Friday, 2019-04-26

*** djhankb has quit IRC00:14
*** jsquare has quit IRC00:14
*** jsquare has joined #openstack-ansible00:15
*** djhankb has joined #openstack-ansible00:17
*** markvoelker has quit IRC00:19
*** cshen has joined #openstack-ansible00:23
*** cshen has quit IRC00:29
*** hwoarang has quit IRC00:37
*** hwoarang has joined #openstack-ansible00:41
*** marst has quit IRC01:24
*** dave-mccowan has quit IRC01:28
*** kplant has quit IRC01:32
*** hwoarang has quit IRC01:59
*** hwoarang has joined #openstack-ansible02:00
*** cshen has joined #openstack-ansible02:24
*** cshen has quit IRC02:30
*** grabes has quit IRC02:32
*** samueldmq has quit IRC03:04
*** nicolasbock has quit IRC03:21
*** udesale has joined #openstack-ansible04:00
*** pcaruana has joined #openstack-ansible04:23
*** cshen has joined #openstack-ansible05:04
*** cshen has quit IRC05:08
*** ivve has joined #openstack-ansible05:43
*** mkuf_ has joined #openstack-ansible05:47
*** mkuf has joined #openstack-ansible05:48
*** mkuf_ has quit IRC05:51
*** DanyC has joined #openstack-ansible06:04
*** kopecmartin|off is now known as kopecmartin06:14
*** electrofelix has joined #openstack-ansible06:16
*** ansmith has quit IRC06:18
*** luksky has joined #openstack-ansible06:20
*** markvoelker has joined #openstack-ansible06:26
*** DanyC has quit IRC06:32
*** ansmith has joined #openstack-ansible06:35
*** pcaruana has quit IRC06:48
*** pcaruana has joined #openstack-ansible06:55
*** DanyC has joined #openstack-ansible07:01
*** witek has joined #openstack-ansible07:15
*** udesale has quit IRC07:20
*** tosky has joined #openstack-ansible07:20
*** jsquare has quit IRC07:20
*** flaper87 has quit IRC07:21
*** DanyC has quit IRC07:22
*** jsquare has joined #openstack-ansible07:23
*** udesale has joined #openstack-ansible07:23
*** cshen has joined #openstack-ansible07:24
*** flaper87 has joined #openstack-ansible07:28
noonedeadpunkmgariepy: actually I wasn't looking into it since my last comment in the bug as had no time for it - decided to postpone futher investigation up to my environments migration to stein, as had a hope, that some workaround might be found and posted to this bug.07:50
evrardjpgod damn it the suse lxc download link changed again07:52
evrardjpI want to rip lxc out of opensuse07:52
noonedeadpunkoh, I see07:56
*** DanyC has joined #openstack-ansible08:19
openstackgerritMerged openstack/openstack-ansible-lxc_hosts master: Use the 'real' name for the LXC image  https://review.opendev.org/65524608:21
*** DanyC has quit IRC08:23
*** DanyC has joined #openstack-ansible08:24
*** luksky has quit IRC08:31
*** cshen has quit IRC08:34
*** gkadam has joined #openstack-ansible08:34
*** hamzaachi has joined #openstack-ansible08:40
*** shardy has joined #openstack-ansible09:07
*** cshen has joined #openstack-ansible09:09
*** cshen has quit IRC09:13
evrardjpnoonedeadpunk: we are still investigating the why but apparently it's a bug in our build system -- so something we couldn't have tracked easily09:21
evrardjpnoonedeadpunk: original image https://download.opensuse.org/repositories/Virtualization:/containers:/images:/openSUSE-Leap-15.0/containers/ is back09:23
noonedeadpunkevrardjp: did I say, that I really love Suse?:))09:36
evrardjphaha09:36
evrardjpI am discussing the best approach forward09:36
noonedeadpunkSo I shouldn't be in a hurry with another patch?09:36
evrardjpyes, but I am tired of these urgent patches :p09:37
evrardjpodyssey4me: are you there?09:37
evrardjpI can't remember why we don't mirror those opensuse containers in infra09:37
*** witek has quit IRC09:40
*** cshen has joined #openstack-ansible09:41
openstackgerritDmitriy Rabotjagov (noonedeadpunk) proposed openstack/openstack-ansible-lxc_hosts master: Returning of the Suse LXC image  https://review.opendev.org/65585109:41
evrardjpnoonedeadpunk: let me handle this :)09:44
noonedeadpunkok, sure09:44
evrardjpthanks for the help though :)09:44
evrardjpand for the quick patch :)09:44
noonedeadpunkthe downside is that we can't backport https://review.opendev.org/#/c/655246/ as is now - will need to wait for confirmed approach and squash changes09:47
*** luksky has joined #openstack-ansible09:48
evrardjpnoonedeadpunk: ok so after further analysis, here is what I think:10:12
evrardjpthe url should be kept as non build specific, so opensuse-leap-image.x86_64-lxc.tar.xz10:13
evrardjpif we keep this, we'll get a 302, which points to multiple duplicate links in headers10:14
evrardjpwhether or not this is mirrored in infra, the redirects should be there, so this would always work10:14
evrardjphowever if it's mirrored in infra, the HEADers sent by the webserver might be different, and not point to a metalink10:14
evrardjpwhich leads me to think that if we mirror in infra we'll not have the metalink issue10:15
evrardjp(the need to rename the file and stuff)10:15
evrardjphowever this means that code by default have to point to infra, which is not recommended10:15
evrardjpwe should still keep an official SUSE url by default, and point to infra in our builds10:16
evrardjp(point to our local mirror seems a normal story for a deployer)10:16
evrardjphowever in that case, it means users not in infra still get that metalink issue.10:16
evrardjpSo....10:17
evrardjpI propose to move to curl, which doesn't respect the first link, as he doesn't know what to do with it, because it's not a known protocoal10:17
noonedeadpunkI'm not sure that issue with metalink is so serious now after Nicolas patch - is's pretty ok for know, imo10:17
evrardjpthe move to curl would be only for opensuse10:17
evrardjpbut it's ugly10:18
evrardjpthe fact we need to do this10:18
evrardjpand I am not sure how well the symlinking works10:18
evrardjpbecause other parts don't follow symlinks10:18
evrardjpthey just consume said file10:18
noonedeadpunkisn't curl even worse?10:19
evrardjpit's slower but safer10:19
noonedeadpunkwe can do "mv" instead of "ln -s"10:19
evrardjphowever in the metalink case it's not slower because aria is respecting metalink specs which in this case might not offer parallel dls10:19
noonedeadpunkbtw, interestingly, that once image is correct one, lxc tests start failing https://review.opendev.org/#/c/655851/10:19
evrardjp(i have to check there)10:20
evrardjpnoonedeadpunk: oh great10:20
evrardjpI guess before things were skipped ?10:21
noonedeadpunkNo, they were passing ok)10:21
noonedeadpunkhttp://logs.openstack.org/46/655246/4/check/openstack-ansible-lxc-dir-opensuse-150/1505ea5/job-output.txt.gz#_2019-04-25_18_13_35_27690310:21
noonedeadpunkvs now http://logs.openstack.org/51/655851/1/check/openstack-ansible-lxc-dir-opensuse-150/1c7f9eb/job-output.txt.gz10:22
noonedeadpunkso that bug with linking was okeyish:)10:22
evrardjpthe apparmor thing10:24
*** nicolasbock has joined #openstack-ansible10:24
noonedeadpunknot really. I've tested this inside my local VM and the problem was in our configs10:25
noonedeadpunkso it was failing at least on this https://review.opendev.org/#/c/652269/1/vars/suse.yml10:26
openstackgerritJean-Philippe Evrard proposed openstack/openstack-ansible-lxc_hosts master: Returning of the Suse LXC image  https://review.opendev.org/65585110:28
evrardjpnoonedeadpunk: I added a note so that we don't change it ever10:28
evrardjpnoonedeadpunk: yes I guess it's the lxc3 issue again10:28
nicolasbockThe image got renamed _again_?10:31
noonedeadpunknicolasbock: yep10:31
nicolasbockThey are not making it very easy :(10:31
noonedeadpunkactually it returned to the previous state10:31
evrardjpnot really because old has 15.0 in it10:32
evrardjpiirc10:32
nicolasbockYes I think you are right evrardjp10:32
evrardjpI am fine with keeping the workaround for the file which follows metalinks10:32
evrardjpI would be more okay with a mv and a ln though10:33
nicolasbockWhich workaround?10:33
noonedeadpunkevrardjp: strange that during this "built bug" there was no lxc problem....10:33
evrardjpit was not using the same image10:34
noonedeadpunkSo I was testing https://download.opensuse.org/repositories/Virtualization:/containers:/images:/openSUSE-Leap-15.1/containers/opensuse-leap-image.x86_64-15.1.0.tar.xz and it redirected to opensuse-leap-image.x86_64-15.1.0-lxc-Build1.1.tar.xz (which is lxc image)10:35
noonedeadpunkStill have an output10:35
evrardjp15.1?10:37
*** cshen has quit IRC10:37
evrardjpwe should stick to 15.0 for now I guess10:37
noonedeadpunkyeah, but still. I don't think that redirect should differ for 15.1 and 15.010:38
evrardjpthe image produced might be different10:39
evrardjpyeah everything is fixed now, so the whole -version.tar.xz and versioneless.tar.xz should work to the lxc image10:39
*** cshen has joined #openstack-ansible10:44
evrardjpcan someone extract the /sr/bin/lxc-create of his machine for me? :)10:47
evrardjp/usr/bin/lxc-create10:47
evrardjpnoonedeadpunk: I am expecting we are not doing the right thing here, and we are using image from images.linuxcontainers.org10:51
evrardjpinstead of using the image we've just created10:51
evrardjpfetched10:51
noonedeadpunkevrardjp: will  /usr/bin/lxc-create from ubuntu suit you?10:51
evrardjpthat would be fine for me10:53
evrardjpand the download template10:53
evrardjpplease :)10:53
noonedeadpunkcan you remind where it's located?:)10:54
noonedeadpunk> /usr/share/lxc/templates/lxc-opensuse ? or the full /usr/share/lxc/templates/ ?10:55
evrardjpin the latter, there must be a download script10:56
evrardjpiirc10:56
evrardjpit's been a while10:57
evrardjpelse I can try spinning up a new machine with this somewhere10:57
noonedeadpunkSo lxc-create is here https://drive.google.com/file/d/1MnW1kErYJSSt-H4qBbReTkkD6B3v-6Kn/ and I'm looking for script11:00
noonedeadpunkand /usr/share/lxc/templates/lxc-download  : http://paste.openstack.org/show/749802/11:04
nicolasbockevrardjp: When you say the 'lxc3 issue', do you mean that the lxc package is breaking again?11:12
nicolasbockCould you have a look at https://review.opendev.org/#/c/651623/ in that case?11:12
*** udesale has quit IRC11:22
evrardjpI agree with Dirk, we should ensure things are mirrored in infra11:27
nicolasbockCan we test it though before we mirror?11:27
nicolasbockIIUC then the mirroring is a performance thing11:28
nicolasbockSo the tests should be green even without the mirror11:28
evrardjpI think here it's not relevant, because I am not sure I see a problem with the mirror artefacts11:28
evrardjpI mean that the problem noonedeadpunk sees is probably unrelated11:28
nicolasbockOk11:29
evrardjpit looks the package is correct11:29
nicolasbockAre we still interested in branching lxc?11:29
noonedeadpunkas far as we're switching to nspawn I'm not sure about that11:29
evrardjpnoonedeadpunk: I haven't seen ppl working actively on nspawn recently11:30
noonedeadpunkfair. but am I right that there was decision about obsoleting lxc in train?11:31
evrardjpwe were all in agreement to move towards it11:31
evrardjpsorry11:31
evrardjpI mean11:31
evrardjpthere was an agreement to move towards it11:32
evrardjpI still have a different opinion :)11:32
noonedeadpunki see:)11:33
fnpanichi11:33
evrardjpwhat matters to me is that we get opensuse leap 15 working11:35
noonedeadpunkand voting?:)11:36
evrardjpthat would be even better11:38
*** hamzaachi has quit IRC11:43
*** witek has joined #openstack-ansible11:48
*** neith has quit IRC12:01
*** neith has joined #openstack-ansible12:03
*** fnpanic has quit IRC12:12
*** joabdearaujo has joined #openstack-ansible12:37
*** markvoelker has quit IRC12:40
evrardjpnicolasbock: do you have a minute?13:08
*** kplant has joined #openstack-ansible13:24
*** hamzaachi has joined #openstack-ansible13:24
*** openstackgerrit has quit IRC13:27
*** ivve has quit IRC13:33
*** marst has joined #openstack-ansible13:40
mnaserevrardjp: I feel your struggles13:43
kplantbefore i dig deeper, anyone else experiencing the task 'python_venv_build: install python packages into the venv' hanging on the utility container? centos7 stable/stein14:06
mnaserkplant: shouldn't be, check /var/log/python_venv_build.log14:19
*** cshen has quit IRC14:20
*** luksky has quit IRC14:29
*** jeh has joined #openstack-ansible14:34
kplantweird.. had a bunch of connection refusals to the repo container, just restarted haproxy and it picked up14:42
kplantthanks14:42
evrardjpmnaser: well I have a good news14:47
evrardjpopensuse leap 15 update channels now have lxc 314:47
evrardjpthe problem is that it is broken14:47
evrardjpbut it also means we're in line with ubuntu14:47
mnaseryay..nay..14:47
evrardjpthe fix is on its way in QA14:48
mnaserevrardjp: I was going to ask if you had some spare time to push rc214:48
evrardjpit's tough today14:48
evrardjpI might do it during the summit14:48
mnaserok no worries14:49
evrardjpI will have more time by then14:49
mnaserstable/stein is in a good place14:49
mnasermore time during the summit?! show me your ways14:49
evrardjpyeah it's gonna be super busy, but today is VERY busy for me :p14:49
nicolasbockYes evrardjp14:51
*** kopecmartin is now known as kopecmartin|off14:53
*** ianychoi_ has joined #openstack-ansible15:00
*** ianychoi has quit IRC15:03
*** Emine has joined #openstack-ansible15:05
*** ansmith has quit IRC15:05
*** ansmith has joined #openstack-ansible15:06
*** mkuf has quit IRC15:06
*** hjohnson1 has joined #openstack-ansible15:08
hjohnson1Can anyone take a look at http://paste.openstack.org/show/749827/... I am having HAProxy errors on a brand new install of Rocky on CentOS and I have been banging my head against the wall all morning with nothing but a sore head to show for it.  Any guidance would be appreciated.15:16
kplantfrontend repo_all-front-115:17
kplant    bind 10.255.61.1:818115:17
kplantshould that not be 10.255.61.206 ?15:17
jamesdenton206 looks to be a backend. The bind would be the VIP in this case. With a .1 IP, I always worry that there is overlap with a gateway address configured somewhere.15:18
noonedeadpunkfrontend should listen on the keepalived VIP15:19
noonedeadpunk+1 :)15:19
hjohnson1The gateway for that subnet is 10.255.61.254... so the .1 is free15:20
kplanthjohnson1: how many infra nodes do you have?15:20
hjohnson1kplant: 315:20
kplantand .1 is the vip?15:20
hjohnson1kplant: Yes... internal_lb_vip_address: 10.255.61.115:21
noonedeadpunkdepends on your configuration:) It's internal_lb_vip_address15:21
jamesdentonIs the .1 address the only one configured? Or do you have a dedicated IP from that network configurd on each infra?15:23
jamesdentonIn other words.. each host would need an addr (.2,.3,.4) and .1 would float15:23
hjohnson1jamesdenton: The infra addresses are 10.255.61.29, .30 and .31.  The 10.255.61.1 address is a second address assigned to infra01 only.15:24
hjohnson1jamesdenton: http://paste.openstack.org/show/749828/15:25
hjohnson1The thing is I can't access the repo from 10.255.61.29.  When i do a wget to 10.255.61.206:8181 I get a no route to host15:28
*** gkadam has quit IRC15:29
*** gkadam has joined #openstack-ansible15:32
*** gyee has joined #openstack-ansible15:36
*** jeh has quit IRC15:45
*** electrofelix has quit IRC16:00
*** jeh has joined #openstack-ansible16:03
*** shardy has quit IRC16:07
*** shardy has joined #openstack-ansible16:09
*** chandankumar is now known as raukadah16:12
*** cshen has joined #openstack-ansible16:14
*** jeh has quit IRC16:15
*** cshen has quit IRC16:19
*** gkadam has quit IRC16:34
*** DanyC has quit IRC16:37
*** marst has quit IRC16:51
joabdearaujohjohnson1: what is the output of the route -n command?16:55
*** Emine has quit IRC17:09
*** Weifan has joined #openstack-ansible17:16
*** hjohnson1 has left #openstack-ansible17:21
mnaserevrardjp, noonedeadpunk: can this be abandoned ? https://review.opendev.org/#/c/642025/17:26
mnasercc nicolasbock too17:26
evrardjpyeah we can abandon17:27
evrardjpI did it17:28
*** Weifan has quit IRC17:33
*** djhankb has quit IRC17:48
*** ivve has joined #openstack-ansible17:50
*** djhankb has joined #openstack-ansible17:51
mnaserodyssey4me: https://bugs.launchpad.net/openstack-ansible/+bug/1826593 let me know if you agree!17:52
openstackLaunchpad bug 1826593 in openstack-ansible "Retire repo_build role" [Undecided,New]17:52
*** openstackgerrit has joined #openstack-ansible17:53
openstackgerritMerged openstack/openstack-ansible master: fix typo  https://review.opendev.org/61613717:53
openstackgerritMerged openstack/openstack-ansible-ops master: Delete the duplicate words in  README.rst  https://review.opendev.org/61756917:55
*** hjohnson1 has joined #openstack-ansible17:57
hjohnson1Thanks all for the help.  I found the problem... it was a dumb mistake on my part :'(17:57
*** shardy has quit IRC17:59
djhankbIs anyone able to offer any advice on a single-NIC configuration with Open vSwitch for  OSA, similar to: https://github.com/openstack/openstack-ansible/blob/master/etc/network/interfaces.d/openstack_interface.cfg.test.example18:00
djhankbI can share my example - something still isn't right, as I can't seem to pass VXLAN traffic18:00
*** shardy has joined #openstack-ansible18:01
*** jeh has joined #openstack-ansible18:02
openstackgerritMerged openstack/openstack-ansible-os_barbican master: Add bugs url link to README  https://review.opendev.org/59852218:04
*** DanyC has joined #openstack-ansible18:04
*** DanyC has quit IRC18:04
*** DanyC has joined #openstack-ansible18:05
*** hamzaachi has quit IRC18:05
openstackgerritMerged openstack/openstack-ansible-os_masakari master: Add bugs url link to README  https://review.opendev.org/59855318:06
*** jbadiapa has quit IRC18:10
*** DanyC has quit IRC18:10
*** cshen has joined #openstack-ansible18:14
*** cshen has quit IRC18:19
*** canori01 has joined #openstack-ansible18:29
canori01hey guys, is there a way to override the mtu for specific interfaces on specific lxc's through the playbooks?18:30
canori01Never mind. I think I found it18:32
openstackgerritAntony Messerli proposed openstack/openstack-ansible-os_keystone master: Remove keystone directory on source installs  https://review.opendev.org/65599218:35
openstackgerritKanishk Godha proposed openstack/openstack-ansible-plugins master: Closes-Bug: #1826246  https://review.opendev.org/65599318:36
openstackbug 1826246 in openstack-ansible "Remove keystone module from plugins" [Undecided,New] https://launchpad.net/bugs/1826246 - Assigned to Kanishk Godha (kanishk-godha)18:36
*** Weifan has joined #openstack-ansible18:49
*** Weifan has quit IRC18:55
*** Weifan has joined #openstack-ansible18:57
openstackgerritMerged openstack/openstack-ansible-os_tempest master: missing document start "---"  https://review.opendev.org/61408819:02
*** hamzaachi has joined #openstack-ansible19:05
openstackgerritKanishk Godha proposed openstack/openstack-ansible-plugins master: Remove keystone module  https://review.opendev.org/65599819:07
*** luksky has joined #openstack-ansible19:09
*** Weifan has quit IRC19:12
*** masterpe has joined #openstack-ansible19:36
*** jbadiapa has joined #openstack-ansible19:45
*** Emine has joined #openstack-ansible19:46
*** Weifan has joined #openstack-ansible20:11
*** cshen has joined #openstack-ansible20:15
*** cshen has quit IRC20:19
*** hjohnson1 has left #openstack-ansible20:22
openstackgerritMerged openstack/openstack-ansible-ceph_client master: yum: add liboath to packages installed from EPEL  https://review.opendev.org/65279620:30
*** DanyC has joined #openstack-ansible20:33
*** witek has quit IRC20:40
*** kplant has quit IRC20:43
*** Emine has quit IRC20:48
*** djhankb has quit IRC21:01
*** DanyC has quit IRC21:01
*** pcaruana has quit IRC21:16
*** KeithMnemonic has quit IRC21:26
*** luksky has quit IRC21:38
*** jeh has quit IRC22:07
*** mhayden has quit IRC22:12
*** cshen has joined #openstack-ansible22:15
*** cshen has quit IRC22:19
*** ivve has quit IRC22:33
*** Weifan has quit IRC22:34
*** mhayden has joined #openstack-ansible22:39
*** Weifan has joined #openstack-ansible22:50
*** kplant has joined #openstack-ansible22:52
*** tosky has quit IRC23:08
*** Weifan has quit IRC23:11
*** hamzaachi has quit IRC23:11
*** Weifan has joined #openstack-ansible23:27
*** Weifan has quit IRC23:32
*** Weifan has joined #openstack-ansible23:34
*** Weifan has quit IRC23:39
*** Weifan has joined #openstack-ansible23:40
*** nicolasbock has quit IRC23:40
*** mhayden has quit IRC23:45
*** gyee has quit IRC23:49

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