysbgpvpn package it is12:59
ihrachyssorry bagpipe13:00
apevecI was taking it as a first step to reassign it, but I guess I could just review it myself :)13:00
apevecI think I have fedorea-review report already13:00
*** jerrygb has quit IRC13:01
dmsimardapevec: I'm requesting a fedora package for the first time, what do I select in "collections" ? >= f24 ?13:02
*** morazi has quit IRC13:02
dmsimardand epel7 ?13:02
rdogerritchkumar246 proposed openstack/neutron-fwaas-distgit: Seperate entrypoint for tempest plugin
*** unclemarc has joined #rdo13:03
*** kgiusti has joined #rdo13:05
amoralejdmsimard, yes we are hitting 90 mins timeouts in p-o-i jobs and even in the runs that pass we usually are pretty close to it. Maybe we should increase it13:05
dmsimardI really don't want to do that13:05
dmsimardJobs finish in <60 minutes in review.r.o on VMs and in <30 minutes in upstream gate13:05
dmsimardIf we're taking ~90 minutes on bare metal there is something we're missing13:06
*** maeca1 has joined #rdo13:06
miscyes, a few more cpu :p13:06
jpenaand network bandwidth maybe13:06
dmsimardI can concede that the AMD CPUs in cico suck, but it's still bare metal.. with 32GB RAM and SSDs...13:06
dmsimardand still 8 cores13:06
amoraleji have the impression that jobs went longer after latest change in ci.centos13:07
miscwhat about a jenkins jobs that run some bechmark on the different env to see how they compare13:07
dmsimardI'm following in shell right now13:07
miscand the I/O sucks on the chassis used by ci.centos :/13:07
dmsimardand puppet is supppppppppperrrrrrrrr slow13:07
amoralejor more variable13:07
dmsimardlook at the timestamps:
ihrachyschandankumar: fwaas and neutron are the only ones in neutron world right?13:07
chandankumarihrachys, nope13:08
*** jhershbe has joined #rdo13:08
chandankumarihrachys, yes13:08
ihrachysvpnaas does not have one I guess13:08
dmsimardiostat has 0% i/o wait13:08
ihrachys(the subproject does suck lately on all fronts)13:08
dmsimardCPU is bored, load 213:08
chandankumarihrachys, ,
chandankumarit has!13:09
chandankumarcurrently i am working on lbaas!13:09
dalvarezsihrachys, would it be a "good practise" to include a bash command to capture a variable such as CPIDS=$(blabla) and use it in a second command separated by ;? i don't know if that's typical.. tried to find something similar though13:09
ihrachysnice of them13:09
miscdmsimard: so network latency ?13:09
ihrachysjlibosva: I blessed*+branch:rpm-master+topic:tempest-plugin-entrypoint would love to hear from you too13:09
jpenadmsimard: are you connected to the slave?13:09
miscdmsimard: memory badnwidht ?13:09
miscpeople randomly adding sleep just to annoy us ?13:09
dmsimardjpena: I'm connected to the duffy node, the slave only runs ansible13:10
ihrachysdalvarezs: I think yes it should but I would need to see the code13:10
apevecdmsimard, yes, >=f24 and epel713:10
jpenadmsimard, try downloading a large rpm file from trunk.rdo, for example and see how fast it is13:11
apevecso we have network bottleneck?13:11
* jpena thinks so13:11
dmsimardjpena: instantaneous13:11
dmsimard2016-11-04 13:11:36 (34.6 MB/s) - ‘openstack-dashboard-10.0.0-0.20161104124155.923e9b2.el7.centos.noarch.rpm’ saved [11468600/11468600]13:11
dmsimardjpena: so I have a theory, and I already talked with EmilienM and mwhahaha about this13:12
*** d0ugal has quit IRC13:12
apevecone thing that is different in AMDs is that they are actually NUMA right?13:12
dmsimardThe thing that sucks with "all in one" controllers is the wsgi implementation in puppet13:12
*** pradk has quit IRC13:12
dmsimardbecause when something behind httpd is scheduled for a refresh, it restarts httpd and every service behind it13:12
dmsimardand then there is a wsgi warm up time13:13
*** pradk has joined #rdo13:13
apevecwhy not just httpd reload ?13:13
dmsimardQ: Why do requests against my application seem to take forever, but then after a bit they all run much quicker?13:13
*** links has quit IRC13:13
dmsimardI'm pretty sure puppet does a restart, not a reload.. I could be proven wrong, though13:14
apevecdmsimard, so we need few fake api requests to avoid timeouts in real tests?13:14
dmsimardI don't follow13:15
apevecjust to get wsgi app loaded13:15
apevecdmsimard, I'm not sure what is your theory13:15
*** zoli|lunch is now known as zoli13:15
apevecis it that due to restarts, we see timeouts in tempest ?13:15
*** morazi has joined #rdo13:15
*** zoli is now known as zoliXXL13:15
dmsimardno, I'm talking about the general length of the jobs13:15
dmsimardtempest doesn't do anything that'd cause httpd to restart13:16
apevecso at that point wsgi should be warm?13:16
apevecit is just initial installation which is slow13:16
dmsimardyeah, but the puppet run is pretty darn slow -- puppet does a lot of call to APIs to authenticate, create endpoints, stuff like that13:16
miscmaybe that's cold becuse winter is here ?13:17
dmsimardand restarts httpd throughout the puppet run13:17
dmsimardI've noticed the issue first hand when setting up the cico cloud13:17
dmsimardthe API is very slow/sluggish when it's not "exercised"13:17
apevecand there isn't a way to avoid all those httpd restarts?13:19
*** akshai has joined #rdo13:19
dmsimardYou can reload individual wsgi services13:20
dmsimard Q: Do I have to restart Apache every time I make a change to the Python code for my WSGI application?13:20
dmsimardbut puppet doesn't do that right now13:20
dmsimardIt may not have been so much of a problem in the past since APIs were running on eventlet a lot13:20
dmsimardbut everything is moving to wsgi/apache nwo13:20
*** d0ugal has joined #rdo13:21
*** jhershbe has quit IRC13:21
miscbut so, that would be seen all across the board, no ?13:21
*** jprovazn has joined #rdo13:22
dmsimardyeah, but slow AMD procs exacerbates the issue :)13:22
*** dustins has joined #rdo13:23
apevecso that's the real question, why is amd that slower?13:23
apevecmisc did suggest mem bandwidth issue, and amd is numa so it could be part of the answer13:24
apevecmisc, do you have a suggestion how to measure/monitor that?13:25
miscapevec: nope13:25
apevecgoogle says
amoralejjust two compare two execution of the same job
amoralejeven restarting the network takes double time13:28
dmsimardI realized the AMD procs are 1.2GHz .. wow13:29
dmsimardah wait, no, it scales up and down13:29
miscso maybe that's the issue, it keep scaling ?13:29
dmsimardit keeps bouncing from 1 Ghz to 2 Ghz13:29
amoralejthat's a good catch dmsimard13:29
dmsimardI'll ask the CentOS guys13:30
amoralejwe should disable cpu scaling13:31
dmsimardCPU bouncing:
amoraleji'm going for lunch13:32
*** amoralej is now known as amoralej|lunch13:32
jpenamy CPU also keeps changing speeds, but that alone shouldn't justify the 50% difference13:32
*** ccamacho is now known as ccamacho|lunch13:32
rdogerritchkumar246 created openstack/neutron-lbaas-distgit: Seperate entrypoint for tempest plugin
jpenado we have any comparison between the two nodes?13:33
dmsimardjpena: yeah but I'm just saying there'd probably be a difference if the turbo whatever would be disabled and pinned to 2 Ghz13:33
jpenadmsimard, yep maybe13:33
dmsimardI haven't seen /anything/ clocked at 1 Ghz for a long time13:33
apeveclaptops go down 800MHz when idle13:34
dmsimardapevec: even my cell phone doesn't drop that low :p13:34
* misc throw a raspberry pi on dmsimard 13:34
dmsimardit's ARM though13:35
*** d0ugal has quit IRC13:35
* dmsimard should run OpenStack on his cell phone13:35
miscyou might need more ram13:36
dmsimardyeah, only 3GB :(13:37
*** eharney has joined #rdo13:37
*** d0ugal has joined #rdo13:37
* number80 is around13:38
number80seems that I'll keep my 10 fingers whole13:38
dmsimardapevec: hm, KB mentioned tuned and I saw that the profile was set to "balanced"13:39
*** jerrygb has joined #rdo13:39
*** jpena is now known as jpena|lunch13:39
dmsimardwhat should we be setting ? throughput-performance ? virtual-host ?13:39
dmsimardI set it to throughput-performance and I don't see any bouncing anymore13:39
*** jerrygb has quit IRC13:39
apevecnumber80, what did you do, play with katana again??13:41
apevecdmsimard, can we make that permanent and see if it helps?13:41
dmsimardapevec: it's a fresh install every time .. so I'd have to hook it into the job13:41
dmsimardor in weirdo13:41
apeveccico nodes are reprovisioned right13:41
apevecyeah, it will be faster in weirdo13:42
dmsimardwhat profile do we want, though ?13:42
dmsimardis that the best one ? throughput-performance ? seems mostly filesystem based tweaks but I don't see the CPU bouncing anymore ..13:42
number80apevec: no, another infection at the same finger13:42
apevecmake-it-fast if there's such :)13:43
dmsimardah wait, throughput-performance seems fine13:43
dmsimardA server profile for typical throughput performance tuning. It disables tuned and ktune power saving mechanisms, enables sysctl settings that improve the throughput performance of your disk and network I/O, and switches to the deadline scheduler. CPU governor is set to performance.13:43
jaygnot sure if this helps, but we used to use birtual_host for compute nodes in the astapor days13:43
* number80 is not bad at swordmanship13:43
*** jerrygb has joined #rdo13:46
*** jerrygb has quit IRC13:48
*** d0ugal_ has joined #rdo13:51
*** d0ugal has quit IRC13:53
number80jruzicka: the last two reviews should be quick
rdogerritMerged puppet/puppet-mongodb-distgit: Update to 0.16.0
number80I haven't experienced Ihar's issue though but I'm using a fork13:56
*** lucas-hungry is now known as lucasagomes13:56
*** Guest13194 has quit IRC13:58
rdogerritDavid Moreau Simard created rdo-infra/ci-config: Stop running a separate playbook for removing firewall and nm
*** fultonj_ has quit IRC14:00
rdogerritDavid Moreau Simard created rdo-infra/weirdo: Setup tuned profile by default in weirdo for optimal performance
rdogerritMerged openstack/rally-distgit: Update to 0.7.0
*** fultonj has quit IRC14:01
rdogerritDavid Moreau Simard proposed rdo-infra/weirdo: Setup tuned profile by default in weirdo for optimal performance
*** Son_Goku has quit IRC14:02
dmsimardweshay: fyi ^
*** fultonj has joined #rdo14:02
dmsimardweshay: *maybe* found one of the factors why AMD nodes were so slow14:02
weshaydmsimard, cool.. will look thanks14:04
dmsimardweshay: I'll let you know how it turns out14:04
dmsimardapevec, number80: do I set co-maintainers for new fedora packages ?14:05
apevecdmsimard, you can add me14:05
rdogerritMerged rdo-infra/ci-config: Stop running a separate playbook for removing firewall and nm
apevecnumber80, is super-hero^Wpackager anyway14:05
number80dmsimard: yep, add apevec at least14:05
* number80 has already access14:05
dmsimardapevec: so just "apevec" ?14:05
dmsimardok, requested, what now ?14:07
pandais haproxy package coming from dlrn repos ?14:07
rdogerritchkumar246 created openstack/neutron-vpnaas-distgit: Seperate entrypoint for tempest plugin
*** eggmaster has joined #rdo14:07
apevecdmsimard, wait :) afaik it's automated so it will be quick14:08
chandankumarihrachys, tempest entry point seperation for neutron and its plugins are done!14:08
dmsimardpanda: definitely not dlrn -- CBS at best14:08
dmsimardpanda: dlrn only packages openstack projects14:08
apevecnope, that's base EL14:08
dmsimardyeah I don't see haproxy in -deps14:08
apevecbest public ref is
dmsimardchandankumar: good job14:08
ihrachyschandankumar: yay. I +2d all of them, thanks14:08
ihrachyschandankumar: you will prolly need jlibosva to chime in with the second one14:09
chandankumaronly 8 more components left!14:09
*** Goneri has joined #rdo14:09
pandadmsimard: ok, who's is to warn if an important update is coming to haproxy ?14:10
apevecpanda, what do you mean?14:11
*** dtantsur is now known as creepy_owlet14:11
apevecpanda, for base RHEL there must be rhbz ?14:11
pandaapevec: this fixed a problem where haproxy 1.5.x reports all good to systemd even if it fails to start.14:12
rdogerritchkumar246 proposed openstack/heat-distgit: Seperate entrypoint for tempest plugins for python-heat-tests
apevecpanda, so that was fix upstream?14:12
apevecyou need to file rhbz14:13
*** Guest13194 has joined #rdo14:13
apevecif you want that fix in EL714:13
pandaapevec: ok, thanks.14:13
*** albertom has quit IRC14:15
miscanyone against a reboot of rdoproject today ?14:16
misc(the webserver, for kernel upgrade)14:16
*** zoliXXL is now known as zoli|brb14:17
dmsimardhm, the only thing that comes to mind is the http redirect for .repo files14:17
*** albertom has joined #rdo14:18
dmsimarder, rpm repository files14:18
dmsimardI would try to do it "off business hours" if possible14:18
*** saneax is now known as saneax-_-|AFK14:19
dmsimardwow, just got like 25 emails from fedora for package creation14:21
miscI can do that tommorow too14:21
*** apetrich has quit IRC14:21
EmilienMdmsimard: woot14:23
dmsimardnow we need to pull that into CBS14:23
*** apetrich has joined #rdo14:23
creepy_owletlucasagomes, have you had a chance to start virtualbmc packaging?14:26
lucasagomescreepy_owlet, no I have not. Only ironic-staging-drivers... I need to update the pyghmi package (vbmc depends on a newer version of pyghmi)14:27
lucasagomesit's on my radar tho14:27
lucasagomesunless someone wants to take the firts tab at it14:27
creepy_owletI see. I might be able to, but not right now14:27
lucasagomesyeah, it has few dependencies tho, it should be grand14:28
*** ccamacho|lunch is now known as ccamacho14:28
dmsimardnumber80, apevec: the next piece of the puzzle I don't know about yet, how do we get in cbs ?14:28
*** jistr is now known as jistr|call14:29
apevecdmsimard,  you need to push it to distgit
*** jaosorior is now known as jaosorior_mtg14:29
apevecthen koji build14:29
apevecfedpkg build14:29
dmsimardapevec: ah, I actually have to push the spec file ?14:29
apevecdmsimard, yes, fedpkg co python-marathon14:29
chandankumardmsimard, fedpkg clone python-marathon14:29
apevecthen commit your latest spec from review14:29
chandankumardmsimard, fedpkg import srpm14:29
number80dmsimard: someone will take care of that14:29
chandankumargit commit <message>14:30
number80it's magic!14:30
lucasagomesbtw, RDO (the project) should kinda of fix the RPM packaging guide. There are two at the moment: and
chandankumargit push14:30
chandankumarfedpkg build14:30
chandankumarthen fedpkg switch-branch14:30
chandankumarand do the same14:30
*** egallen has joined #rdo14:30
chandankumaryou are done!14:30
number80lucasagomes: nope, one is process, the other is about guidelines14:30
chandankumardmsimard, do fedpkg update on f25, f24 and epel7 branches14:30
apevecdmsimard, once you have it in fedora distgit, I'll rebuild it from there in cbs14:30
*** toanju has quit IRC14:31
*** jaosorior_mtg has quit IRC14:31
*** jaosorior_mtg has joined #rdo14:32
rdogerritMerged openstack/neutron-distgit: Seperate entrypoint for tempest plugin
*** paragan has joined #rdo14:32
*** paragan has joined #rdo14:32
apevecnova 14.0.2 coming
dmsimardchandankumar: hmm, I'm missing something I think:
dmsimardit looks like it doesn't download the source properly ?14:35
apeveclooks like import srpm didn't work well14:35
dmsimardI didn't actually import srpm cause that didn't work14:35
apevecdmsimard, fedpkg new-sources <tarball>14:35
dmsimardCould not execute import_srpm: File not found..14:36
apevecah then you need new-sources14:36
*** zoli|brb is now known as zoli14:36
dmsimardok, trying14:36
apevecto u/l to look-aside cache14:36
*** d0ugal_ has quit IRC14:39
*** d0ugal has joined #rdo14:39
dmsimardapevec: it's weird, I need to download the source files ?14:40
dmsimarddoesn't mock/koji take care of that ?14:40
*** jmelvin has joined #rdo14:40
dmsimardlike the README for docs, and the .tar.gz source14:40
*** fragatina has quit IRC14:42
*** fragatina has joined #rdo14:42
*** sudipto has joined #rdo14:43
*** sudipto_ has joined #rdo14:43
chandankumardmsimard, How to add a new project to r.o.o?14:44
*** toanju has joined #rdo14:45
dmsimardchandankumar: you can look at an example here:
*** amoralej|lunch is now known as amoralej14:46
dmsimardok I have a new fedpkg build going.. let's see14:47
rdogerritrdo-trunk created openstack/rally-distgit: openstack-rally: failed to build 7d76719
chandankumarsorry dmsimard it was taking about review.rdoproject.org14:48
dmsimardchandankumar: what kind of project ?14:48
chandankumardmsimard, opstools was looking to add a new project14:49
dmsimardchandankumar: yeah, mrunge was supposed to get back to me on that14:49
*** cliles has joined #rdo14:49
*** dmacpher has joined #rdo14:49
dmsimardrawhide successful \o/
*** jpena|lunch is now known as jpena14:50
dmsimardso I just push to all other branches and do a build there too ?14:51
mrungedmsimard, I got a bit pushback on enabling sf on opstools-ansible . let's try with a different project then?14:51
*** jerrygb_ has joined #rdo14:52
dmsimardmrunge: up to you, let's do it monday ?14:52
mrungedmsimard, sounds fine with me14:53
mrungedmsimard, has no pull requests now. that'll be a good candidate for moving to sf14:54
mrungebut we can do in on Monday too14:54
dmsimardmonday would be better for me14:55
* chandankumar will connect from home after an hour!14:58
dmsimardapevec: I was able to do it:
ihrachysapevec: I already hit the 3rd bug in rdopkg today :-|14:59
dmsimardapevec: if you can pull that into cbs, I'll add the dependency in magnum for it14:59
dmsimardihrachys: maybe jruzicka can help15:00
ihrachysapevec: now it does not want to regenerate patches with rdopkg patch saying there is nothing to regenerate; but update-patches does15:00
ihrachysdmsimard: I haven't seen him today15:00
*** jerrygb has joined #rdo15:01
*** jistr|call is now known as jistr15:03
ihrachysI will report a bug I guess...15:03
*** Son_Goku has joined #rdo15:04
ihrachysactually, maybe is related15:04
*** jerrygb_ has quit IRC15:04
*** sadasu has joined #rdo15:06
*** pbandark has quit IRC15:07
rdogerritrdo-trunk created openstack/rally-distgit: openstack-rally: failed to build 7d76719
rdogerritchkumar246 created openstack/tempest-distgit: Update to post 13.0.0 (296cca79678463dd90ca2694245bb7b89f8405d8)
*** jistr is now known as jistr|biab15:10
chandankumarapevec, we need a new package os_faults for fixing rally ftbfs!15:10
chandankumarapevec, i will get the package review up by tonight!15:10
chandankumarafazekas, tempest newton spec updated with the backports landed in tempest newton branch
chandankumarbrb after an hour!15:12
*** chandankumar has quit IRC15:12
rdogerritIhar Hrachyshka proposed rdopkg: Fixed -c argument parsing
imcsk8good morning!15:14
*** jistr|biab is now known as jistr15:14
*** vimal has quit IRC15:18
*** READ10 has joined #rdo15:19
*** sudipto_ has quit IRC15:19
*** sudipto has quit IRC15:19
*** emalikov has joined #rdo15:20
sadasuHello! I am trying to bring networking-cisco tag 4.0.0 into RDO so that it can make it into OSP1015:22
dmsimardsadasu: OSP10 is newton, newton has been out for a while already15:23
sadasusome questions: 1.which repo should I be looking at to determine the current state of networking-cisco in RDO?15:23
sadasudmsimrad: that is correct and networking-cisco is also part of it15:24
dmsimardsadasu: this is the versions we are pinning at:
dmsimardlack of versions means it's unpinned (builds from master) -- but that's for trunk builds15:24
sadasuwe recently came out with a tagged version of networking-cisco for newton, and want to make sure RDO is pointing to that tag15:24
dmsimardit looks like we are using 3.2.1 for newton right now
dmsimardIt sounds like 3.2.1 to 4.0.0 is a significant bump15:25
*** rcernin has quit IRC15:25
*** rcernin has joined #rdo15:27
sadasuthat is correct, yet, we do want to move to 4.0.015:27
*** _elmiko is now known as elmiko15:27
dmsimardso basically you want to clone, check out newton-rdo and then update the spec to use 4.0.015:28
dmsimardit'll create a code review on https://review.rdoproject.org15:28
dmsimardwhere it'll be tested and approved if everything looks okay15:28
sadasuyes, that is correct. Also, want to make sure that this route will ensure we will make it to OSP 1015:29
dmsimardjschlueter: ^ ?15:29
rdogerritEmilien Macchi created rdoinfo: osprofiler: pin 1.4.0 for Liberty and Mitaka too
EmilienMslagle, panda ^15:30
EmilienMapevec: please review when you have time ^15:30
*** eaguilar has quit IRC15:32
*** absubram has joined #rdo15:32
*** jprovazn has quit IRC15:35
*** vimal has joined #rdo15:37
*** emalikov has quit IRC15:38
*** Guest13194 is now known as tesseract-15:38
*** eaguilar has joined #rdo15:39
jschlueterdmsimard: /me reading back15:40
*** karimb has quit IRC15:40
jschlueterdmsimard: ack sounds good to me15:41
jschlueterI'll keep an eye out for it15:41
*** karimb has joined #rdo15:41
jschluetersadasu: it's a bit late for these changes to be coming in but yea ... better late than never15:41
sadasujschlueter, agreed!15:42
jschlueterdmsimard: did you get patch up yet or should I get it for python-networking-cisco?15:43
jschluetersadasu: you can propose the tag as well if you have things setup ...15:43
dmsimardjschlueter: I figured sadasu would be taking care of it15:43
jschluetersadasu: is there a stable/newton branch planned?15:43
jschlueterdid we miss stable/mitaka as well?15:44
sadasujschlueter: yes, there is a plan for it. But, we haven't done it yet15:44
jschluetersadasu: what should CI be testing against for RDO Newton Trunk and RDO Mitaka Trunk15:44
jschluetersadasu: ack will it be based off of master or from the 4.0.0 tag?15:45
*** seanatcisco has quit IRC15:45
* jschlueter wants to make sure rdo.yaml in rdoinfo is sane for python-networking-cisco15:45
*** artom has quit IRC15:45
*** seanatcisco has joined #rdo15:46
*** artom_ has joined #rdo15:46
dmsimardyeah, interestingly enough, only mitaka is pinned right now15:46
dmsimardthere is nothing for liberty, newton and ocata15:46
sadasujschlueter: from 4.0.0 tag and not master15:46
jschlueterdmsimard: /me didn't peek at rdoinfo15:46
jschluetersadasu: ack15:46
jschlueterdmsimard: newton then needs 4.0.0 for source-branch ... liberty has a stable branch15:47
* jschlueter will get that next then15:47
sadasujschlueter: that is correct. Liberty has stable branch, mitaka is already pinned, newton to 4.0.0, Ocata I don't know15:48
dmsimardthe challenge here is *maintenance*, because since networking-cisco isn't tracked in upstream upper-constraints (for example) we won't automatically be updating tags15:49
dmsimardsadasu: we have and listed as maintainers, is that ok ?15:51
*** sdake has quit IRC15:52
*** karimb_ has joined #rdo15:53
*** rbowen has quit IRC15:54
sadasudmsimrad: that is fine for the maintainers15:54
*** athomas has quit IRC15:55
*** numans has quit IRC15:55
*** karimb has quit IRC15:56
*** rcernin has quit IRC15:56
*** vimal has quit IRC15:56
*** dprince has quit IRC15:58
*** pcaruana has quit IRC15:58
*** anilvenkata has quit IRC15:59
*** tesseract- has quit IRC15:59
rdogerritJon Schlueter created rdoinfo: Pin python-networking-cisco for newton to 4.0.0
amoralejwe have a new library requirement requirement for rally15:59
jschlueterdmsimard: ^^16:00
amoraleji'll prepare the package review16:00
*** athomas has joined #rdo16:01
jpenaamoralej: chandankumar said he would do it16:02
amoralejthanks jpena, i looked for os-fault in #rdo and he used os_fault  :D16:03
rdogerritJon Schlueter created openstack/networking-cisco-distgit: Update to 4.0.0 (dbb4b8d30accf4747ecfee0d989b57cb672ce7ee)
jschluetersadasu: ^^16:03
sadasuyes, I just looked at it. I have made changes to python-networking-cisco.spec but have questions about the changelog16:05
jschlueteramoralej: several more added to stable_build topic just fyi16:05
jschluetersadasu: sure what's the question16:05
jschlueteramoralej: just so we can keep tabs on the related newton stable build related things I've been pushing this week16:06
sadasuhow is this id "1:3.2.1-0.1.509eb2cgit" generated? I am manually editing this that OK?16:06
jschluetersadasu: on the commit log ... did you look at
jschluetersadasu: /me personally used rdopkg for helping generate reasonable changelogs for doing this work but that is the combination of the <epoch>:<version>-<revision> from the spec file16:08
sadasujschlueter: made the changes for me16:08
jschlueterif we want to tweak things we can do that16:08
jschluetersadasu: it was on my list and I've done just a few of these in the past week for other newton packages as well16:09
dmsimardjpena: Would cause nvr to go back in time ?16:09
sadasujschlueter: thanks! Looks good to me.16:09
dmsimardjpena: and thus be ineffective in, i.e, CDN repos because they carry all packages16:09
dmsimardjschlueter: ^16:10
jschluetersadasu: hmm one bit I think we want to tweak is to drop %{?alphatag} as it's from a clean tag upstream16:10
*** mlammon|afk has quit IRC16:10
*** mlammon has joined #rdo16:10
jschlueterjpena, dmsimard: that's part of the problem we have .. but for this one it's 2 patches right now ahead of the tag16:11
jpenadmsimard: I'm afraid so, but let me check16:11
*** jaosorior_mtg is now known as jaosorior16:11
jschlueterand from one of them I think they should consider a point release of 4.0.1 here soon ... fix typeo in rootwrap.conf16:11
jschluetersadasu: can they get the stable/newton branch created sooner rather than later?16:12
jschlueterjpena: all trivial changes since tag16:12
sadasujschlueter: sure! Let me get things started internally on that front16:12
jschlueterbut not wanting to let it creep farther in RDO Newton trunk16:12
jschluetersadasu: if we have stable/newton that would help out16:13
sadasubut, we can still proceed with moving to the 4.0.0 page on RDO or do you want that stable/newton to be created before you proceed?16:13
*** tiswanso has joined #rdo16:14
amoralejjschlueter, do we have any feedback if networking-odl version in stable/mitaka works for newton?16:14
jpenadmsimard: confirmed, the nvr will be lower after pinning16:15
*** akshai has quit IRC16:15
*** Son_Goku has quit IRC16:15
rdogerritMerged rdoinfo: osprofiler: pin 1.4.0 for Liberty and Mitaka too
sadasujschlueter: we don't plan to cut a newton/stable branch on networking-cisco16:16
sadasuso, we should just proceed with the 4.0.0 tag for networking-cisco newton on RDO16:17
*** percevalbot has quit IRC16:17
jschluetersadasu: ack16:18
jschlueteramoralej: ack that was the recommended version from mike when I was talking with him about the issue for python-networking-odl16:19
amoralejok, version in cbs is based on a later git snapshot16:19
*** saneax-_-|AFK is now known as saneax16:20
rdogerritJon Schlueter proposed openstack/networking-cisco-distgit: Update to 4.0.0 (dbb4b8d30accf4747ecfee0d989b57cb672ce7ee)
*** eaguilar has quit IRC16:21
*** dprince has joined #rdo16:21
*** percevalbot has joined #rdo16:21
jschlueteramoralej: yep...  I know and ick but that's what was being asked ... not sure if we can do much about that at this point though16:21
rdogerritrdo-trunk created openstack/rally-distgit: openstack-rally: failed to build 7d76719
*** tshefi has joined #rdo16:22
amoralejwe could build 2.0.1-0.2 based on current commit in stable/mitaka for stable repo16:23
jschlueteramoralej: yea  probably would be good to do as well!16:24
*** smeyer has quit IRC16:26
rdogerritJon Schlueter proposed rdoinfo: Pin python-networking-cisco for newton to 4.0.0
*** abehl has joined #rdo16:27
*** bnemec is now known as beekneemech16:28
*** panda is now known as panda|bbl16:29
*** jcoufal has quit IRC16:29
*** jcoufal has joined #rdo16:30
rdogerritJon Schlueter proposed rdoinfo: Add stable branch for python-networking-odl for newton
rdogerritJon Schlueter proposed rdoinfo: Pin python-networking-cisco for newton to 4.0.0
rdogerritMerged rdo-infra/weirdo: Setup tuned profile by default in weirdo for optimal performance
dmsimardapevec, amoralej, jpena: ^ merged, will be effective for new runs on cico.. let's see if it helps (it definitely can't hurt)16:33
amoralejcool dmsimard, let's see how it goes...16:37
*** dustins has quit IRC16:37
*** apevec has quit IRC16:38
dmsimardapevec: the -pending .repo is not available anywhere right? no rpm or .repo ?16:38
dmsimardoh, he's not there16:39
dmsimardnumber80: ^ ?16:39
rdogerritMerged rdoinfo: Pin python-networking-cisco for newton to 4.0.0
*** READ10 has quit IRC16:41
*** zoli is now known as zoli|gone16:42
*** chandankumar has joined #rdo16:42
number80I have a .repo for weirdo generic job though16:42
amoralej ?16:42
dmsimardamoralej: that's something I created momentarily yesterday16:42
dmsimardjpena: should we add a -pending.repo on trunk.rdo ? I just don't want to carry that in weirdo.. it's something that oooq will have to use as well16:42
dmsimardor maybe add it to -deps but disabled16:43
dmsimardyeah, I like that idea16:43
*** jaosorior has quit IRC16:43
number80well, disabled by default16:45
number80or else, we'll need to create another "testing" repo16:45
dmsimardnumber80: yeah, that's what I said, disabled16:46
rdogerritDavid Moreau Simard created rdo-infra/puppet-dlrn: Add -pending repository to delorean-deps.repo
dmsimardamoralej, jpena ^16:47
*** apevec has joined #rdo16:50
*** mcornea has quit IRC16:51
*** ayoung has joined #rdo16:52
rdogerritJon Schlueter proposed openstack/networking-cisco-distgit: Update to 4.0.0 (dbb4b8d30accf4747ecfee0d989b57cb672ce7ee)
dmsimardjpena: thanks, I'll add it manually for the time being so I can get the weirdo implementation done16:53
jschlueteramoralej, apevec: thanks for reminder to cleanup where it's coming from ...16:53
jpenadmsimard, I can add it to the dlrn instance right now16:53
amoralejwait for a while jpena16:54
*** paramite has quit IRC16:54
amoralejit'll break
jpenaamoralej: ack16:55
amoraleji'll send an update to repoquery command16:55
dmsimardamoralej: oh, good catch16:55
dmsimardamoralej: let me fix it16:55
dmsimardamoralej: I'll bundle it into the -pending patch16:55
dmsimardso we don't run twice the gate jobs16:55
dmsimardamoralej: it's set to ignore_errors and it won't fail as-is anyway I think16:56
amoralejyes, but the info from the repoquery may be unaccurate16:56
rdogerritMerged openstack/oslo-rootwrap-distgit: Update to 5.1.1
rdogerritMerged rdo-infra/puppet-dlrn: Add -pending repository to delorean-deps.repo
dmsimardamoralej: so the regex becomes "delorean*testing" ?16:58
rdogerrithguemar created openstack/horizon-distgit: Update to 10.0.1
amoralejwe'll need to add two --enablerepo16:58
amoralejone for "delorean" and one with "delorean*testing"16:59
amoralejlet me test it16:59
rdogerrithguemar created openstack/neutron-distgit: Update to 9.1.0
*** emalikov has joined #rdo17:00
*** satya4ever has quit IRC17:00
dalvarezsihrachys, instead of submitting a new patch i suggested a solution as a comment. please check it out when you can and if so, i will submit a new patch set17:00
*** kgiusti has left #rdo17:00
dalvarezsthanks!! :)17:00
dalvarezsnow im off, 10 hours in a row are just too much for me :)17:01
rdogerritMerged openstack/ceilometerclient-distgit: Update to 2.6.2
*** tvignaud has quit IRC17:03
*** danpawlik has quit IRC17:03
ihrachysdalvarezs: replied17:03
*** flepied has joined #rdo17:04
*** tvignaud has joined #rdo17:04
dmsimardnumber80: hmm, setting up the -pending repo is awkward.. it has to be used with -testing and can be tested with either trunk or stable release I guess ?17:05
dmsimardor just trunk ?17:05
dmsimardI would guess just trunk because afterwards we have -testing support with stable17:05
dmsimardapevec: ^17:06
*** egallen_ has joined #rdo17:06
*** egallen has quit IRC17:06
*** egallen_ is now known as egallen17:06
*** hewbrocca is now known as hewbrocca_afk17:07
amoralejdmsimard, i think the combinations could be "trunk + -testing + -pending" or "stable + pending" , both may make sense17:08
*** eaguilar has joined #rdo17:08
amoralejbut if the question is how to gate changes in deps, i'm not sure what whould be better, to be honest17:09
sadasuthanks so much for your help!17:09
amoraleji think trunk17:09
jschluetersadasu: not a problem and glad to have more people interested17:09
*** Son_Goku has joined #rdo17:10
*** Son_Goku has quit IRC17:10
number80dmsimard: with common-testing17:11
number80-pending only exists for common17:11
*** fragatina has quit IRC17:11
rdogerritJon Schlueter proposed rdoinfo: Add stable branch for python-networking-odl for newton
*** jpich has quit IRC17:13
*** dustins has joined #rdo17:15
amoralejdmsimard, if the change in weirdo improves performance, we should propose it to oooq also, we are having performance issues there as well17:16
amoraleji mean the tuned profile17:16
*** jprovazn has joined #rdo17:17
dmsimardamoralej: yup, definitely17:21
dmsimardnumber80: common-testing? is that ?17:22
*** ohamada has quit IRC17:23
*** saneax is now known as saneax-_-|AFK17:23
*** rbowen has joined #rdo17:23
*** ChanServ sets mode: +o rbowen17:23
*** florianf has quit IRC17:23
rdogerritDaniel Alvarez created openstack/neutron-distgit: Now the cleanup function iterates overa all keepalived-state-changed processes killing them and all their children
number80dmsimard: common-testing is mashed with buildlogs yes17:25
rdogerritJon Schlueter created puppet/puppet-pacemaker-distgit: bump version to 0.3.0 to match metadata.json upstream
dalvarezsihrachys, thanks i have uploaded a new patchset17:25
*** markd_ has quit IRC17:25
dmsimardnumber80: ok, good .. last question, for the testing pipeline, we want to test -pending with trunk only, right? Since they'll eventually be caught in the stable pipelines that include -testing17:26
*** imcleod has quit IRC17:26
dmsimards/the testing pipeline/the pipeline/17:26
dmsimardor actually pending+testing+trunk17:26
ihrachysdalvarezs: you were going to go home ;)17:26
* dmsimard confused with all the repos, need more caffeine17:27
dalvarezsihrachys, you're right :( hehe17:27
ihrachysdalvarezs: enjoy!17:27
dalvarezsihrachys, thanks! i am double checking because i had to change the commit message and i bet we lost the comments :(17:29
*** florianf has joined #rdo17:29
*** aludwar has quit IRC17:30
*** aludwar has joined #rdo17:31
*** toanju has quit IRC17:31
*** lucasagomes is now known as lucas-afk17:32
number80dmsimard: no, all of them, -pending is staging for -testing17:36
number80if he fails with a stable release, it won't in common-testing but ocata-testing for instance17:36
dmsimardnumber80: we have limited resources and can't test everything.... do you want to test with trunk or with stable ?17:37
number80dmsimard: it can be on-demand jobs17:37
rdogerritJon Schlueter proposed puppet/puppet-pacemaker-distgit: bump version to 0.3.0 to match metadata.json upstream
number80if we want to run periodically => trunk17:38
number80but i need the possibility to run it on stable branches from time to time17:38
dmsimardnumber80: so in summary so we're on the same wavelength, this is what you want ?
*** d0ugal has quit IRC17:39
number80yes,should be good17:39
number80trunk is daredevil17:39
number80we need to test things that could break17:39
*** rasca has quit IRC17:40
number80stable are cautious kittens that needs to be sure before jumping in water17:40
dmsimardthis is complicated ...17:42
*** imcleod has joined #rdo17:42
number80well, it'd be simpler if -testing were used as a testing playground and not source of truth :(17:42
*** jkilpatr has quit IRC17:42
dmsimard-testing is already used in trunk promotion jobs17:42
dmsimardso we don't need to touch that17:42
dmsimardbut we need to have a version that's able to hook in -pending17:43
number80yep, we don't need pending there17:43
*** iniazi has joined #rdo17:43
dmsimardand -pending is provided by delorean-deps, no less17:43
*** jkilpatr has joined #rdo17:43
*** ihrachys has quit IRC17:43
number80dmsimard: compromise would be a weekly job, that can be launched on-demand17:43
dmsimardnumber80: yeah that's fine, but handling the possible matrix in weirdo is getting complex17:44
dmsimardand I don't want weirdo to become complex :D17:44
dmsimardtrying to think of something17:44
*** creepy_owlet is now known as dtantsur|afk17:44
dmsimardamoralej: did you end up figuring out repoquery string ?17:45
*** dustins has quit IRC17:46
amoralejsorry dmsimard, something crossed in my way17:46
amoralejjust a while17:47
number80pretty much understandable17:47
*** markd_ has joined #rdo17:47
amoralejdmsimard, look at this, response time difference between RDO-CI and "other" ci17:47
*** emalikov has quit IRC17:51
*** shardy has quit IRC17:51
dmsimardamoralej: not surprising17:52
dmsimardamoralej: look at this:
*** gszasz has quit IRC17:54
dmsimardamoralej: everything just passed, before that everything was timing out >1h3017:54
amoralejdmsimard --enablerepo="delorean" --enablerepo="delorean*testing" works like a charm17:54
amoralejwe have two repoquery commands17:55
dmsimardamoralej: ah, maybe it's a coincidence, because those jobs didn't run with tuned enabled17:55
amoraleji sent a job17:55
amoralejlet me check17:55
amoralejstill building the image17:56
dmsimardamoralej: there's another thing that'd probably improve oooq perf17:56
dmsimardamoralej: but I haven't had the chance to merge it yet17:56
*** dustins has joined #rdo17:57
dmsimardtl;dr, oooq thought it had nested virt but not really17:58
dmsimardbecause the module needs to be reloaded17:58
*** akshai has joined #rdo18:00
amoraleji see18:00
dmsimardamoralej: with tuned, proper nested virt and ansible 2.2 I think performance would be improved and jobs would be more reliable :D18:00
amoralejthat'd explain slowness when creating instances in overcloud, as in tempest or ping test runs18:00
dmsimardyeah, slowness or outright failures18:02
amoralejbut it's being veeeery slow even in pure undercloud actions !18:02
amoraleji'm confident with the cpu tunning also18:03
amoralejis there a way to run tripleo jobs with a specific hash repo where we already have the images created?18:03
amoralejwaiting for jobs is killing me... :(18:04
rdogerritDaniel Alvarez proposed openstack/neutron-distgit: Killing all child processes of keepalived-state-change
dmsimardamoralej: the images are discarded on every new build18:07
amoralejwell we may re-run it against the current-passed-ci18:07
dmsimardwell, yeah18:08
amoralejjust to check performance of the job18:08
dmsimardamoralej: I think I fixed the failures in .. let's see what happens18:08
*** jlibosva has quit IRC18:09
dmsimardamoralej: wow completed in 19 minutes18:09
dmsimardran stable/mitaka18:09
*** links has joined #rdo18:09
dmsimardon dusty (intel?)18:10
*** links has quit IRC18:10
amoralejyes, but still pretty fast18:10
amoralej is on crusty18:11
dmsimard 39 minutes on AMD18:11
rdogerritDaniel Alvarez proposed openstack/neutron-distgit: Killing all child processes of keepalived-state-change
amoralejsounds good...18:11
dmsimardok I'll submit a tuned implementation in quickstart18:12
jpenahave a nice weekend everyone!18:15
amoralejyou too jpena!18:15
*** jpena is now known as jpena|off18:16
*** karimb_ has quit IRC18:16
*** karimb has joined #rdo18:17
dmsimardjpena: o/18:18
dmsimardamoralej, apevec: interesting coincidence how we chatted about part of this issue today
*** sshnaidm has quit IRC18:24
*** sshnaidm has joined #rdo18:27
*** rhefner has quit IRC18:27
*** karimb has quit IRC18:27
*** karimb has joined #rdo18:28
*** pblaho has quit IRC18:33
*** sambetts is now known as sambetts|afk18:35
*** anilvenkata has joined #rdo18:35
*** milan has quit IRC18:38
*** saneax-_-|AFK has quit IRC18:38
*** gszasz has joined #rdo18:43
*** dpeacock has quit IRC18:44
*** jerrygb has quit IRC18:45
*** jerrygb has joined #rdo18:45
*** akshai has quit IRC18:46
*** karimb has quit IRC18:46
rdogerritMerged openstack/tempest-distgit: Update to post 13.0.0 (296cca79678463dd90ca2694245bb7b89f8405d8)
*** jkilpatr has quit IRC18:50
rdogerritrdo-trunk created openstack/vmware-nsx-distgit: python-networking-vmware-nsx: failed to build ed75979
*** imcleod has quit IRC19:00
rdogerritAlfredo Moralejo created openstack/neutron-lbaas-distgit: Update to 9.1.0
*** jkilpatr has joined #rdo19:03
*** fzdarsky__ is now known as fzdarsky|afk19:08
*** sadasu has quit IRC19:08
*** paragan has quit IRC19:11
*** amoralej is now known as amoralej|off19:11
*** jerrygb has quit IRC19:14
rdogerritMerged openstack/horizon-distgit: Update to 10.0.1
*** imcleod has joined #rdo19:16
*** saneax-_-|AFK has joined #rdo19:16
*** jerrygb has joined #rdo19:16
rdogerritJon Schlueter created openstack/networking-bigswitch-distgit: Update to 7.0.0
apevecnumber80, I know you nacked it before, but we should enable at least cbs-check job on older *-rdo branches19:27
apevecnewton 7.2.0 on liberty-rdo failed due to missing patch  from rpm-liberty (I'm cherry picking it now)19:27
*** jcoufal has quit IRC19:28
*** akshai has joined #rdo19:33
*** akshai has quit IRC19:35
apevechrmpf liberty-rdo and rpm-liberty are not in sync19:35
*** fragatina has joined #rdo19:44
*** eaguilar has quit IRC19:44
*** eaguilar has joined #rdo19:52
*** chandankumar has quit IRC19:54
*** jerrygb_ has joined #rdo20:01
*** fragatina has quit IRC20:03
*** jerrygb has quit IRC20:05
*** jerrygb_ has quit IRC20:06
*** tiswanso has quit IRC20:06
*** coolsvap has quit IRC20:07
*** morazi has quit IRC20:09
*** alexismonville has quit IRC20:11
*** dprince has quit IRC20:12
*** rwmjones has quit IRC20:12
*** rwmjones has joined #rdo20:14
*** abregman has joined #rdo20:14
*** jayg is now known as jayg|g0n320:14
*** karimb has joined #rdo20:14
*** florianf has quit IRC20:15
*** jprovazn has quit IRC20:15
rdogerritMerged openstack/networking-cisco-distgit: Update to 4.0.0 (dbb4b8d30accf4747ecfee0d989b57cb672ce7ee)
*** sileht has quit IRC20:21
rdogerritMerged openstack/neutron-lbaas-distgit: Update to 9.1.0
*** jkilpatr has quit IRC20:25
*** athomas has quit IRC20:25
*** derekh has joined #rdo20:27
*** sileht has joined #rdo20:28
dmsimardweshay: do you know if the failures from are documented ?20:29
*** panda|bbl is now known as panda20:31
apeveccheck etherpad20:31
apevecI guess Fred's analyse script didn't run yet since pipeline has not finished yet20:32
apevecdmsimard, so looks like weirdo are faster now?20:33
apeveconly one barely missed 90min timeout20:33
apevecoh 806 did finish20:33
apevecboth oooq failed20:34
dmsimardapevec: yeah, faster .. but there's still something weird, especially with p-o-i scenario00220:35
apevecwas there something new added recently in that scenario?20:36
dmsimardI don't know, it's probably not that20:36
dmsimardFor example, the stable/mitaka packstack scenarios can finish in under 20 minutes20:36
dmsimardbut the trunk ones are sometimes taking an hour20:37
dmsimardI have some ideas but hard to test. tuned was one, nested virt was another.. those are fixed now and we're still seeing slowness20:38
apevecso in 806 oooq min failed in overcloud validation
dmsimardI submitted in upstream p-o-i but I don't think that's it either20:38
apeveci.e. pingtest failed20:39
dmsimardpuppet is just darn slow, I can't explain it20:39
dmsimardlike, for example, puppet runs twice in p-o-i, to check for idempotency20:39
dmsimardsecond catalog compilation took ~20 seconds and actual second run took >400 seconds20:40
*** karimb has quit IRC20:40
apevecwhat was it doing?20:41
*** tiswanso has joined #rdo20:45
*** afazekas_ has joined #rdo20:45
*** jkilpatr has joined #rdo20:46
dmsimardI don't think it's puppet4 because if it was puppet4 we would not see the issue in trunk mitaka/newton20:48
dmsimardThe AMD nodes do not look not particularly loaded, stressed or bottlenecked afaict, it's just .... slow20:48
*** afazekas has quit IRC20:48
dmsimardcpu usage, i/o, network and ram are fine20:48
dmsimardthat's why I mentioned the httpd slowness this morning20:48
dmsimardif httpd is slow, everything puppet will be doing will be slow because of all the required API calls20:49
*** tiswanso has quit IRC20:50
*** karimb has joined #rdo20:50
*** dbecker has joined #rdo20:54
*** dbecker has quit IRC20:54
*** pradk has quit IRC20:55
rdogerritrdo-trunk created openstack/tripleoclient-distgit: python-tripleoclient: failed to build 207bd4a
*** steveg_afk has quit IRC21:03
*** flepied has quit IRC21:03
rdogerritrdo-trunk created openstack/vmware-nsx-distgit: python-networking-vmware-nsx: failed to build ed75979
EmilienMapevec: I need your expertise21:07
EmilienMI think (but not 100% sure) that Redis 3.2.4 broke TripleO when deploying without HA21:08
EmilienMpanda: is it with HA too? i don't recall21:08
EmilienMapevec: I saw some tripleo jobs working well with redis 3.2.321:08
EmilienMapevec: how can we look packaging changes ?21:08
pandaEmilienM: in HA the resource agent takes care of creating the dir /var/run/redis every time redis is started, so the daemon is still broken, but pacemaker works around it21:10
*** abregman has quit IRC21:13
*** absubram has quit IRC21:13
*** steveg_afk has joined #rdo21:16
EmilienMdmsimard: thx man21:17
EmilienMdmsimard: do you know where can we see the .spec file?21:18
EmilienMI'm interested if we changed something in systemd script, etc21:18
dmsimardEmilienM: the src rpm are in the builds, not sure about spec21:18
*** eaguilar has quit IRC21:19
dmsimardand for 3.2.321:19
EmilienMnumber80: ^ fyi21:19
EmilienMnumber80: you tagged redis end of september and it doesn't work for us21:20
dmsimardit hasn't worked in two months ?21:21
*** lv_ has joined #rdo21:23
dmsimardEmilienM: that shows 3.2.3, not 3.2.4 .. but
*** unclemarc has quit IRC21:24
*** rhallisey has quit IRC21:25
*** rlandy has quit IRC21:32
*** anilvenkata has quit IRC21:32
*** ihrachys has joined #rdo21:34
*** maeca1 has quit IRC21:34
*** paramite has joined #rdo21:37
*** abehl has quit IRC21:40
*** sdake has joined #rdo21:42
*** jeckersb is now known as jeckersb_gone21:44
*** dustins has quit IRC21:51
*** jerrygb has joined #rdo21:51
*** morazi has joined #rdo21:52
*** dtrainor has quit IRC21:54
*** jerrygb has quit IRC21:56
*** paramite has quit IRC21:56
*** jmelvin has quit IRC21:57
*** derekh has quit IRC21:58
*** aludwar has quit IRC21:58
weshaydmsimard, apevec yes.. we've escalated the the ping test issues22:03
*** derekh has joined #rdo22:09
*** ihrachys has quit IRC22:11
*** aludwar has joined #rdo22:13
*** dneary has quit IRC22:14
*** fultonj has quit IRC22:19
*** artom_ has quit IRC22:19
*** gfidente has quit IRC22:19
*** noslzzp has quit IRC22:24
*** noslzzp has joined #rdo22:24
*** flepied has joined #rdo22:27
*** tshefi has quit IRC22:27
apevecEmilienM, dmsimard - 3.2.3..3.2.4 is just upstream rebase, no real changes in spec afaict22:34
apevecit was security update, but it must be some additional upstream change22:34
apevecEmilienM, what's the failure?22:34
apevecspec diff:
amoralej|offapevec, /var/run/redis was removed from the package bug 1374728 in redis "/var/run/redis and /usr/lib/tmpfiles/redis.conf not used" [Unspecified,New] - Assigned to fpercoco22:36
amoralej|offso when it was created by resource agent had the wrong selinux label22:36
*** artom has joined #rdo22:36
apevecamoralej|off, you're off or not :)22:36
amoralej|offkind of :)22:37
apevecbut having it in the package is wrong, it's tmpfs22:37
apevecon reboot it would be gone22:37
apevecso it _seemed_ like it was working but not really22:37
amoralej|offno, it was done through tmpfiles22:37
amoralej|offlet me look for it22:37
*** sdake has quit IRC22:37
*** sdake has joined #rdo22:40
pandaamoralej|off: oh22:43
apevecfound it
apevecso logic is that /run is not used with systemd22:44
amoralej|offyes apevec22:44
amoralej|offthat was the logic22:44
amoralej|offi had to fix packstack to disable unixsocket22:45
apevecpanda, so what doesn't work in non-HA ooo ?22:45
pandawe're doing the same thing in tripleo in the end22:45
*** amoralej|off is now known as amoralej22:46
amoralejbut in pcmk configuration unixsocket is required22:46
pandaapevec: unix socket is configured in tripleo, but nothing creates it, so redis does not start22:46
pandaamoralej: for what ?22:47
amoralejresource agent uses unixsocket to check the status22:47
apevecto communicate with the daemon22:47
pandamwhahaha: ^22:47
pandamwhahaha: we need the socket22:47
apevecsystemd is not forking it so it talks directly to the child22:47
amoralejonly in HA config22:47
apevecpanda, for HA22:47
amoralejnot in non-ha22:48
mwhahahawe need to update the systemd unit file to create that directory then22:48
pandafor HA we already have a fix22:48
apevecmwhahaha, which dir?22:48
mwhahahawe need /var/run/redis22:49
apevecsystemd does not use /run22:49
amoralejbut puppet-redis enables unixsocket by default22:49
apevecpacemaker should create it22:49
mwhahahasystemd can create a /var/run/redis as part of the unit22:49
apevecyeah, so just change parameter when calling puppet-redis22:49
mwhahahabut then we need a selinux acceptable path22:49
apevecmwhahaha, but why? systemd is not using it22:49
mwhahahaand /var/run/redis is already configured22:50
pandaso, we have two cases:22:50
pandain HA, the socket is needed and the /var/run/redis dir is created by the resource agent22:50
pandain nonHA, the socket is not needed and may be disabled22:51
apevecand for nonHA unixsocket => false22:51
mwhahahabut i should be able to configure the socket (as an end user)22:51
apevecsure but then BYOdir :)22:51
mwhahahaand manage the selinux around that dir?22:51
amoralejor make puppet-redis to do it via /usr/lib/tmpfiles.d/redis.conf file22:52
mwhahahathis is a packaging thing not something to be done in puppet later22:52
apevecyou can create /var/run/redis which is covered by selinux22:52
mwhahahajust because systemd has taken over /run you're kicking stuff out of it?22:52
apevecwell, package and distro maintainers decided to go systemd22:52
mwhahahawhy can't the redis systemd unit file (from packaging) have a patch to create this dir?22:52
*** jerrygb has joined #rdo22:52
amoralejpackaging usually takes care of making it work with default configuration22:53
*** elmiko is now known as _elmiko22:53
amoralejand in this case is systemd without unixsocket22:53
mwhahahaso this existed in packaging until people who didn't understand the usecase removed it22:53
pandamwhahaha: because it doesn't know what's the configured socket dir22:53
mwhahahaall this argument over asking for a directory that used to exist to be put back22:53
apevecmwhahaha, from packaging PoV this was a bug that it kept dir22:53
pandadefault redis.conf uses /tmp/redis.sock22:54
apevecyou relied on a bug :)22:54
mwhahahayea cause it was used22:54
amoralejdefault redis.conf doesn't enable unixsocket iirc22:54
apevecit doesn't22:54
apevecit was a patch in rpm22:54
pandauh, maybe default example then :) I've seen thi /tmp/redis.sock somewhere22:55
apevecanyway, we are not redis maintainers,
apevecoh number80 is co-maint :)22:56
pandaare we able to differentiate when we are building HA and wehn we're not in puppet-tripleo ? Or do e have to modify the templates ?22:56
apevecand flaper87 ^ !22:56
mwhahahaor we just configure it to /tmp22:56
mwhahahabecause that's a great place for stuff22:57
amoralejwill fail because of selinux22:57
pandayeah, probably ...22:57
apevecalso CVE opportunity22:57
apevecpanda, alright, so we can blame Flavio, he's maintainer :)22:58
pandait's 1:30 hour of train from here.22:58
*** jerrygb has quit IRC22:58
pandaI'll go yell at him22:58
apevecfix it Italian-way? :)22:58
apevecso I thing we could revert back tmpfiles into rpm22:59
apevecbut keep default config w/o pidfile22:59
apevecnumber80, ^ wdyt ?22:59
apevecI'll add a comment in bug 1374728 in redis "/var/run/redis and /usr/lib/tmpfiles/redis.conf not used" [Unspecified,New] - Assigned to fpercoco23:00
*** cliles has quit IRC23:01
pandaSo this is the root cause of almost everything that broke CI in the last week23:02
apevec bug 1374728 in redis "/var/run/redis and /usr/lib/tmpfiles/redis.conf not used" [Unspecified,New] - Assigned to fpercoco23:02
apevecmwhahaha, seen your comment: not unit file but we just need tmpfiles/redis.conf back23:03
apevecamoralej, rgiht? ^23:03
amoralejyeap, i can test it if you want23:03
mwhahahawell, i was saying a fix could be instead of managing the state folder from packaging, do it in the redis.service via an ExecStartPre23:03
apevecnope, systemd is not using it so it doesn't make sense there23:04
pandaapevec: yeah, unit file is too static in this case, and it's tmpfiles job anyway ..23:04
apevecamoralej, I think you should leave for w/e :)23:04
mwhahahai'll defer to you guys on that, was just looking for a minor solution23:04
amoralejlet me send the change for magnum23:05
apevecah redis.tmpfiles is still in distgit, so revert will be just 2-3 lines in spec :)23:06
pandathis render completely useless my work for the past week. yay.23:10
apevecit's for greater good!23:11
apevecsorry, I didn't look at this redis issue in detail, only saw it mentioned in the passing23:12
apevecI guess nobody tried to make a compromise23:12
apevecI'm good at doing rotten ones23:13
pandawell, when I found the selinux problem, I thought I was going deep enough.23:15
rdogerritAlfredo Moralejo created openstack/magnum-distgit: Add python-marathon as req and BR
pandabut it's always the upstream source, or the package...23:16
apevec bug 1374728 in redis "/var/run/redis and /usr/lib/tmpfiles/redis.conf not used" [Unspecified,Post] - Assigned to apevec23:17
apevecpanda, ^ proposed patch23:17
apevecI'll make a scratch build in CBS23:17
pandaapevec: you can ? you don't need to wait for all RHEL process ?23:19
*** saneax-_-|AFK is now known as saneax23:19
pandaapevec: because the source was not modified ?23:19
amoralejredis is not in rhel panda23:19
apevecredis is not in base EL23:19
amoralejis in openstack repos23:19
pandais there something I understand in all this stuff ? :(23:19
apevecI posted earlier: best (only?) public reference for what is "base EL7" is
apevecso if not there, it must be in our deps repo or openstack project23:20
pandalas update: 2014  ... but ok, I get it.23:22
apevecyeah, it's good approximation23:23
apevecbut indeed there were few packages added to base EL7 in minor .Y updates23:23
apevecotoh nothing major23:23
pandaThe funny thing now will be to track down all the work on the bugs that were caused by this and shot it down.23:24
pandanext time I'll go for the package first :)23:25
*** imcleod has quit IRC23:25
apevecdepends :)23:25
apevecanyway, here's scratch build
apevecok, might not be working23:26
pandaapevec: is this only for local testing ? not exaclty in the mood to start a local test at midnight on frday night ..23:27
amoralejyeap, no tmpfiles file23:27
apeveconly local but hold on23:27
amoralejit's not under etc23:27
apevecthere is /usr/lib/tmpfiles.d/redis.conf but that's used only on boot ?23:28
amoralejsorry, you are right23:28
amoralejusr is right23:28
amoraleji'm checking if we need to trigger systemd-tmpfiles --create23:29
apevecyeah, I got dir after reboot23:29
apevecok, so it won't work OOB...23:29
apevecpanda, re. friday night - I suggest to start drinking vs testing :)23:30
pandaapevec: I could drink while testing23:31
pandaI suppose tmpfiles are created before selinux right ?23:31
pandaso that context can be restored correctly23:32
amoralejsystemd-tmpfiles creates dir with the right selinux label23:32
*** scorcoran has joined #rdo23:32
apevecbut  that's not automatically executed23:33
apeveconly on boot23:33
*** apevec has left #rdo23:33
*** apevec has joined #rdo23:33
*** scorcoran is now known as scorcoran_biab23:33
apevecso we still don't have a complete solution23:33
apevecmaybe trigger it in %post  ?23:34
apevecamoralej, what's the full command line?23:34
amoralejnot, we may need to add a systemd-tmpfiles --create in %post23:34
apevecalright, updating the patch23:34
amoralejbut, why it worked before?23:34
apevecmust be systemd triggering it somehow23:35
*** gszasz has quit IRC23:35
*** akshai has joined #rdo23:35
amoralejlet me try with old version23:36
*** akshai has quit IRC23:36
amoralejin old version, directory gets automatically created23:40
apevecamoralej, ah yes, 3.2.3-1 includes it in rpm %files23:40
apevecok, adding it23:40
amoraleji used 3.0.623:41
*** akshai has joined #rdo23:44
*** Son_Goku has joined #rdo23:46
*** akshai has quit IRC23:48
amoralejapevec, i'm leaving23:53
amoralejare you preparing a new scratch build?23:53
apevecyep, running it now23:53
apevecbut go to w/e :)23:53
apevecI'm a bad example...23:53
apevecwe haz /var/run/redis/ back23:54
amoralejthat works fine23:55
apevecok, I'll push cbs build to common-pending23:55
pandaapevec: thanks!23:55
amoralejhave a nice weekend !23:55
amoralejanyway i think puppet-redis shouldn't enable unixsocket by default and resource-agent should restorecon a dir if it creates it :)23:56
apevecamoralej, yes, but this can be seen as a regression, it would break existing installations right?23:57
amoralejyes, that's right, backwards compatibility is a good reason23:57
apevecso for backward compat, rpm should keep tmpfiles23:57
amoralejyou convinced me23:58
pandaamoralej: resource-agent fix is in place23:58
amoralejyes, i saw it was merged and built already23:58
pandafor the puppet-redis we'll see ..23:58
apevecnow just need to convince pkg slackers^W maintainers number80 and flaper87  :)23:59
pandathe troublemakers23:59
amoralejenjoy your weekend!23:59
*** amoralej is now known as amoralej|off23:59

Generated by 2.14.0 by Marius Gedminas - find it at!