sdake_ | oslo.i18n doesn't work from the pip install | 00:00 |
---|---|---|
*** g3ek has quit IRC | 00:00 | |
sdake_ | might be an upper constraints problem | 00:00 |
*** saneax is now known as saneax-_-|AFK | 00:00 | |
sdake_ | most people would never notice - however- I have a fresh box | 00:00 |
*** schwicht has joined #openstack-kolla | 00:01 | |
sbezverk | srwilkers_: ping | 00:02 |
*** sayantani01 has quit IRC | 00:07 | |
sbezverk | sdake_: do you have by chance srwilkers_ email? | 00:07 |
sdake_ | sbezverk you can find it in gerrit | 00:07 |
sdake_ | moemnt | 00:07 |
sdake_ | https://review.openstack.org/#/admin/groups/1380,members | 00:08 |
kfox1111 | sdake_: worst worst case, just having a 3.0.2 for kolla-toolbox and kubernetes-entrypoint would really help. | 00:08 |
sdake_ | kfox1111 let me build and push and then we can see if they are busted via the gate ;) | 00:09 |
*** g3ek has joined #openstack-kolla | 00:09 | |
kfox1111 | sdake_: sounds good. | 00:09 |
sdake_ | built binary - now buildilng source | 00:09 |
sdake_ | then i'll push them | 00:09 |
sdake_ | binary built correctly - yay :) | 00:09 |
kfox1111 | nice. | 00:11 |
sbezverk | sdake_: thanks | 00:12 |
*** Pavo has quit IRC | 00:14 | |
*** Pavo has joined #openstack-kolla | 00:14 | |
*** hfu has joined #openstack-kolla | 00:16 | |
*** tonanhngo has quit IRC | 00:16 | |
*** saneax-_-|AFK is now known as saneax | 00:16 | |
*** schwicht has quit IRC | 00:19 | |
*** schwicht has joined #openstack-kolla | 00:22 | |
*** schwicht has quit IRC | 00:24 | |
*** TxGirlGeek has quit IRC | 00:26 | |
kfox1111 | bbiab | 00:26 |
*** zhurong has joined #openstack-kolla | 00:28 | |
*** inc0 has quit IRC | 00:32 | |
*** celo_r has quit IRC | 00:34 | |
*** liyifeng has joined #openstack-kolla | 00:39 | |
*** sdake has joined #openstack-kolla | 00:40 | |
*** sdake_ has quit IRC | 00:41 | |
*** duonghq_ has joined #openstack-kolla | 00:41 | |
duonghq_ | morning guys | 00:41 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Glance service orchestration with entrypoint https://review.openstack.org/410736 | 00:43 |
sbezverk | duonghq_: morning, had any luck with halcyon? | 00:43 |
*** sayantani01 has joined #openstack-kolla | 00:44 | |
sdake | duonghq_ sbezverk has started work on operators | 00:44 |
sdake | duonghq_ if you are looking for some cool work - sbezverk can use a partner :) | 00:44 |
sdake | kfox1111 images are built - pushing shortly | 00:44 |
duonghq_ | sdake, I saw that, trying to finish ceph microservices, and will move to operator soon | 00:45 |
duonghq_ | thank you sdake | 00:45 |
sdake | duonghq_ nice ;) | 00:45 |
duonghq_ | sbezverk, I still keep the vagrant VM running | 00:45 |
duonghq_ | sdake, which image are you building? (base and type) | 00:45 |
*** hfu has quit IRC | 00:46 | |
*** saneax is now known as saneax-_-|AFK | 00:46 | |
*** hfu has joined #openstack-kolla | 00:47 | |
*** sdake_ has joined #openstack-kolla | 00:47 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 00:47 |
*** duonghq has left #openstack-kolla | 00:48 | |
*** duonghq_ is now known as duonghq | 00:49 | |
*** sdake has quit IRC | 00:49 | |
*** schwicht has joined #openstack-kolla | 00:52 | |
sdake_ | any idea who is building master on docker hub? | 00:54 |
*** schwicht has quit IRC | 00:55 | |
*** schwicht has joined #openstack-kolla | 00:56 | |
Pavo | does swift need anything special in kolla-ansible other than just enabling it in globals? | 00:57 |
sdake_ | pavo you have to create the disks | 00:59 |
sdake_ | or register them or something | 00:59 |
sdake_ | it is in the swift guide | 00:59 |
Pavo | so can't use cepf as a backend for it | 00:59 |
sdake_ | nope | 00:59 |
Pavo | damn | 00:59 |
sdake_ | kfox1111 - binary is half pushed | 01:00 |
Pavo | ok can those disk be made on the storage node and not the compute node? | 01:01 |
sdake_ | pushing source centos next | 01:01 |
sdake_ | pavo - I htink they cna be anywhere a storage network has access to - so it doesn't matter in particular | 01:01 |
duonghq | sdake_, how often do we push master image? | 01:01 |
sdake_ | but it may have to only be on the storage nodes - not certain | 01:01 |
sdake_ | duonghq no idea - I didn't think we were pushing master at all | 01:01 |
Pavo | ok cool will make one of my ceph drives into swift | 01:02 |
duonghq | so how do we deal with kolla-k8s? | 01:02 |
sdake_ | duonghq we are using 3.0.2 images at present | 01:02 |
sdake_ | rather 3.0.1 - soon to be 3.0.2 | 01:02 |
sdake_ | which I think is wrong - we need to be testing master of openstack images | 01:02 |
duonghq | sdake_, agree on this point | 01:03 |
duonghq | so I wonder how we achieve that | 01:03 |
sdake_ | we build in the gate instead of pulling from docker hub and push to a local registry in the gate | 01:03 |
*** PramodJ has joined #openstack-kolla | 01:03 | |
sdake_ | this is called cross repo gating :) | 01:04 |
sdake_ | we need it done - haven't had time to address it and kfox1111 hasn't either | 01:04 |
duonghq | We can make a repo persistent in gate? | 01:04 |
sdake_ | no it dies every time - so it needs to be rebuilt every time | 01:04 |
duonghq | I heard somebody talk about this, just wonder if we can make it persistent | 01:04 |
sdake_ | inc0 is working on a patch which puts a registry somewhere in the middle of that | 01:04 |
sdake_ | so itis persistent | 01:04 |
sdake_ | not sure how it works :) | 01:05 |
sdake_ | atm its based on puppet - so I am unable to spend 6-8 weeks to learn how to puppetize something like docker hub | 01:05 |
*** MasterOfBugs has quit IRC | 01:05 | |
duonghq | ah, good news, building image is painful if we must do this every time | 01:05 |
sdake_ | full moon and friday the 13th all in one | 01:06 |
sdake_ | should be an interesting friday :) | 01:06 |
duonghq | didn't notice it | 01:07 |
*** pramodrj07 has quit IRC | 01:07 | |
duonghq | what should happen when both beautiful things happen? | 01:07 |
sdake_ | who knows | 01:08 |
*** PramodJ has quit IRC | 01:08 | |
*** lrensing has quit IRC | 01:10 | |
*** pramodrj07 has joined #openstack-kolla | 01:10 | |
*** MasterOfBugs has joined #openstack-kolla | 01:10 | |
*** pramodrj07 has quit IRC | 01:14 | |
*** MasterOfBugs has quit IRC | 01:14 | |
kfox1111 | back. | 01:15 |
duonghq | sup kfox1111 | 01:15 |
kfox1111 | duonghq: hey. :) | 01:15 |
kfox1111 | sdake_: nice. :) | 01:15 |
*** zhubingbing has joined #openstack-kolla | 01:16 | |
jemcevoy | Mech422 ... Question on your vlan setup with OVS. Does that network config file you sent me in http://paste.openstack.org/show/594621/ run in one of the openvswitch containers or on the bare metal server? | 01:19 |
jemcevoy | I ask because I'll need some of this setup before the openvswitch_db is created.... | 01:21 |
*** sayantan_ has joined #openstack-kolla | 01:25 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 01:27 |
*** Pavo has quit IRC | 01:27 | |
*** mannamne has quit IRC | 01:27 | |
*** sayantani01 has quit IRC | 01:27 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Glance service orchestration with entrypoint https://review.openstack.org/410736 | 01:32 |
*** jrobinson has joined #openstack-kolla | 01:32 | |
sdake_ | kfox1111 - what work are we waiting on to tag? | 01:32 |
sdake_ | kfox1111 ubuntu binary images? | 01:33 |
kfox1111 | centos-binary and ubuntu-binary | 01:33 |
sdake_ | kfox1111 - I've got to jet soon and feed the kids- so won't be able to build ubuntu atleast unitil I get back | 01:33 |
sdake_ | kfox1111 i don't htink inc0 finished buildign ubuntu binary | 01:33 |
kfox1111 | inc0 said he's pushing ubuntu | 01:33 |
sdake_ | yes but we need it tonight to tag tomorrow right? | 01:33 |
kfox1111 | asap, yeah. | 01:34 |
kfox1111 | yay! docker pull kolla/centos-binary-kolla-toolbox:3.0.2 worked. | 01:34 |
kfox1111 | getting close now. | 01:35 |
sdake_ | kfox1111 can you run that through the gate | 01:35 |
sdake_ | kfox1111 and see if centos binary is working properly | 01:35 |
kfox1111 | did all upload? | 01:36 |
kfox1111 | rechecked. | 01:37 |
kfox1111 | https://review.openstack.org/#/c/419082/ | 01:37 |
kfox1111 | once that goes green, we're good to go. | 01:37 |
sdake_ | kfox1111 - ubuntu binary is building now (on centos) | 01:38 |
sdake_ | it may not be totally working - so inc0 will have to push again | 01:38 |
sbezverk | sdake_: I believe inc0 was taking care of ubuntu | 01:39 |
sdake_ | sbezverk yes but it needs to be built now | 01:39 |
sdake_ | sbezverk and it hasn't been built | 01:39 |
sbezverk | sdake_: I see, did not know that it is so much trouble to start building | 01:39 |
sdake_ | i don't have an ubuntu system to build on | 01:39 |
sdake_ | so its a struggle ;) | 01:40 |
sbezverk | sdake_: hm I build ubuntu on centos machine, works like a charm | 01:40 |
sdake_ | ok dinner time - images are pushing as we speak | 01:40 |
sdake_ | i'll try to get on irc at dinner if possible | 01:40 |
kfox1111 | k. thx. | 01:42 |
sbezverk | kfox1111: after renaming objects we have bunch of left overs folders | 01:43 |
*** lrensing has joined #openstack-kolla | 01:43 | |
*** tovin07_ is now known as tovin07 | 01:45 | |
kfox1111 | yeah, side affect of generateing values files. :/ | 01:45 |
kfox1111 | fresh checkout's easiest way to get rid of them. :/ | 01:46 |
sbezverk | kfox1111: but not after resolving merge conflicts.. :( | 01:47 |
kfox1111 | yeah. those are going to be a bit of a pain. dealing with that myself. :/ | 01:48 |
kfox1111 | the neutron service package is quite a ways away from tunk now. | 01:48 |
*** Pavo has joined #openstack-kolla | 01:48 | |
sbezverk | kfox1111: yeah mariadb and glance are all screwed up too, I mean in services.. but well at least when they green they will get merged ;-) | 01:50 |
kfox1111 | yeah. the neutron one's looking really nice to config after the merge though. | 01:51 |
kfox1111 | went from a loop across all of the packages with config for each, to like one general config file. :) | 01:51 |
kfox1111 | will post in a few. | 01:51 |
sbezverk | kfox1111: check it out tomorrow.. I am off for now, good night | 01:53 |
*** adrian_otto has quit IRC | 01:53 | |
kfox1111 | sbezverk: ok. good nite. | 01:53 |
openstackgerrit | desheng proposed openstack/kolla-ansible: Move etcd precheck into its own role https://review.openstack.org/419238 | 01:58 |
*** tovin07_ has joined #openstack-kolla | 02:01 | |
*** pramodrj07 has joined #openstack-kolla | 02:02 | |
*** MasterOfBugs has joined #openstack-kolla | 02:02 | |
*** unicell has quit IRC | 02:02 | |
openstackgerrit | Kuo-tung Kao proposed openstack/kolla-ansible: provide a universal collectd template https://review.openstack.org/414499 | 02:06 |
*** schwicht has quit IRC | 02:15 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: WIP: neutron service package https://review.openstack.org/412239 | 02:15 |
*** schwicht has joined #openstack-kolla | 02:16 | |
*** yingjun has joined #openstack-kolla | 02:21 | |
kfox1111 | sdake_: this is new: http://logs.openstack.org/82/419082/4/check/gate-kolla-kubernetes-deploy-centos-binary-ceph-nv/ad1f5a5/console.html | 02:23 |
*** negronjl has quit IRC | 02:28 | |
*** Pavo has quit IRC | 02:28 | |
*** Pavo has joined #openstack-kolla | 02:29 | |
zhubingbing | hello | 02:33 |
zhubingbing | https://review.openstack.org/#/c/418811/ | 02:33 |
Pavo | has anyone ever seen this before after a deployment https://www.dropbox.com/s/47z1k6nqvxkse66/Screenshot%202017-01-11%2021.38.38.png?dl=0 | 02:38 |
Pavo | getting that when I try to login | 02:39 |
Pavo | [Wed Jan 11 21:37:34.711409 2017] [:error] [pid 33] RuntimeError: Unable to create a new session key. It is likely that the cache is unavailable. | 02:42 |
Pavo | thats in horizon.log | 02:42 |
duonghq | kfox1111, should kolla-kubernetes work with master image? | 02:48 |
kfox1111 | duonghq: should, but untested so far. | 02:51 |
duonghq | kfox1111, copy | 02:51 |
*** schwicht_at_work has joined #openstack-kolla | 02:51 | |
*** schwicht has quit IRC | 02:52 | |
*** tovin07 has quit IRC | 02:52 | |
*** schwicht_at_work has quit IRC | 02:56 | |
*** schwicht has joined #openstack-kolla | 02:56 | |
*** tovin07 has joined #openstack-kolla | 02:56 | |
*** cliles has quit IRC | 03:02 | |
*** goldyfruit has joined #openstack-kolla | 03:07 | |
*** duonghq has quit IRC | 03:09 | |
*** tovin07 has quit IRC | 03:09 | |
*** sdake_ has quit IRC | 03:10 | |
*** tovin07 has joined #openstack-kolla | 03:10 | |
*** hieulq has quit IRC | 03:10 | |
*** tovin07_ has quit IRC | 03:11 | |
*** sdake has joined #openstack-kolla | 03:11 | |
*** lrensing has quit IRC | 03:12 | |
sdake | kfox1111 my kernel oopsed during building | 03:13 |
*** cliles has joined #openstack-kolla | 03:13 | |
*** mgiles has quit IRC | 03:14 | |
*** sdake_ has joined #openstack-kolla | 03:15 | |
*** adrian_otto has joined #openstack-kolla | 03:15 | |
*** adrian_otto has joined #openstack-kolla | 03:16 | |
*** ocaner has quit IRC | 03:16 | |
*** sdake has quit IRC | 03:17 | |
*** sdake_ has quit IRC | 03:19 | |
*** sdake has joined #openstack-kolla | 03:22 | |
*** adrian_otto has quit IRC | 03:22 | |
*** duonghq has joined #openstack-kolla | 03:24 | |
*** tovin07_ has joined #openstack-kolla | 03:24 | |
*** hieulq has joined #openstack-kolla | 03:25 | |
sdake | Jeffrey4l_ ping | 03:29 |
sdake | Jeffrey4l_ when your about - did you happen to test 3.0.2 with rbd? | 03:29 |
sdake | rather with ceph | 03:29 |
sdake | Jeffrey4l_ the k8s gate (which uses ceph unlike the rest of the gates) returns http://logs.openstack.org/82/419082/4/check/gate-kolla-kubernetes-deploy-centos-binary-ceph-nv/ad1f5a5/console.html#_2017-01-12_02_21_11_929763 | 03:30 |
Mech422 | jemcevoy: those configs run on the bare metal - also, kolla will 'restart' ovs which caused me to lose config - I had to comment out the restart or something | 03:31 |
*** lrensing has joined #openstack-kolla | 03:34 | |
*** zhurong has quit IRC | 03:39 | |
*** zhubingbing has quit IRC | 03:39 | |
sp_ | sdake: hi... | 03:42 |
*** schwicht_at_work has joined #openstack-kolla | 03:42 | |
*** schwicht has quit IRC | 03:44 | |
*** sayantan_ has quit IRC | 03:45 | |
*** prameswar has joined #openstack-kolla | 03:45 | |
sdake | sup sp_ | 03:45 |
sp_ | sdake: how kolla-k8s differ from Stackanetes ? | 03:46 |
Jeffrey4l_ | sdake, no. haven't test with ceph. | 03:46 |
sp_ | sdake: i am confused here. | 03:46 |
sdake | Jeffrey4l_ can you do so - I think 3.0.2 is doa in multiple ways | 03:47 |
sdake | sp_ confused how | 03:47 |
sp_ | sdake: Stacknetes also provide life cycle management using k8s as kolla-k8s does. | 03:48 |
Jeffrey4l_ | ok. will try. | 03:48 |
sdake | Jeffrey4l_ i just filed this bug: https://bugs.launchpad.net/kolla/+bug/1655817 | 03:49 |
openstack | Launchpad bug 1655817 in kolla "pip install kolla==3.0.2 followed by kolla-build fails" [Critical,Triaged] | 03:49 |
openstackgerrit | rajiv proposed openstack/kolla-kubernetes: Use upper-constraints for tox envs https://review.openstack.org/419253 | 03:49 |
sdake | Jeffrey4l_ also if ceph is busted - that is another DOA point | 03:49 |
sdake | Jeffrey4l_ also I created 3.0.3 :) | 03:50 |
Jeffrey4l_ | roger. | 03:50 |
Jeffrey4l_ | thansk. | 03:50 |
sdake | if 3.0.2 is really DOA we should strive to release newton update asap to fix the doa-ness | 03:50 |
Jeffrey4l_ | yes. | 03:50 |
sdake | sp_ not sure where the question is coming from | 03:50 |
sdake | sp_ i never said stackanetes didn't provide life cycle management | 03:51 |
sdake | nor did I say it does - I have made no statements about stackanetes at all | 03:51 |
sp_ | sdake: of-course you didn't. | 03:53 |
sp_ | sdake thanks | 03:53 |
sdake | sp_ happy to help | 03:53 |
sp_ | sdake: may be i should go little more dip down into this. | 03:53 |
sdake | i dont know what dip down means sorry doesn't pass parser after 15 hours of workday :( | 03:54 |
*** qiliang27 has joined #openstack-kolla | 03:55 | |
sp_ | sdake: I going through Stacknetes for better understanding and looking for exactly what is difference between stacknetes and kolla-k8s. | 03:55 |
sdake | sp_ i have not examined stackanetes all that much | 03:56 |
*** sayantani01 has joined #openstack-kolla | 03:56 | |
sdake | and by all that much I mean i looked at it once long ago and havne't really looked since | 03:56 |
sdake | if there isn't much delta - we should figure out how to fold the communities together | 03:57 |
sdake | if that is a thing that can be done :) | 03:57 |
*** Pavo has quit IRC | 03:57 | |
sdake | 190 bugs in ocata 3 | 03:57 |
* sdake sadfaces | 03:57 | |
sdake | the repo split thing really screwed up kolla's mojo with implementation | 03:58 |
sp_ | sdake: You are right. We should fold the communities if not much difference. | 03:59 |
sp_ | sdake: thanks | 03:59 |
*** schwicht has joined #openstack-kolla | 04:01 | |
*** schwicht_at_work has quit IRC | 04:01 | |
sdake | kfox1111 all of the 3.0.2 images have been pushed | 04:10 |
duonghq | kfox1111, around? does we add kubernetes-entrypoint just in ocata? | 04:11 |
sdake | kfox1111 if they are broken - it may be that 3.0.2 is doa with regard to ceph | 04:11 |
*** goldyfruit has quit IRC | 04:11 | |
*** Jeffrey4l__ has joined #openstack-kolla | 04:11 | |
*** l4yerffeJ has joined #openstack-kolla | 04:11 | |
*** l4yerffeJ_ has quit IRC | 04:13 | |
*** Jeffrey4l_ has quit IRC | 04:14 | |
*** mdnadeem has joined #openstack-kolla | 04:18 | |
sdake | Jeffrey4l__ http://tracker.ceph.com/issues/16289 | 04:18 |
Jeffrey4l__ | yep. it looks the same. | 04:18 |
Jeffrey4l__ | what the ceph client version in k8s gate? | 04:19 |
*** jrobinson is now known as jrobinson-afk | 04:20 | |
sdake | no clue | 04:22 |
*** adrian_otto has joined #openstack-kolla | 04:28 | |
*** zhubingbing has joined #openstack-kolla | 04:28 | |
*** adrian_otto has quit IRC | 04:28 | |
*** skramaja has joined #openstack-kolla | 04:30 | |
*** jtriley has joined #openstack-kolla | 04:32 | |
Mech422 | Morning | 04:41 |
*** sp__ has joined #openstack-kolla | 04:41 | |
*** jrobinson-afk is now known as jrobinson | 04:46 | |
*** jtriley has quit IRC | 04:47 | |
sdake | Jeffrey4l__ this looks like the ceph client call: | 04:49 |
sdake | 2017-01-12 02:21:10.080105 | + kubectl exec ceph-admin -c main --namespace=kolla -- /bin/bash -c 'timeout 240s rbd create kollavolumes/mariadb --size 1024' | 04:49 |
sdake | looks like its calling ceph-admin | 04:49 |
Jeffrey4l__ | ceph-admin should be the resource/pod name. | 04:50 |
sdake | it is | 04:51 |
Jeffrey4l__ | it is calling "rbd create ... " | 04:51 |
Jeffrey4l__ | sdake, are the 3.0.2 image pushed into hub.docker.com already? | 04:53 |
sdake | Jeffrey4l__ i've been working on it | 04:53 |
sdake | and that gate job was 3.0.2 | 04:53 |
sdake | iiuc | 04:53 |
sdake | i have pushed centos binary | 04:53 |
sdake | and ubuntu binary | 04:53 |
sdake | now doing centos source | 04:53 |
Jeffrey4l__ | so k8s's gate is using the centos-binary u just pushed? | 04:54 |
sdake | yup | 04:54 |
Jeffrey4l__ | ok. i am trying to set up a env based on centos-binary 3.0.2 | 04:54 |
*** unicell has joined #openstack-kolla | 04:55 | |
sdake | i used sudo pip install kolla==3.0.2 | 04:55 |
Jeffrey4l__ | to build the images? | 04:56 |
sdake | yup | 04:56 |
sdake | that is how you build images pushed to docker hub | 04:56 |
sdake | that way the tags are correct | 04:56 |
Jeffrey4l__ | roger. that is correct. | 04:56 |
*** saneax-_-|AFK is now known as saneax | 05:01 | |
*** v1k0d3n has quit IRC | 05:07 | |
*** v1k0d3n has joined #openstack-kolla | 05:07 | |
*** v1k0d3n has quit IRC | 05:12 | |
*** sayantani01 has quit IRC | 05:14 | |
*** sayantani01 has joined #openstack-kolla | 05:14 | |
*** sayantani01 has quit IRC | 05:19 | |
*** schwicht has quit IRC | 05:21 | |
*** sacharya has quit IRC | 05:25 | |
*** salv-orlando has joined #openstack-kolla | 05:26 | |
*** salv-orlando has quit IRC | 05:31 | |
*** zhubingbing has quit IRC | 05:39 | |
*** zhubingbing has joined #openstack-kolla | 05:39 | |
*** cliles has quit IRC | 05:45 | |
*** awidders_ has joined #openstack-kolla | 05:48 | |
*** awiddersheim has quit IRC | 05:51 | |
*** suka_ has joined #openstack-kolla | 06:09 | |
*** unicell1 has joined #openstack-kolla | 06:16 | |
*** unicell has quit IRC | 06:17 | |
*** absubram has quit IRC | 06:17 | |
suka_ | Can anyone help me solve the "pod (xxx) failed to fit in any node fit failure summary on nodes : MatchNodeSelector (1)" error on kubernetes during the final steps of kolla-kubernetes configuration/installation | 06:20 |
*** unicell1 has quit IRC | 06:21 | |
*** unicell has joined #openstack-kolla | 06:21 | |
*** mdnadeem has quit IRC | 06:22 | |
*** zhurong has joined #openstack-kolla | 06:23 | |
*** sacharya has joined #openstack-kolla | 06:25 | |
openstackgerrit | HongZhi XU proposed openstack/kolla-ansible: Move swift precheck into its own role https://review.openstack.org/418762 | 06:25 |
*** skramaja has quit IRC | 06:26 | |
*** skramaja has joined #openstack-kolla | 06:26 | |
suka_ | Can anyone help me solve the "pod (xxx) failed to fit in any node fit failure summary on nodes : MatchNodeSelector (1)" error on kubernetes during the final steps of kolla-kubernetes configuration/installation | 06:27 |
jascott1 | suka_ it cant schedule the pod, maybe its asking for CPU shares or other resource thats not available? | 06:27 |
*** salv-orlando has joined #openstack-kolla | 06:28 | |
jascott1 | is that the same error when node is missing required label? cant recall | 06:30 |
*** sacharya has quit IRC | 06:30 | |
suka_ | @jascott1 I have made sure that I given enough resources. Kubernetes is able to recognize all the openstack services. The problem is just with scheduling the pods | 06:31 |
jascott1 | suka_ do you have labels on the nodes? | 06:31 |
suka_ | Yes. I do | 06:31 |
jascott1 | it should tell you exact resource problem, if its not in describe or logs maybe events | 06:32 |
*** salv-orlando has quit IRC | 06:32 | |
*** mdnadeem has joined #openstack-kolla | 06:34 | |
openstackgerrit | Wei Cao proposed openstack/kolla-ansible: Fix cleanup-containers https://review.openstack.org/419296 | 06:35 |
*** suka_ has quit IRC | 06:36 | |
*** msimonin has quit IRC | 06:39 | |
*** msimonin has joined #openstack-kolla | 06:39 | |
*** msimonin has quit IRC | 06:40 | |
*** msimonin has joined #openstack-kolla | 06:40 | |
*** msimonin has quit IRC | 06:41 | |
*** msimonin has joined #openstack-kolla | 06:41 | |
*** msimonin has quit IRC | 06:42 | |
*** msimonin has joined #openstack-kolla | 06:42 | |
*** msimonin has quit IRC | 06:42 | |
*** msimonin has joined #openstack-kolla | 06:43 | |
*** msimonin has quit IRC | 06:43 | |
*** satyar has joined #openstack-kolla | 06:46 | |
*** salv-orlando has joined #openstack-kolla | 06:57 | |
*** lrensing has quit IRC | 06:59 | |
openstackgerrit | Wei Cao proposed openstack/kolla-ansible: Change 1-subnet to public1-subnet in tools/init-runonce https://review.openstack.org/419309 | 07:13 |
*** duonghq has quit IRC | 07:16 | |
*** mdnadeem has quit IRC | 07:24 | |
*** mdnadeem has joined #openstack-kolla | 07:25 | |
*** tovin07 has quit IRC | 07:28 | |
*** logan- has quit IRC | 07:47 | |
*** logan- has joined #openstack-kolla | 07:56 | |
*** matrohon has joined #openstack-kolla | 07:59 | |
*** sacharya has joined #openstack-kolla | 08:01 | |
*** haplo37_ has quit IRC | 08:05 | |
*** sacharya has quit IRC | 08:06 | |
*** haplo37_ has joined #openstack-kolla | 08:09 | |
*** jrobinson has quit IRC | 08:12 | |
*** PramodJ has joined #openstack-kolla | 08:15 | |
*** pramodrj07 has quit IRC | 08:18 | |
*** MasterOfBugs has quit IRC | 08:18 | |
*** MasterOfBugs has joined #openstack-kolla | 08:18 | |
openstackgerrit | Surya Prakash Singh proposed openstack/kolla-ansible: Docs update related to hostname resolution in RabbitMQ https://review.openstack.org/418252 | 08:22 |
*** strigazi_AFK is now known as strigazi | 08:28 | |
*** mgoddard has joined #openstack-kolla | 08:32 | |
*** caowei has quit IRC | 08:39 | |
*** salv-orl_ has joined #openstack-kolla | 08:43 | |
*** salv-orlando has quit IRC | 08:43 | |
*** egonzalez has joined #openstack-kolla | 08:44 | |
egonzalez | morning | 08:44 |
*** DanyC has joined #openstack-kolla | 08:46 | |
*** DanyC has quit IRC | 08:50 | |
*** skramaja has quit IRC | 08:52 | |
Jeffrey4l__ | sdake, around? | 08:53 |
*** Serlex has joined #openstack-kolla | 08:55 | |
Jeffrey4l__ | sdake, i tested 3.0.2 with ceph, it works as expected. | 08:57 |
*** athomas has joined #openstack-kolla | 08:58 | |
openstackgerrit | jianyi proposed openstack/kolla-ansible: Add zun ansible role https://review.openstack.org/416179 | 08:58 |
*** yingjun has quit IRC | 09:00 | |
*** david-lyle has quit IRC | 09:01 | |
*** sacharya has joined #openstack-kolla | 09:01 | |
*** mgoddard has quit IRC | 09:02 | |
*** prameswar has quit IRC | 09:03 | |
openstackgerrit | jianyi proposed openstack/kolla-ansible: Add zun ansible role https://review.openstack.org/416179 | 09:04 |
*** sacharya has quit IRC | 09:06 | |
openstackgerrit | Surya Prakash Singh proposed openstack/kolla: Heat configuration param change for heat-api section https://review.openstack.org/419330 | 09:10 |
*** DanyC has joined #openstack-kolla | 09:11 | |
*** PramodJ has quit IRC | 09:13 | |
*** MasterOfBugs has quit IRC | 09:13 | |
*** DanyC has quit IRC | 09:15 | |
*** DanyC has joined #openstack-kolla | 09:16 | |
openstackgerrit | Surya Prakash Singh proposed openstack/kolla: Heat configuration param change for heat-api section https://review.openstack.org/419330 | 09:16 |
openstackgerrit | Wei Cao proposed openstack/kolla-ansible: Fix cleanup-containers https://review.openstack.org/419296 | 09:18 |
*** DanyC has quit IRC | 09:21 | |
*** david-lyle has joined #openstack-kolla | 09:24 | |
satyar | Hi jeffrey4l__ | 09:24 |
*** shardy has joined #openstack-kolla | 09:25 | |
*** salv-orl_ has quit IRC | 09:26 | |
*** gfidente has joined #openstack-kolla | 09:27 | |
*** gfidente has joined #openstack-kolla | 09:27 | |
*** caowei has joined #openstack-kolla | 09:30 | |
*** qiliang28 has joined #openstack-kolla | 09:37 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla: Updated from global requirements https://review.openstack.org/419341 | 09:37 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla-ansible: Updated from global requirements https://review.openstack.org/419342 | 09:37 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla-kubernetes: Updated from global requirements https://review.openstack.org/419343 | 09:37 |
*** qiliang27 has quit IRC | 09:37 | |
*** qiliang28 is now known as qiliang27 | 09:37 | |
*** zhurong has quit IRC | 09:44 | |
*** mgoddard has joined #openstack-kolla | 09:44 | |
*** hfu has quit IRC | 09:48 | |
openstackgerrit | Surya Prakash Singh proposed openstack/kolla: Heat configuration param change for heat-api section https://review.openstack.org/419330 | 09:53 |
*** qiliang27 has quit IRC | 09:53 | |
*** mannamne has joined #openstack-kolla | 09:57 | |
*** prameswar has joined #openstack-kolla | 10:04 | |
*** nwik has quit IRC | 10:12 | |
Jeffrey4l__ | sup satyar, | 10:21 |
*** ccesario has joined #openstack-kolla | 10:22 | |
openstackgerrit | Wei Cao proposed openstack/kolla-ansible: Fix cleanup-containers https://review.openstack.org/419296 | 10:22 |
*** zhurong has joined #openstack-kolla | 10:25 | |
openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Ansible config for nova-serialproxy console https://review.openstack.org/415870 | 10:27 |
*** satyar has quit IRC | 10:35 | |
sp_ | pbourke: it would be great if you review it again https://review.openstack.org/#/c/411313/ | 10:36 |
*** caowei has quit IRC | 10:41 | |
*** gfidente has quit IRC | 10:42 | |
sp_ | pbourke: thanks for review | 10:43 |
*** JoseMello has joined #openstack-kolla | 10:43 | |
pbourke | sp_: no problem | 10:44 |
*** pbourke has quit IRC | 10:45 | |
*** pbourke has joined #openstack-kolla | 10:45 | |
openstackgerrit | Merged openstack/kolla-ansible: Move etcd precheck into its own role https://review.openstack.org/416063 | 10:47 |
sp_ | egonzalez: pbourke: Jeffrey4l__: zhubingbing: I have incorporated your comments please review again https://review.openstack.org/#/c/418252/ | 10:48 |
sp_ | need to focus on BP too | 10:49 |
zhubingbing | ok | 10:49 |
zhubingbing | fix now | 10:49 |
sp_ | yes | 10:49 |
zhubingbing | done | 10:50 |
*** shardy has quit IRC | 10:51 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move kibana precheck into its own role https://review.openstack.org/418834 | 10:51 |
Jeffrey4l__ | sp_, nit :) | 10:53 |
openstackgerrit | Eduardo Gonzalez proposed openstack/kolla-ansible: Unify jinja syntax https://review.openstack.org/419415 | 10:53 |
openstackgerrit | Merged openstack/kolla-ansible: Move murano precheck into its own role https://review.openstack.org/415113 | 10:53 |
*** saneax is now known as saneax-_-|AFK | 10:55 | |
sp_ | Jeffrey4l__: it was a bit confusing that time to decide the exact section for this. After the egonzalez comment i moved it to that section. | 10:56 |
sp_ | Jeffrey4l__: though np, I will create new section and will move it to that section | 10:57 |
egonzalez | pbourke: were you yesterday in the meeting? | 10:57 |
Jeffrey4l__ | but full qulified domain name is not related rabbitmq hostname issue. | 10:57 |
Jeffrey4l__ | they are totally different. fqdn is not required in most of cause. | 10:58 |
Jeffrey4l__ | but rabbitmq hostname is essential . | 10:58 |
*** Jeffrey4l__ is now known as Jeffrey4l | 10:58 | |
Jeffrey4l | egonzalez, pbourke could u review this https://review.openstack.org/415504 | 11:00 |
*** sacharya has joined #openstack-kolla | 11:02 | |
pbourke | egonzalez: yes | 11:03 |
pbourke | egonzalez: but wasn't paying close attention | 11:03 |
*** qiliang27 has joined #openstack-kolla | 11:03 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Bump openstack_release to 3.0.3 in group_vars https://review.openstack.org/419417 | 11:03 |
egonzalez | pbourke: just was no make you know ptg brainstorming in case you missed https://etherpad.openstack.org/p/kolla-ocata-ptg | 11:04 |
pbourke | egonzalez: ah thanks | 11:04 |
*** mgoddard has quit IRC | 11:05 | |
openstackgerrit | Surya Prakash Singh proposed openstack/kolla-ansible: Docs update related to hostname resolution in RabbitMQ https://review.openstack.org/418252 | 11:05 |
openstackgerrit | Merged openstack/kolla: Fix telegraf contaiers https://review.openstack.org/418811 | 11:06 |
sp_ | Jeffrey4l: done | 11:06 |
Jeffrey4l | LGTM. thanks. | 11:06 |
*** sacharya has quit IRC | 11:07 | |
*** qiliang27 has quit IRC | 11:09 | |
*** qiliang27 has joined #openstack-kolla | 11:10 | |
*** shardy has joined #openstack-kolla | 11:14 | |
*** zhurong has quit IRC | 11:15 | |
openstackgerrit | Merged openstack/kolla: Cover environment for kolla https://review.openstack.org/411313 | 11:16 |
pbourke | Jeffrey4l: will tryi tout soon | 11:19 |
*** jmccarthy has joined #openstack-kolla | 11:19 | |
pbourke | *it out | 11:19 |
pbourke | Jeffrey4l: will it improve error reporting or is it more to neaten the code | 11:20 |
*** gfidente has joined #openstack-kolla | 11:20 | |
*** gfidente has joined #openstack-kolla | 11:20 | |
Jeffrey4l | pbourke, it is more to neaten the code | 11:22 |
pbourke | ok | 11:22 |
pbourke | optimise is probably not accurate in the commit message then | 11:23 |
pbourke | but its fine | 11:23 |
openstackgerrit | shaofeng cheng proposed openstack/kolla-ansible: Update README.rst in kolla-ansible https://review.openstack.org/410528 | 11:23 |
sp_ | egonzalez: pbourke incorporated Jeffrey4 comments too. would be great if you review now. https://review.openstack.org/418252 | 11:24 |
Jeffrey4l | pbourke, hrm. which world can explain this more accuracy? refactor? | 11:27 |
openstackgerrit | Merged openstack/kolla-ansible: Update kolla-ansible links on README https://review.openstack.org/418467 | 11:27 |
pbourke | Jeffrey4l: refactor is the correct term | 11:28 |
Jeffrey4l | ok. i can fix it if necessary. | 11:29 |
egonzalez | Jeffrey4l: issues in newton gates are dependent of this change? https://review.openstack.org/#/c/419417/ | 11:30 |
Jeffrey4l | egonzalez, yep. | 11:30 |
Jeffrey4l | this patch should fix the gate. | 11:30 |
egonzalez | OK, thanks | 11:31 |
openstackgerrit | Merged openstack/kolla-ansible: Docs update related to hostname resolution in RabbitMQ https://review.openstack.org/418252 | 11:31 |
*** zhubingbing has quit IRC | 11:32 | |
sp_ | Jeffrey4l: pbourke egonzalez thanks for review :) | 11:34 |
*** sayantani01 has joined #openstack-kolla | 11:35 | |
*** sayantani01 has quit IRC | 11:40 | |
*** infinitydon has joined #openstack-kolla | 11:45 | |
*** sp__ has quit IRC | 11:47 | |
*** Pavo has joined #openstack-kolla | 11:52 | |
berendt | [Openstack-operators] [kolla] deprecation/removal of Debian images: http://lists.openstack.org/pipermail/openstack-operators/2017-January/012427.html | 11:52 |
*** shardy has quit IRC | 11:55 | |
*** gfidente has quit IRC | 11:55 | |
*** shardy has joined #openstack-kolla | 11:56 | |
*** gfidente has joined #openstack-kolla | 11:56 | |
*** gfidente has joined #openstack-kolla | 11:56 | |
*** masber has quit IRC | 11:56 | |
*** saneax-_-|AFK is now known as saneax | 11:58 | |
*** gfidente has quit IRC | 12:01 | |
*** gfidente has joined #openstack-kolla | 12:02 | |
*** gfidente has quit IRC | 12:02 | |
pbourke | Jeffrey4l: im seeing passwords being output from ansible with the refactor where they werent before | 12:02 |
pbourke | Jeffrey4l: e.g http://paste.openstack.org/show/594682/ | 12:03 |
*** dave-mccowan has joined #openstack-kolla | 12:13 | |
*** msimonin has joined #openstack-kolla | 12:18 | |
*** mdnadeem has quit IRC | 12:19 | |
openstackgerrit | jianyi proposed openstack/kolla-ansible: Add zun ansible role https://review.openstack.org/416179 | 12:22 |
sdake | morning folks | 12:23 |
sdake | Jeffrey4l any update on ceph 3.0.2? | 12:23 |
sdake | egonzalez - grats dude I noticed on linked in you got a new gig at ibm :) | 12:24 |
*** sdake_ has joined #openstack-kolla | 12:27 | |
*** ipsecguy has quit IRC | 12:29 | |
*** sdake has quit IRC | 12:29 | |
*** bjolo has joined #openstack-kolla | 12:33 | |
bjolo | good afternoon | 12:33 |
bjolo | can someone please help me check out one thing. Trying to determine if it is a bug or corrupt database. please run command: openstack role assignment list --names | 12:34 |
bjolo | paste the output | 12:35 |
*** caowei has joined #openstack-kolla | 12:35 | |
*** pomac has joined #openstack-kolla | 12:37 | |
pomac | We changed, default domain id in the keystone configuration and we want to know if we have a serius bugfix or just something that went wrong with our installation | 12:38 |
pomac | (continuation of bjolo's message above) | 12:38 |
*** sdake_ has quit IRC | 12:43 | |
egonzalez | bjolo pomac http://paste.openstack.org/show/594686/ | 12:44 |
egonzalez | just glance and keystone deployed | 12:44 |
bjolo | tnx egonzalez | 12:46 |
egonzalez | bjolo: what happened? | 12:47 |
pomac | We got domain_id == None | 12:47 |
pomac | so eventually it fails with NoneType has no len() | 12:47 |
openstackgerrit | Surya Prakash Singh proposed openstack/kolla-ansible: Cover environment for kolla-ansible https://review.openstack.org/419450 | 12:47 |
pomac | egonzalez: so question is, did we miss something, or is it a real bug | 12:48 |
bjolo | history is that this used to work, but to enable v2 authentications we set default_domain_id = foobar in keystone.conf | 12:48 |
bjolo | that seems to have broken the database in some way | 12:48 |
*** sdake has joined #openstack-kolla | 12:48 | |
*** zhubingbing has joined #openstack-kolla | 12:49 | |
sdake | sup bjolo | 12:53 |
bjolo | hi sdake | 12:54 |
bjolo | hows it going | 12:54 |
sdake | busy | 12:54 |
bjolo | up late or early? | 12:54 |
sdake | 6am | 12:54 |
sdake | always up by 5:30-6am | 12:55 |
sdake | went to bed around 9 or 10pm | 12:55 |
sdake | so up at a normal time I think is the correct answer :0 | 12:55 |
sdake | although it may be early for others | 12:55 |
*** schwicht has joined #openstack-kolla | 12:55 | |
sdake | i have two kids in school - and my daughter starts at 7:30 | 12:55 |
sdake | so need to wake up by 6 to get the house rolling | 12:56 |
bjolo | pretty much the same for me | 12:58 |
*** dgonzalez has quit IRC | 12:58 | |
bjolo | egonzalez, can i ask you to run one more thing? | 12:58 |
egonzalez | bjolo: yes | 12:58 |
bjolo | mysql | 12:58 |
bjolo | keystone db | 12:58 |
*** liyifeng has quit IRC | 12:58 | |
bjolo | select * from project; | 12:58 |
pomac | egonzalez: basically, we don't know where this NULL/None value comes form | 12:59 |
*** mdnadeem has joined #openstack-kolla | 13:00 | |
egonzalez | http://paste.openstack.org/show/594690/ | 13:01 |
*** sacharya has joined #openstack-kolla | 13:03 | |
*** salv-orlando has joined #openstack-kolla | 13:04 | |
*** dgonzalez has joined #openstack-kolla | 13:05 | |
*** sacharya has quit IRC | 13:08 | |
mliima | good morning awesome guys | 13:08 |
mliima | and girls | 13:08 |
bjolo | egonzalez, tnx. looks the same as ours tho | 13:08 |
mliima | ;) | 13:08 |
*** yangzb09 has joined #openstack-kolla | 13:09 | |
*** mdnadeem has quit IRC | 13:09 | |
*** gfidente has joined #openstack-kolla | 13:11 | |
*** mliima_ has joined #openstack-kolla | 13:14 | |
openstackgerrit | Dan Prince proposed openstack/kolla: Add Apache packages to nova-api container https://review.openstack.org/418131 | 13:17 |
*** mliima has quit IRC | 13:17 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Glance service orchestration with entrypoint https://review.openstack.org/410736 | 13:18 |
*** caowei has quit IRC | 13:21 | |
egonzalez | bjolo pomac all my keystone db http://paste.openstack.org/show/594694/ | 13:21 |
openstackgerrit | Dan Prince proposed openstack/kolla: Add Apache packages to cinder-api container https://review.openstack.org/419076 | 13:22 |
egonzalez | services enabled glance heat and keystone only | 13:22 |
openstackgerrit | Dan Prince proposed openstack/kolla: Add Apache packages to nova-api container https://review.openstack.org/418131 | 13:22 |
*** ipsecguy has joined #openstack-kolla | 13:23 | |
pomac | egonzalez: thanks, have you ever been looking at the keystone code? =) | 13:23 |
egonzalez | took a look, keystone removed part of v2.0 code as part of deprecation | 13:24 |
pomac | egonzalez: i think i need to have someone explain parts of it to me - the domain_id bit that we're having problems with is actually explicitly set to "None" | 13:25 |
openstackgerrit | Dan Prince proposed openstack/kolla: Add Apache packages to nova-api container https://review.openstack.org/418131 | 13:26 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 13:26 |
openstackgerrit | Dan Prince proposed openstack/kolla: Add Apache packages to cinder-api container https://review.openstack.org/419076 | 13:26 |
pomac | egonzalez: keystone/resource/backends/base.py, get_project (line 564) ends with calling get_project_from_domain --- And that is funny! since it actually sets domain_id to None and this is before we want to use it in our lookup | 13:27 |
pomac | egonzalez: if you could play with adding domains and see if it starts to fail, or just tell me that my reading of the code is wrong - hold on, will provide a stacktrace | 13:28 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 13:28 |
pomac | egonzalez: http://paste.openstack.org/show/594697/ | 13:29 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Glance service orchestration with entrypoint https://review.openstack.org/410736 | 13:30 |
*** schwicht has quit IRC | 13:30 | |
pomac | egonzalez: if you don't have the time/energy then don't worry, will be pushing a real bug on it as well - we're deploying a new environment to test with currently | 13:32 |
*** yangzb09 has quit IRC | 13:32 | |
sdake | sup mliima_ | 13:38 |
*** salv-orlando has quit IRC | 13:39 | |
sdake | say egonzalez - I think you missed my last msg | 13:39 |
sdake | egonzalez - grants on the new gig at IBM :) | 13:39 |
openstackgerrit | Merged openstack/kolla: Bump openstack_release to 3.0.3 in group_vars https://review.openstack.org/419417 | 13:41 |
openstackgerrit | Merged openstack/kolla: Updated from global requirements https://review.openstack.org/419341 | 13:41 |
egonzalez | sdake: thanks, was at lunch time then | 13:42 |
egonzalez | also moved to Ireland (someday i'll learn english) | 13:42 |
sdake | egonzalez I recall when you started on kolla you weren't really allowed to work on openstack during work hours | 13:42 |
sdake | has that changed? | 13:42 |
sdake | egonzalez - you seem to know english to me? or you mean spoken english | 13:42 |
egonzalez | spoken and better gramar | 13:43 |
sdake | egonzalez my spanish is terrible :) | 13:43 |
sdake | egonzalez even living in AZ on the border with mexico, I struggle to compose basic 2 word statements :) | 13:43 |
sdake | your english is better then my spanish :) | 13:44 |
egonzalez | yeah, In my previous company i wasn't able to work, now I can do when Im not too busy | 13:44 |
mliima_ | improve my english is one goal for this year | 13:46 |
mliima_ | :) | 13:46 |
pomac | egonzalez: heh, have fun in Ireland, I really like it everytime i go there to visit family | 13:47 |
*** srwilkers has joined #openstack-kolla | 13:48 | |
*** absubram has joined #openstack-kolla | 13:51 | |
egonzalez | zhubingbing: were you able to fix mariadb logs? | 13:53 |
zhubingbing | np | 13:54 |
*** schwicht has joined #openstack-kolla | 13:54 | |
zhubingbing | egonzalez i fix it now, wait | 13:54 |
openstackgerrit | Merged openstack/kolla: Add Apache packages to ironic-pxe container https://review.openstack.org/419092 | 13:54 |
sdake | egonzalez i have been to dublin before - nice city | 13:55 |
sdake | Jeffrey4l - any word on the ceph gate problem? | 13:55 |
sdake | I am taking pto today and tomorrow and wanted to get kolla-kubernetes 0.4.0 sorted | 13:56 |
sdake | for tagging | 13:56 |
sdake | we need a ceph gate for kolla-ansible as well - so that we catch these problems up front | 13:56 |
sdake | (if anyone looking for some challenging work to do beyond the 190 bugs in the kolla bug tracker) | 13:56 |
egonzalez | yeah, the only thing I miss from spain is food | 13:57 |
sdake | i have been to barcelona | 13:57 |
sdake | very nice | 13:57 |
sdake | i could definately have a condo there someday if I could afford it ;) | 13:57 |
sdake | and if they permit ownership from other countries | 13:57 |
*** lamt has joined #openstack-kolla | 13:58 | |
sdake | what was crazy about ireland is everyone there doesn't steal and stuff | 13:58 |
sdake | I put my bags in a truck | 13:58 |
sdake | rather in a bus | 13:58 |
sdake | and like 15 or 20 people got off | 13:58 |
sdake | and I was worried my bags would be stolen | 13:58 |
sdake | and when I got done, the bags were still there | 13:58 |
sdake | that isn't something that would work in the US | 13:58 |
sdake | the 15-20 people got off at different stops | 13:59 |
sdake | real eye opener about the irish culture | 13:59 |
*** v1k0d3n has joined #openstack-kolla | 13:59 | |
*** absubram has quit IRC | 14:00 | |
*** absubram has joined #openstack-kolla | 14:00 | |
*** sdake_ has joined #openstack-kolla | 14:01 | |
*** schwicht has quit IRC | 14:02 | |
srwilkers | yeah, i could see living in barcelona -- really enjoyed it | 14:03 |
*** rhallisey has joined #openstack-kolla | 14:03 | |
*** schwicht has joined #openstack-kolla | 14:04 | |
*** sdake has quit IRC | 14:05 | |
*** shardy has quit IRC | 14:05 | |
*** shardy has joined #openstack-kolla | 14:06 | |
*** srwilkers is now known as srwilkers_coffee | 14:09 | |
*** saneax is now known as saneax-_-|AFK | 14:16 | |
pbourke | sdake_: they could very easily be stolen ;) | 14:24 |
pbourke | sdake_: i would be careful about that next time | 14:24 |
pbourke | its not too likely but about as likely as the US I imagine | 14:25 |
*** schwicht has quit IRC | 14:25 | |
Jeffrey4l | sdake_, i tested kolla 3.0.2 docker images with ceph enabled, it works. | 14:26 |
Jeffrey4l | but no idea why kolla-k8s failed. | 14:26 |
*** satyar has joined #openstack-kolla | 14:30 | |
*** jtriley has joined #openstack-kolla | 14:35 | |
*** schwicht has joined #openstack-kolla | 14:38 | |
*** salv-orlando has joined #openstack-kolla | 14:39 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move mongodb precheck into its own role https://review.openstack.org/416060 | 14:41 |
*** salv-orlando has quit IRC | 14:41 | |
*** salv-orlando has joined #openstack-kolla | 14:41 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move rabbitmq precheck into its own role https://review.openstack.org/417731 | 14:43 |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move kuryr precheck into its own role https://review.openstack.org/419022 | 14:46 |
*** srwilkers_coffee is now known as srwilkers | 14:46 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move Influxdb precheck into its own role https://review.openstack.org/418875 | 14:54 |
*** yingjun has joined #openstack-kolla | 14:56 | |
*** lrensing has joined #openstack-kolla | 14:57 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Glance service orchestration with entrypoint https://review.openstack.org/410736 | 14:59 |
sbezverk | Jeffrey4l: which failure you are refering to? Do you have any logs/links? | 15:00 |
Jeffrey4l | sbezverk, this one https://review.openstack.org/#/c/419082/ | 15:01 |
egonzalez | Jeffrey4l: remeber libvirt 2.0 issue in gates with xen nodes? Not only affect xen, i'm facing the same with qemu/kvm | 15:01 |
Jeffrey4l | seem k8s do not work with kolla 3.0.2 | 15:01 |
Jeffrey4l | egonzalez, yep. this issue happen in different cause. | 15:01 |
*** lamt has quit IRC | 15:02 | |
Jeffrey4l | i saw this issue when running qemu in kvm too. | 15:02 |
egonzalez | though was only with xen | 15:02 |
Jeffrey4l | it seem related to how qemu ( without kvm ) emulate the cpu. | 15:03 |
*** lamt has joined #openstack-kolla | 15:03 | |
*** sacharya has joined #openstack-kolla | 15:04 | |
sbezverk | Jeffrey4l: that must be related to packaging for ceph | 15:04 |
Jeffrey4l | sbezverk, i tested 3.0.2 with kolla-ansible. it works. | 15:04 |
sbezverk | http://tracker.ceph.com/issues/16289 | 15:05 |
openstackgerrit | OpenStack Proposal Bot proposed openstack/kolla-ansible: Updated from global requirements https://review.openstack.org/419342 | 15:05 |
sbezverk | Jeffrey4l: I am not say it is the image problem, I know kfox1111 did some stuff for ceph client in the container used at the gate.. | 15:06 |
Jeffrey4l | sbezverk, seems need add some parameter when calling `create command` | 15:06 |
Jeffrey4l | ok. sound should make some change in the kolla-k8s gate script. | 15:07 |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Add elasticsearch precheck into its own role https://review.openstack.org/419507 | 15:07 |
*** absubram has quit IRC | 15:08 | |
*** lrensing has quit IRC | 15:08 | |
*** sacharya has quit IRC | 15:09 | |
*** lrensing has joined #openstack-kolla | 15:09 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move elasticsearch precheck into its own role https://review.openstack.org/419507 | 15:09 |
*** adrian_otto has joined #openstack-kolla | 15:09 | |
*** yingjun has quit IRC | 15:11 | |
*** yingjun has joined #openstack-kolla | 15:11 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Mariadb service deployment https://review.openstack.org/412237 | 15:16 |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move rally precheck into its own role https://review.openstack.org/419513 | 15:19 |
sdake_ | sbezverk right - isaid it was a client problem | 15:19 |
sdake_ | sbezverk - but I could be wrong as I don't know precisely how the k8s gate woroks - just a curosory understanding of it | 15:19 |
sdake_ | sbezverk this is why i asked Jeffrey4l to test kolla-kubernetes with ceph - to validate 3.0.2 wasn't DOA | 15:19 |
sdake_ | the i18n problem can be easily fixed | 15:20 |
sdake_ | a ceph problem - harder to fix | 15:20 |
sdake_ | btw folks I am on PTO today and tomororw but around a bit | 15:20 |
sdake_ | if you ping me I may respond | 15:20 |
openstackgerrit | Edmund Rhudy proposed openstack/kolla: Enable no-Internet build for most Ubuntu binary containers https://review.openstack.org/414639 | 15:21 |
Pavo | anyone know why ceph-osd would exit with Container exited with non-zero return code | 15:22 |
Pavo | I am getting errors during TASK [ceph : Bootstrapping Ceph OSDs] | 15:23 |
*** yingjun has quit IRC | 15:23 | |
Pavo | I have labeled all my disk for ceph | 15:24 |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move telegraf precheck into its own role https://review.openstack.org/419520 | 15:24 |
Pavo | using parted /dev/sdb -s -- mklabel gpt mkpart KOLLA_CEPH_OSD_BOOTSTRAP 1 -1 | 15:24 |
Pavo | but during deployment I am getting failed: [compute2] (item=(0, {u'fs_uuid': u'', u'journal_device': u'/dev/sdc', u'journal': u'/dev/sdc2', u'partition': u'/dev/sdc1', u'partition_num': u'1', u'journal_num': 2, u'fs_label': u'', u'device': u'/dev/sdc', u'external_journal': False})) => {"changed": true, "failed": true, "item": [0, {"device": "/dev/sdc", "external_journal": false, "fs_label": "", "fs_uuid": "", "journal": | 15:24 |
Pavo | "/dev/sdc2", "journal_device": "/dev/sdc", "journal_num": 2, "partition": "/dev/sdc1", "partition_num": "1"}], "msg": "Container exited with non-zero return code"} | 15:24 |
Pavo | only on compute nodes | 15:24 |
Pavo | anyone have a suggestion? | 15:25 |
*** adrian_otto has quit IRC | 15:27 | |
*** adrian_otto1 has joined #openstack-kolla | 15:27 | |
*** yingjun has joined #openstack-kolla | 15:29 | |
*** yingjun has joined #openstack-kolla | 15:29 | |
*** adrian_otto1 has quit IRC | 15:30 | |
*** mgiles has joined #openstack-kolla | 15:32 | |
*** infinitydon has quit IRC | 15:32 | |
*** inc0 has joined #openstack-kolla | 15:32 | |
inc0 | good morning | 15:32 |
Pavo | inc0 morning | 15:33 |
inc0 | erhudy, I saw your email, I'll look into it | 15:34 |
erhudy | inc0: thanks, let me know what i can do to help | 15:34 |
Pavo | any reason why bootstrap ceph osd wouldn't be able to partition drives that are labeled during deployment? | 15:34 |
inc0 | also you and Pavo should talk:) both of you are deploying hyper security sensitive OS | 15:34 |
inc0 | Pavo, it does partition them | 15:34 |
Pavo | it failing for me | 15:35 |
inc0 | hmm | 15:35 |
Pavo | only on my computes though | 15:35 |
erhudy | do you have leftover ceph structures on the drives from before? | 15:35 |
inc0 | interesting, with what error? | 15:35 |
Pavo | I did a fdisk and wiped them all out prior | 15:35 |
Pavo | but during deployment I am getting failed: [compute2] (item=(0, {u'fs_uuid': u'', u'journal_device': u'/dev/sdc', u'journal': u'/dev/sdc2', u'partition': u'/dev/sdc1', u'partition_num': u'1', u'journal_num': 2, u'fs_label': u'', u'device': u'/dev/sdc', u'external_journal': False})) => {"changed": true, "failed": true, "item": [0, {"device": "/dev/sdc", "external_journal": false, "fs_label": "", "fs_uuid": "", "journal": | 15:35 |
Pavo | "/dev/sdc2", "journal_device": "/dev/sdc", "journal_num": 2, "partition": "/dev/sdc1", "partition_num": "1"}], "msg": "Container exited with non-zero return code"} | 15:35 |
erhudy | i've had problems with ceph trying to re-deploy to drives that were previously used for ceph, sgdisk -Z to destroy the GPT structures and a reboot so the kernel would forget the partition table was necessary | 15:35 |
Pavo | ok will try that | 15:36 |
Pavo | only thing I have changed now using btrfs instead of lvm for host OS | 15:36 |
Pavo | whats the cmd erhudy? | 15:37 |
erhudy | a bold move | 15:37 |
erhudy | sgdisk -Z /dev/XXXX will destroy GPT structures on that disk | 15:37 |
Pavo | I don't have sgdisk in CentOS | 15:37 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla-ansible: Refactor bootstrap.yml file https://review.openstack.org/415504 | 15:38 |
inc0 | erhudy, so immediate fix to your problem would be using template override mechanism | 15:38 |
*** jrich523 has quit IRC | 15:40 | |
inc0 | http://docs.openstack.org/developer/kolla/image-building.html#dockerfile-customisation | 15:40 |
erhudy | Pavo: try gdisk package, it should come wit that | 15:40 |
Pavo | trying to install it now | 15:40 |
Pavo | wouldn't fdisk do the same thing though? | 15:41 |
erhudy | fdisk probably only understands MBR | 15:41 |
*** sacharya has joined #openstack-kolla | 15:41 | |
erhudy | inc0: there presently is not an override block in the base image dockerfile for what i care about, are you suggesting that your preference is that i rework the PR to add blocks for any sections where they don't already exist | 15:42 |
erhudy | there are still some places where kolla itself would need modified in order to copy additional things into the build context | 15:42 |
inc0 | you can use base_apt_packages block | 15:42 |
inc0 | http://git.openstack.org/cgit/openstack/kolla/tree/docker/base/Dockerfile.j2#n259 | 15:42 |
erhudy | one other part of the PR re-works how apt keys are set up so that i can just give a directory of keys and then they're all copied into the context, instead of trying to reach an external keyserver | 15:42 |
inc0 | sorry, block name is base_ubuntu_package_installation | 15:43 |
erhudy | okay, i wasn't sure if i had added that block or if it was upstream in master | 15:43 |
inc0 | you can extend this, write your code, modify sources.list/keys and do {{ super }} | 15:43 |
inc0 | but adding block in sources list is not bad as well | 15:43 |
inc0 | we can do as many blocks as it's reasonable | 15:44 |
inc0 | but regardless, http://docs.openstack.org/developer/kolla/image-building.html#ubuntu <- this should work | 15:44 |
erhudy | also the base_ubuntu_package_installation doesn't cover the section above where it still tries to use canonical's mirrors to add a few necessary packages | 15:44 |
inc0 | if it isn't it's a bug | 15:44 |
inc0 | I'll look into it | 15:44 |
erhudy | specifically the "apt-get -y install --no-install-recommends apt-transport-https ca-certificates" part | 15:44 |
erhudy | so, yeah, you can see how this quickly turns into some yakshaving | 15:45 |
inc0 | base_header will do, you'd need to override both | 15:45 |
Pavo | inc0 why would I get https://www.dropbox.com/s/62twu7a8nwxc4so/Screenshot%202017-01-12%2010.44.59.png?dl=0 only thing I have changed is now using btrfs instead of lvm | 15:45 |
inc0 | also good point, I'm not sure why we modify sources list *after* installation of ca-certs and stuff | 15:46 |
dmsimard | inc0: oh hai, been meaning to talk to you -- I wanted to know what was the status on the openstack-infra/upstream registry | 15:46 |
*** yingjun has quit IRC | 15:46 | |
inc0 | dmsimard, we don't have registry per se | 15:46 |
erhudy | i think it was done that way initially because canonical's bare image lacks a few pieces that are then required for accessing https-only repos | 15:46 |
inc0 | Pavo, check logs man:) | 15:46 |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move octavia precheck into its own role https://review.openstack.org/419531 | 15:46 |
dmsimard | inc0: so raw image hosting then? | 15:47 |
inc0 | erhudy, yeah, but that would mean someone would need to provide it over http... | 15:47 |
erhudy | yeah, which canonical does, their repos can be accessed via either | 15:47 |
inc0 | dmsimard, it's a bit more hacky, hold on | 15:47 |
inc0 | dmsimard, http://tarballs.openstack.org/kolla/images/ | 15:47 |
inc0 | this is literally tar.gz of registry | 15:47 |
inc0 | docker run registry:2 -v /tmp/somedir | 15:48 |
inc0 | then we tar.gz /tmp/somedir | 15:48 |
dmsimard | inc0: interesting, that's what you settled for ? I remember you vigorously wanting a proper registry in infra :) | 15:48 |
inc0 | settled? no. | 15:48 |
inc0 | but that's something we can do to achieve immediate result of having gates working like they should | 15:49 |
*** jmccarthy has quit IRC | 15:49 | |
inc0 | without infra changing anything | 15:49 |
erhudy | basically my PR comes out of the fact that certain places in the bloomberg infra have absolutely no internet access, and assumptions of internet access are very pervasive and hard to stamp out | 15:49 |
inc0 | we still need to provide proper registry, but that would mean figuring few things out | 15:49 |
inc0 | mirroring, auth and such | 15:50 |
*** adrian_otto has joined #openstack-kolla | 15:50 | |
inc0 | erhudy, that's not just you | 15:50 |
*** sp__ has joined #openstack-kolla | 15:50 | |
inc0 | this isnt unusual use case | 15:50 |
dmsimard | inc0: ok -- so how would I use one of those? I'm assuming I would download that, extract it and somehow point a private registry to that ? | 15:50 |
inc0 | but I think all the people who tried that uses centos so far | 15:50 |
inc0 | so ubuntu might be buggy on that front | 15:50 |
inc0 | when you run registry, run it with flag -v /var/lib/registry:/yourtempdir | 15:51 |
inc0 | dmsimard, https://github.com/openstack/kolla/blob/master/tools/setup_gate.sh#L156 | 15:51 |
dmsimard | inc0: okay, last question before I let you go :) | 15:51 |
*** adrian_otto has quit IRC | 15:51 | |
dmsimard | inc0: when are those generated ? on post merge ? on tags ? | 15:51 |
inc0 | post merge | 15:52 |
erhudy | right, so, i guess my main point is that right now the ubuntu checks rely on a particular bug in kolla to work, and since my PR is failing on that it's impeding discussion about the approach because everyone sees "oh it failed jenkins, dev broke something" | 15:52 |
inc0 | this is master | 15:52 |
*** mliima has joined #openstack-kolla | 15:52 | |
*** adrian_otto has joined #openstack-kolla | 15:52 | |
erhudy | i can sever that bugfix from the rest of the PR and present it separately so that it's not coming as part of a big package of change | 15:52 |
inc0 | yeah, thanks for bringing it up | 15:52 |
*** mliima_ has quit IRC | 15:53 | |
erhudy | because, yeah, my PR does kind of change a lot of crap, so if it helps to reduce confusion, i'll re-do just that sources.list piece as a separate review | 15:53 |
dmsimard | inc0: ok, ty | 15:53 |
erhudy | inc0: thanks for responding on the issue, in any event | 15:56 |
openstackgerrit | Merged openstack/kolla: Heat configuration param change for heat-api section https://review.openstack.org/419330 | 15:58 |
*** pcaruana has joined #openstack-kolla | 15:58 | |
*** absubram has joined #openstack-kolla | 15:59 | |
openstackgerrit | caoyuan proposed openstack/kolla-ansible: Move panko precheck into its own role https://review.openstack.org/419539 | 16:00 |
inc0 | erhudy, as I mentioned, afaik Pavo is doing same thing, but with centos (right Pavo? builidng without internet at all?) | 16:00 |
Pavo | roger | 16:01 |
Pavo | its not hard, just gotta add your repo files to /usr/share/kolla/docker/base directory and add rpm_setup_config = nameofcustomrepofile in kolla-build.conf | 16:02 |
erhudy | yeah, but we're using ubuntu, not centos/RHEL, so | 16:02 |
Pavo | that tells kolla-build to point to whatever repo you have internally | 16:02 |
erhudy | maybe we're forging the path on the ubuntu stuff | 16:03 |
inc0 | yeah, you're first that I know of tbh | 16:03 |
erhudy | RIP me then | 16:03 |
Pavo | http://docs.openstack.org/developer/kolla/image-building.html | 16:04 |
Pavo | under custom repos | 16:04 |
inc0 | erhudy, well truth be told when we deal with this one problem | 16:04 |
inc0 | which we need in any case | 16:04 |
inc0 | I don't think there will be anything potentially wrong in the future | 16:04 |
erhudy | to accommodate ubuntu/debian it needs to deal with both the sources.list bit and getting the keys in without contacting an external keyserver, unless we're willing to just install packages unauthenticated | 16:05 |
erhudy | which could be done by clobbering the apt preferences, but it's not my favorite solution | 16:06 |
inc0 | yeah, but it will get us there fastest with least fallout | 16:06 |
inc0 | apt.prefernces + sources list above RUN apt-get install ca-stuff | 16:06 |
erhudy | yeah, the problem there is the sources.list would have to come in two pieces | 16:07 |
erhudy | because if you add the full override sources.list | 16:07 |
erhudy | and then run apt-get update | 16:07 |
inc0 | ahh | 16:07 |
inc0 | it won't even update? | 16:07 |
erhudy | it will probably detonate because of HTTPS-only repos, which it can't contact because it lacks the CA certs and the apt HTTPS transport | 16:07 |
erhudy | right | 16:07 |
inc0 | crap | 16:07 |
erhudy | blame canonical for not putting them in their bare image! | 16:08 |
inc0 | I do | 16:08 |
*** goldyfruit has joined #openstack-kolla | 16:08 | |
inc0 | problem is, out of 2 solutions other than that none of them would really fix it | 16:09 |
inc0 | but, I know what would | 16:09 |
erhudy | so, if you look at what i did to the dockerfile in my PR, that's why i took the approach with a couple of embedded ifs | 16:09 |
inc0 | so how about wrapping apt-get install ca-stuff into block | 16:09 |
inc0 | and then you can do it manually somehow | 16:09 |
inc0 | even from raw package over dpkg -i | 16:10 |
erhudy | uh, let me look at the dockerfile in master to see what would need to be put into blocks | 16:10 |
inc0 | so adding new block is free | 16:10 |
inc0 | not changing anything | 16:10 |
inc0 | you can propose patch with as many blocks as you want as long as names are unique | 16:10 |
erhudy | i guess i still have questions about how i could copy additional files into the build context, because that seems like it requires kolla code changes no matter what | 16:10 |
inc0 | yeah I don't think docker understands COPY from absolute path | 16:11 |
erhudy | it doesn't | 16:11 |
erhudy | which is why i ended up adding another config option that says "here's a directory of files that will get fed to apt-key, since you can't contact an external keyserver" | 16:11 |
*** sacharya_ has joined #openstack-kolla | 16:11 | |
inc0 | so you'd need to copy deb into kolla source tree | 16:11 |
inc0 | alternatively we can do it in our build script | 16:12 |
inc0 | have additional cp during build | 16:12 |
*** jrich523 has joined #openstack-kolla | 16:12 | |
erhudy | maybe we can compromise on my PR and leave some of the extra config options in that i added, but remove some of the convoluted embedded ifs and replace them with straightforward blocks that we can override | 16:13 |
*** sayantani01 has joined #openstack-kolla | 16:13 | |
inc0 | yeah, blocks would be much cleaner | 16:13 |
inc0 | we still need to solve this issue tho | 16:13 |
erhudy | then there's not all sorts of logic being embedded in the template, but we can still add the necessary files to the build context | 16:13 |
erhudy | this issue being the bug about not reading sources.list at all, or something else? | 16:13 |
inc0 | it does read sources.list, has to - we do use mirrors in gates | 16:14 |
*** sacharya has quit IRC | 16:14 | |
erhudy | in my e-mail i linked a log file from another build that rebuilt base | 16:14 |
erhudy | it goes straight to canonical, never touches OSIC | 16:14 |
*** rbergeron has quit IRC | 16:14 | |
Pavo | inc0 this is what horizon log says http://paste.openstack.org/show/594729/ | 16:14 |
*** jmccarthy has joined #openstack-kolla | 16:15 | |
Pavo | any idea | 16:15 |
inc0 | erhudy, https://github.com/openstack/kolla/blob/2c887565d4299462a29b98210c2a7ce233e71039/kolla/image/build.py#L587 | 16:15 |
erhudy | inc0: i know, it copies sources.list into the context | 16:15 |
erhudy | but | 16:15 |
inc0 | Pavo, check keystone log | 16:15 |
erhudy | https://github.com/openstack/kolla/blob/2c887565d4299462a29b98210c2a7ce233e71039/docker/base/Dockerfile.j2#L214 | 16:15 |
erhudy | it doesn't ever get used | 16:16 |
inc0 | no, hold on | 16:16 |
inc0 | our build script does something else | 16:16 |
erhudy | it will always copy sources.list.ubuntu or sources.list.debian to /etc/apt/sources.list | 16:16 |
inc0 | ahh | 16:16 |
inc0 | yeah, good point | 16:16 |
inc0 | so THAT is a bug | 16:16 |
inc0 | we need to copy sources.list with name suffix | 16:16 |
erhudy | so, yeah, when i realized that and fixed that, now the sources.list in the gate _does_ get used, and it's not signed, and so everything falls apart | 16:16 |
inc0 | fun life | 16:17 |
inc0 | so we should remove gate mirrors for now | 16:17 |
inc0 | can you do it in your patchset? | 16:17 |
inc0 | I'll comment why | 16:17 |
inc0 | (or you can;)) | 16:17 |
erhudy | can you do it so that it doesn't just look like i'm shouting into the abyss | 16:17 |
inc0 | it's funny because literally we announced debian deprecation today | 16:18 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Use ZUUL_REFNAME to get branch info during pack registry https://review.openstack.org/419548 | 16:18 |
inc0 | ok | 16:18 |
Pavo | inc0 not seeing any errors in keystone.log | 16:18 |
inc0 | mind if I do it in your patchset? | 16:18 |
erhudy | oh, if we ignore debian, that's great, idc about debian | 16:18 |
erhudy | not at all | 16:18 |
inc0 | Jeffrey4l, ^ | 16:18 |
inc0 | did you know that our ubuntu gates never used mirrors? | 16:18 |
Pavo | inc0 only getting Warning of deprication | 16:19 |
Jeffrey4l | inc0, yep. | 16:19 |
Jeffrey4l | wait. | 16:19 |
*** sayantani01 has quit IRC | 16:20 | |
*** dag_ has joined #openstack-kolla | 16:20 | |
Jeffrey4l | inc0, , centos doesn't and ubuntu does. | 16:20 |
*** dag_ is now known as rbergeron | 16:21 | |
Jeffrey4l | could u review this https://review.openstack.org/#/c/419548/ , we are using a wrong variable when packaging registry. | 16:21 |
Jeffrey4l | inc0, ^ | 16:21 |
inc0 | Jeffrey4l, no, what I mean, custom apt soruces.list doesn't work at all | 16:21 |
Jeffrey4l | file name is wrong here http://tarballs.openstack.org/kolla/images/ | 16:21 |
*** rbergeron has quit IRC | 16:21 | |
*** rbergeron has joined #openstack-kolla | 16:21 | |
Jeffrey4l | maybe. i saw some fix in queue. but i never used ubuntu ;/ | 16:22 |
inc0 | yeah we were just discussing that | 16:22 |
erhudy | i guess i'm learning that kolla is a centos world and i'm the lone crazy person using ubuntu :v | 16:22 |
sdake_ | morning folks | 16:22 |
inc0 | erhudy, I'm ubuntu too | 16:22 |
sdake_ | looks like my calendar was off - tomorrow is dentisst torture day | 16:22 |
*** salv-orl_ has joined #openstack-kolla | 16:23 | |
Jeffrey4l | inc0, i think we can move to dockerfile customization solution for this. | 16:23 |
openstackgerrit | Eduardo Gonzalez proposed openstack/kolla: Add missing pull task for neutron-vpnaas-agent https://review.openstack.org/419550 | 16:24 |
Jeffrey4l | no need use apt_sources-list and apt_preferences parameter | 16:24 |
inc0 | Jeffrey4l, not exactly unfortunately, docker can't COPY absolute path | 16:24 |
Jeffrey4l | we can use something trick, base64 ;) | 16:25 |
erhudy | noooooooooooooo | 16:25 |
*** salv-orlando has quit IRC | 16:26 | |
Jeffrey4l | erhudy, inc0 check here, not that bad https://github.com/openstack/kolla/blob/master/tools/setup_gate.sh#L37 | 16:26 |
sdake_ | Jeffrey4l did we have any conclusion with sbezverk on the gate problem with the 3.0.2 images? | 16:27 |
Jeffrey4l | sdake_, no. may need kfox1111 to check the script. | 16:27 |
Jeffrey4l | gate scripts. | 16:27 |
Jeffrey4l | i am not familiar with that. | 16:28 |
sdake_ | Jeffrey4l roger | 16:28 |
sdake_ | Jeffrey4l so you believe the gate scripts are incorrect then? | 16:28 |
sdake_ | Jeffrey4l one q i'd have about that is why did it work with 3.0.1 images? | 16:28 |
Jeffrey4l | sdake_, either gate script is wrong or k8s use some ceph command in wrong. | 16:29 |
sdake_ | Jeffrey4l ok thanks dude for the hep with the triage | 16:29 |
Jeffrey4l | i guess ceph package is updated? | 16:29 |
bjolo | keystone and memcached | 16:29 |
sdake_ | Jeffrey4l it is possible we went from 0.94.0 to 0.94.1 | 16:29 |
sdake_ | I am uncertain on that point | 16:29 |
bjolo | any reason why kolla newton is not using memcached for tokens by default_ | 16:30 |
sdake_ | although librdb may have changed as well | 16:30 |
bjolo | ? | 16:30 |
sdake_ | bjolo iirc it had something to do with security | 16:30 |
sdake_ | bjolo SamYaple might know more about it | 16:30 |
bjolo | ok tnx | 16:30 |
Jeffrey4l | 3.0.2 is using ceph 10.2.2 | 16:30 |
sdake_ | huh? | 16:30 |
sdake_ | what is 3.0.1 using? | 16:30 |
sdake_ | let me connect and check myself | 16:31 |
sdake_ | brb | 16:31 |
Jeffrey4l | pulling the image. | 16:31 |
*** adrian_otto1 has joined #openstack-kolla | 16:32 | |
*** adrian_otto has quit IRC | 16:32 | |
*** sdake has joined #openstack-kolla | 16:33 | |
jemcevoy | Morning Everyone, I am trying to figure a way to get a smooth install to work using the vlan provider | 16:33 |
jemcevoy | that Mech422 shared with me yesterday http://www.opencloudblog.com/?p=614 ... I see a chicken and | 16:33 |
jemcevoy | egg issue with OVS. I will need a OVS bridge setup for the kolla config but the kolla install will | 16:33 |
jemcevoy | wipe the ovs database when the openvswitch_db get loaded. Is there a way to install the openvswitch containers in my prep palybooks? | 16:33 |
erhudy | inc0 Jeffrey4l: i will go back to my PR and look at where to add additional blocks to accommodate what we want to do | 16:33 |
erhudy | i can't say base64-encoding apt keys into a block override is my favorite solution of all time, but, hacks make the world go around | 16:34 |
sdake | erhudy ya - its imperct - although i htink best we have come up with atm :) | 16:34 |
Mech422 | jemcevoy: I just hacked the kolla container not to restart openvswitch | 16:34 |
sdake | erhudy that is also how the gate works (or used to work) | 16:35 |
*** sdake_ has quit IRC | 16:35 | |
*** zhubingbing has quit IRC | 16:36 | |
sdake | Jeffrey4l 3.0.1: ceph-10.2.2-0.el7.x86_64 | 16:36 |
sdake | 3.0.2: ceph-10.2.2-0.el7.x86_64 | 16:36 |
Jeffrey4l | 3.0.2 ceph-common-10.2.2-0.el7.x86_64 ok it is the same. | 16:36 |
sdake | (pulled from hub) | 16:36 |
Jeffrey4l | then no idea :( | 16:37 |
sdake | there is a rdb library | 16:37 |
sdake | tryign to find the package that is in | 16:37 |
sdake | maybe that has changed | 16:37 |
sdake | 3.0.2: librbd1-10.2.2-0.el7.x86_64 | 16:37 |
sdake | 3.0.1: librbd1-10.2.2-0.el7.x86_64 | 16:38 |
sdake | theoretically - nothing has changed :) | 16:38 |
sdake | kfox1111 are you around? | 16:38 |
sdake | or sbezverk around? | 16:38 |
Jeffrey4l | yep. weird | 16:38 |
sdake | could be a kernel changef issue | 16:38 |
sdake | i'm not sure which patchset produced that librdb error | 16:39 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: TEST NEWTON BRANCH https://review.openstack.org/417033 | 16:40 |
sbezverk | sdake: hey what is up? | 16:41 |
jemcevoy | Mech422, So where does your openvswitch data base end up in the container or on the bare metal? I will access to a VLAN on the host for the hypervisors to work with ceph | 16:41 |
Mech422 | mine was on the metal | 16:42 |
kfox1111 | morning. | 16:42 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Fix ceilometer idempotent https://review.openstack.org/386314 | 16:43 |
sbezverk | sdake: if it is about ceph issue I found this bug: http://tracker.ceph.com/issues/16289 | 16:43 |
sbezverk | kfox1111: I might hit global api issue with service level packages | 16:43 |
kfox1111 | sbezverk: should work. I was able to munge the neutron stuff in I think: https://review.openstack.org/#/c/412239/ | 16:44 |
kfox1111 | made for a much easier helm_entrypoint_general: https://review.openstack.org/#/c/412239/22/tests/bin/setup_helm_entrypint_config.sh | 16:45 |
kfox1111 | sdake: I'm going to try reverting to 2.0.2 for everything but the previous 3.0.1 supporting stuff. | 16:45 |
kfox1111 | its been a very long time since we tried 3.0.1 for everything I think. | 16:45 |
sbezverk | kfox1111: the issue is with all the same settings , glance-api does not get logger container but glance-registry does.. | 16:46 |
kfox1111 | maybe we never had ceph in the gate working with 3.0.1. don't remember. its been too long. | 16:46 |
kfox1111 | sbezverk: weird. that seems like a bug... | 16:46 |
sdake | kfox1111 sound sgood so some images (like toolbox) will be 3.0.2? | 16:46 |
sbezverk | kfox1111: http://logs.openstack.org/36/410736/29/experimental/gate-kolla-kubernetes-deploy-centos-binary-helm-entrypoint-nv/34bfc23/console.html | 16:46 |
kfox1111 | a config like https://review.openstack.org/#/c/412239/22/tests/bin/setup_helm_entrypint_config.sh should work for that. | 16:46 |
sbezverk | when you have a chance to check.. | 16:47 |
kfox1111 | sdake: yeah. I'll leave those at 3.0.2 as it fixes bugs. | 16:47 |
sbezverk | bb in 30 minutes | 16:47 |
kfox1111 | sbezverk: ok. give me a few to try and fix up the gate. | 16:47 |
kfox1111 | k | 16:47 |
sdake | kfox1111 sounds good | 16:47 |
openstackgerrit | sayantani proposed openstack/kolla: Improve Quick Start documentation https://review.openstack.org/404993 | 16:47 |
sdake | kfox1111 btw i did manage to push ubuntu and centos last night, although ubuntu was built on centos host | 16:47 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Set versions to 0.4.0 and 3.0.2 https://review.openstack.org/419082 | 16:48 |
kfox1111 | sdake: cool. thanks. | 16:48 |
sdake | kfox1111 so inc0 may want to push the proper 3.0.2 images for ubuntu as I am out of disk space on my laptop where I run vms | 16:48 |
kfox1111 | we should know in about 20 min. | 16:48 |
kfox1111 | k. | 16:48 |
sdake | kfox1111 sweet - then we are good to tag 0.40? | 16:48 |
inc0 | I think they're there | 16:48 |
kfox1111 | if you don't mind having a 2.0.2 based 0.4.0. | 16:49 |
kfox1111 | works well though. | 16:49 |
inc0 | https://hub.docker.com/r/kolla/ubuntu-binary-kolla-toolbox/tags/ <- 3.0.2 is there | 16:49 |
*** jmccarthy has quit IRC | 16:50 | |
kfox1111 | heh. their dates are off: https://hub.docker.com/r/kolla/ubuntu-binary-kubernetes-entrypoint/tags/ | 16:50 |
kfox1111 | there wasn't a 3.0.2 until yesterday. | 16:50 |
sdake | inc0 i uploaded that last night | 16:51 |
sdake | inc0 if you uploaded as well, i may have ovewritten yours | 16:51 |
sdake | inc0 so you may need to do a push again if its not too painful | 16:51 |
sdake | btw- who is uploading master images to dockerhub | 16:52 |
sdake | anyone know? | 16:52 |
sdake | it almost looks like a gate job | 16:52 |
sdake | master tags i mean | 16:52 |
*** sayantani01 has joined #openstack-kolla | 16:53 | |
sdake | everyone: there is a lag between docker push and the hub tag metadata being updated in the web ui | 16:53 |
sdake | it seems about a 1-2 hour lag | 16:53 |
sdake | inc0 so I thought you didn't upload (if you did) and uploaded as it was blocking kfox1111 | 16:53 |
sdake | inc0 hope that adds a bit of clarity | 16:54 |
Jeffrey4l | why kolla-k8s do not build images in gate ;/ | 16:54 |
kfox1111 | sdake: portdirect is uploading master images. :) | 16:54 |
kfox1111 | he's got a cron job going, so it will keep them up to date. :) | 16:55 |
sdake | kfox1111 ok - we weren't doing that in the past because they are test images - however seems fine to me | 16:55 |
kfox1111 | yeah. I think they are just intended for testing. but still useful. | 16:55 |
sdake | kfox1111 medium term once the fires are put out we should be using master images from hub rather then 3.0.z | 16:55 |
Jeffrey4l | it is super easy now to build the image you want. | 16:55 |
sdake | Jeffrey4l everyoen wants the stuff on docker hub for convience | 16:56 |
Jeffrey4l | but for master branch gate testing. it is not convience to depends on hub.docker.com | 16:56 |
portdirect | Jeffrey4l: I try to develop on my own workstation rather than just reply on the gate - for that it is easier to pull, than have to build each time | 16:57 |
portdirect | *rely | 16:57 |
openstackgerrit | sayantani proposed openstack/kolla: Improve Quick Start documentation https://review.openstack.org/404993 | 16:58 |
Jeffrey4l | i just wanna to talk the gate. kolla-ansible build the image in gate now. rather than use a hub.docker.com edition. and the issue is there is no master branch image too in docker. | 16:59 |
*** fooliouno has joined #openstack-kolla | 17:00 | |
*** sdake_ has joined #openstack-kolla | 17:02 | |
sayantani01 | Hi guys, as discussed, the documentation team wants to add the quickstart guide to the installation guides section of openstack. It will be great if I can have some eyes on https://review.openstack.org/#/c/404993/ . | 17:02 |
*** sdake has quit IRC | 17:04 | |
openstackgerrit | Dan Prince proposed openstack/kolla: Add Apache packages to cinder-api container https://review.openstack.org/419076 | 17:05 |
*** mliima_ has joined #openstack-kolla | 17:06 | |
klindgren | Hey all - why was this change abandoned? https://review.openstack.org/#/c/412231/ I didn't sense any descent in the reviews - so I am more than willing to rebase/take it over. But it seemed like it was at a pretty good state. | 17:06 |
*** unicell has quit IRC | 17:07 | |
*** wcb has joined #openstack-kolla | 17:07 | |
openstackgerrit | Dan Prince proposed openstack/kolla: Add Apache packages to nova-api container https://review.openstack.org/418131 | 17:07 |
*** mliima has quit IRC | 17:07 | |
*** williamcaban has quit IRC | 17:07 | |
*** afranc has quit IRC | 17:07 | |
*** afranc has joined #openstack-kolla | 17:10 | |
sayantani01 | Guys, this is where the kolla deployment guide will be after the doc team takes over http://docs.openstack.org/project-deploy-guide/newton/ | 17:10 |
sdake_ | klindgren let me unabandon it | 17:10 |
sdake_ | moment | 17:10 |
sdake_ | klindgren not sure why that change was abandoned - the last statement from SamYaple was that someone could take it over | 17:12 |
sdake_ | klindgren all yorus :) | 17:12 |
*** fallingsky has joined #openstack-kolla | 17:12 | |
klindgren | kk | 17:12 |
sdake_ | klindgren note Jeffrey4l I think was doing some work in this area - might sync with him | 17:12 |
sdake_ | Jeffrey4l or SamYaple may be able to add more context a to why it was abandoned | 17:13 |
Jeffrey4l | kfox1111, it is not abandoned. just sam -w to wait for feedback. | 17:13 |
Jeffrey4l | sorry, ping the wrong person. sdake_ klindgren ^^ | 17:14 |
sdake_ | Jeffrey4l it was abaonded - i just restored it | 17:14 |
Jeffrey4l | ok got it. sam abandonded it. | 17:14 |
sdake_ | Jeffrey4l sam indicated in the comments section that anyone else was free to take it over | 17:14 |
Jeffrey4l | we should continue that works. | 17:14 |
*** fooliouno has quit IRC | 17:14 | |
sdake_ | Jeffrey4l cool - looks like a green light to me klindgren :) | 17:15 |
Jeffrey4l | no idea why Sam did that ;( | 17:15 |
*** sayantan_ has joined #openstack-kolla | 17:15 | |
*** bjolo has quit IRC | 17:16 | |
klindgren | kk - I have no problem shepherding this. | 17:17 |
*** sayantani01 has quit IRC | 17:19 | |
*** bmace has quit IRC | 17:19 | |
*** bmace has joined #openstack-kolla | 17:20 | |
sdake_ | sayantan_ looks great - can't wait for that work to finish :) | 17:20 |
sayantan_ | sdake_ Thank you. Couldn't agree more :P | 17:20 |
*** portdirect is now known as portdirect_eatin | 17:21 | |
Pavo | wow firewalld was stopping from Bootstrap Ceph OSD from working | 17:22 |
Pavo | *facepalm* | 17:22 |
inc0 | sdake_, can I have your eyes on this one? https://review.openstack.org/#/c/404993/ | 17:24 |
inc0 | to help sayantan_ with deploy guide | 17:24 |
inc0 | we wanted it to be kinda similar/exactly the same | 17:24 |
sdake_ | inc0 yup i actually have it open now | 17:24 |
*** msimonin has quit IRC | 17:24 | |
sayantan_ | inc0 , sdake_ Thanks :) | 17:25 |
*** matrohon has quit IRC | 17:26 | |
*** jmccarthy has joined #openstack-kolla | 17:26 | |
kfox1111 | hmm... the hub seems to be still a bit flaky. | 17:26 |
kfox1111 | sdake_: but the 2.0.2 ceph does seem happy. | 17:27 |
mliima_ | just fix merge conflict and it'll be ready to go https://review.openstack.org/#/c/412231/ klindgren | 17:27 |
klindgren | yep - working on that now | 17:27 |
sdake_ | kfox1111 using 3.0.1 images? | 17:27 |
kfox1111 | using 2.0.2 images except where 3.0.2 is required. | 17:27 |
sdake_ | kfox1111 I looked at centos binary - and it looks to me like its the same exact ceph release from the virtualization sig | 17:27 |
sdake_ | kfox1111 oh - your jumping from 2.0.2 to 3.0.2 ? | 17:28 |
kfox1111 | sdake_: yeah. I don't know if we ever tested the ceph deployment parts of the gate with 3.0.x. | 17:28 |
sdake_ | 3.0.2 uses a completely different version of ceph (jewel) then 2.0.2 (hammer) | 17:28 |
kfox1111 | so its possible there is an issue with ceph in 3.0.x package. | 17:28 |
sdake_ | kfox1111 Jeffrey4l tested ceph multinode ansible and that works | 17:28 |
kfox1111 | yeah. | 17:28 |
kfox1111 | its just that one weird issue I think. | 17:29 |
inc0 | yeah I have jewel deployed even now | 17:29 |
kfox1111 | where it won't take the image creation request. | 17:29 |
sdake_ | my guess is from hammer to jewel, the CLI changed in some odd way | 17:29 |
inc0 | ceph changed way it uses permissions between versions | 17:29 |
sdake_ | inc0++ :) | 17:29 |
sdake_ | kfox1111 the permissions change would do it | 17:29 |
sdake_ | had forgotten about that | 17:29 |
sdake_ | kfox1111 also your probably running with selinux enabled | 17:30 |
sdake_ | kfox1111 which we do in the gate jobs | 17:30 |
kfox1111 | no. selinux is disabled | 17:30 |
kfox1111 | if we can figure out: | 17:30 |
kfox1111 | 2017-01-12 02:21:10.080105 | + kubectl exec ceph-admin -c main --namespace=kolla -- /bin/bash -c 'timeout 240s rbd create kollavolumes/mariadb --size 1024' | 17:30 |
kfox1111 | 2017-01-12 02:21:11.929763 | 2017-01-12 02:21:11.928333 7f40cc4d3d80 -1 librbd: error writing header: (38) Function not implemented | 17:30 |
*** jmccarthy has quit IRC | 17:30 | |
kfox1111 | 2017-01-12 02:21:11.933492 | rbd: create error: (38) Function not implemented | 17:30 |
sdake_ | kfox1111 oen thing to try is to disable selinux - although thats not a long term solution | 17:30 |
sdake_ | inc0 what do you make of that above? | 17:30 |
inc0 | kfox1111, you're using 2.0.2 across the board? | 17:30 |
kfox1111 | inc0: mostly. | 17:31 |
sdake_ | kfox1111 i did see that yesterday when you posted teh full log - but not sure if inc0 has seen it or SamYaple has seen it, and they are both experts in ceph | 17:31 |
inc0 | looks like librbd version conflict | 17:31 |
kfox1111 | all the other ceph stuff seems to be working. | 17:31 |
*** egonzalez has left #openstack-kolla | 17:31 | |
kfox1111 | shouldn't be. everything should be 3.0.2 in that gate. | 17:31 |
inc0 | isn't librbd also kernel module? | 17:32 |
inc0 | it changed too I think | 17:32 |
inc0 | at some point | 17:32 |
sbezverk | kfox1111: have you had a chance to check the log I pasted? | 17:33 |
kfox1111 | the kernel module hasn't changed in a long time. | 17:33 |
kfox1111 | sbezverk: not yet. sorry. trying to figure out why 3.0.2 isn't working. | 17:33 |
sbezverk | kfox1111: here is the bug with exact error we are getting: http://tracker.ceph.com/issues/16289 | 17:33 |
inc0 | what I'm saying is they used kernel module but now they push librbd lib instead | 17:33 |
sbezverk | I suspect you have done some magic with ceph, and somewhere some binaries version maybe got hardcoded.. | 17:34 |
kfox1111 | sbezverk: yeah, that looks about right. but I'm not sure how to parse it. | 17:34 |
inc0 | I'm fairly sure glance is supposed to use image layering | 17:34 |
kfox1111 | inc0: this code is creating volumes for k8s, not for glance. | 17:35 |
kfox1111 | its possible we need some flags maybe for jewel that weren't required for hammer? | 17:36 |
sbezverk | kfox1111: the thing is I use jewel in my cluster and it is wroking fine with kube | 17:36 |
sbezverk | the only diff is I am not using ceph containers you introduced | 17:37 |
inc0 | yeah...maybe you could use kolla containers for ceph?;) | 17:37 |
kfox1111 | you did the same command I did? rbd create kollavolumes/mariadb --size 1024 | 17:37 |
sbezverk | rbd create --pool volumes --image-feature layering --size 10240 mariadb | 17:39 |
inc0 | rbd create rbd/mariadb --size 1024 <- works just fine | 17:39 |
inc0 | works with layering too | 17:39 |
sbezverk | inc0: this is from local cluster | 17:39 |
sbezverk | it was/is and will be working ;-) | 17:40 |
inc0 | ubuntu@inc0-kubernetes-1:~$ rbd --version | 17:40 |
inc0 | ceph version 10.2.2 (45107e21c568dd033c2f0a3107dec8f0b0e58374) | 17:40 |
kfox1111 | so, maybe --image-feature layering is requried now? | 17:40 |
sbezverk | kfox1111: I think so | 17:40 |
inc0 | kfox1111, I tried both with and without | 17:40 |
inc0 | works just fine both ways | 17:40 |
sbezverk | I remember reading it somewhere now that you mentioend it | 17:40 |
*** mgiles has quit IRC | 17:40 | |
inc0 | it does look like client error tho | 17:40 |
sbezverk | rbd --version | 17:41 |
sbezverk | ceph version 10.2.5 | 17:41 |
inc0 | ubuntu@inc0-kubernetes-1:~$ dpkg -l | grep librbd | 17:41 |
inc0 | ii librbd1 10.2.2-0ubuntu0.16.04.2 amd64 RADOS block device client library | 17:41 |
inc0 | ii python-rbd 10.2.2-0ubuntu0.16.04.2 amd64 Python libraries for the Ceph librbd library | 17:41 |
kfox1111 | easy enough to test.. lets try it. | 17:42 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Set versions to 0.4.0 and 3.0.2 https://review.openstack.org/419082 | 17:42 |
kfox1111 | sbezverk: I dug into the ps attached to the issue you posted. it looks like it was about adding the option to specify extra flags. so I think that does imply it needs an extra flag. | 17:44 |
kfox1111 | brb | 17:45 |
sdake_ | sayantan_ - NICE WORK | 17:52 |
sdake_ | sayantan_ first time the quickstart has been rewritten (many people have tried) without totally breaking how it works ;) | 17:53 |
kfox1111 | back | 17:53 |
sayantan_ | Thanks sdake_ Will address your comments now :) | 17:53 |
sbezverk | kfox1111: have you seen it? http://paste.openstack.org/show/594749/ | 17:56 |
inc0 | sayantan_, I made few comments too, but it's almost ready | 17:58 |
sayantan_ | inc0 Thanks, working on all the comments now. :) | 17:58 |
kfox1111 | sbezverk: usuauly when there's a left over directory from switching branches. :/ | 17:59 |
kfox1111 | see if /helm/service/neutron looks incomplete. if it is, can just rm -rf it. | 18:00 |
kfox1111 | if we really wanted to fix that, we'd have to put a list of packages into the prebuild scripts. | 18:00 |
kfox1111 | rather then just doing all found dirs. | 18:00 |
kfox1111 | as switching git branches can sometimes leave some dirs around. | 18:01 |
*** msimonin has joined #openstack-kolla | 18:01 | |
sbezverk | kfox1111: let me manually clean them up, hopefully it will help.. | 18:01 |
*** Serlex has quit IRC | 18:03 | |
kfox1111 | hmm... the layering flag let it get much farther. | 18:04 |
kfox1111 | http://logs.openstack.org/82/419082/6/check/gate-kolla-kubernetes-deploy-centos-binary-external-ovs-nv/c28da21/console.html | 18:04 |
sbezverk | :-) | 18:04 |
kfox1111 | http://logs.openstack.org/82/419082/6/check/gate-kolla-kubernetes-deploy-centos-binary-external-ovs-nv/c28da21/logs/openstack/kolla-glance-api-251924186-0r8z3.txt | 18:05 |
kfox1111 | glance error creating image... | 18:05 |
*** unicell has joined #openstack-kolla | 18:07 | |
*** msimonin has quit IRC | 18:07 | |
sbezverk | kfox1111: that would mean older ceph backend binaries | 18:07 |
kfox1111 | it all should be 3.0.2 though. | 18:08 |
sbezverk | kfox1111: we should be able to see it in pod definition if it uses 3.0.2 image | 18:10 |
kfox1111 | yeah. let me look. | 18:11 |
kfox1111 | ah... I forgot not all of the ceph stuff is in microservice packages yet. its still using kollakube for it. gotta tweak versions there too. | 18:13 |
*** msimonin has joined #openstack-kolla | 18:14 | |
*** srwilkers is now known as srwilkers_aaway | 18:14 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Set versions to 0.4.0 and 3.0.2 https://review.openstack.org/419082 | 18:14 |
*** sacharya has joined #openstack-kolla | 18:17 | |
*** sacharya_ has quit IRC | 18:17 | |
*** sp__ has quit IRC | 18:17 | |
sbezverk | kfox1111: once ceph gets sorted out, appreciate if we could take a look at global api.. as glance-api and glance-registry use identical variables, at least for general things, glance-api correcty see that no logger and hapoxy is needed, but glance-registry keep adding these containers. The main difference is glance-api uses its own template but glance-registry uses common api python | 18:19 |
sbezverk | template.. | 18:19 |
kfox1111 | K. I'm waiting on zuul now, so I'll have a look now. | 18:20 |
*** msimonin has quit IRC | 18:21 | |
*** msimonin has joined #openstack-kolla | 18:21 | |
*** msimonin has quit IRC | 18:21 | |
*** msimonin has joined #openstack-kolla | 18:22 | |
*** msimonin has quit IRC | 18:22 | |
*** msimonin has joined #openstack-kolla | 18:22 | |
*** msimonin has quit IRC | 18:23 | |
*** msimonin has joined #openstack-kolla | 18:23 | |
*** msimonin has quit IRC | 18:24 | |
*** msimonin has joined #openstack-kolla | 18:24 | |
*** msimonin has quit IRC | 18:25 | |
*** msimonin has joined #openstack-kolla | 18:25 | |
*** msimonin has quit IRC | 18:25 | |
*** jmccarthy has joined #openstack-kolla | 18:29 | |
openstackgerrit | Kris Lindgren proposed openstack/kolla: use static uid/gid in images https://review.openstack.org/412231 | 18:31 |
*** portdirect_eatin is now known as portdirect | 18:32 | |
kfox1111 | sbezverk: are you working on: https://review.openstack.org/#/c/410736 ? | 18:32 |
sbezverk | kfox1111: yep | 18:32 |
*** openstackgerrit has quit IRC | 18:33 | |
kfox1111 | I'm still debugging, but tyr removing the helm_entyrpoint_glance section from https://review.openstack.org/#/c/410736/29/tests/bin/ceph_workflow_service.sh, and pulling in this: https://review.openstack.org/#/c/412239/22/tests/bin/setup_helm_entrypint_config.sh | 18:34 |
*** pcaruana has quit IRC | 18:34 | |
kfox1111 | and you can add the ceph bits in there too. | 18:34 |
kfox1111 | shouldn't have to do individual package config's anymore most of the time with the global api. | 18:35 |
kfox1111 | as for the kubernetes_entrypoint vars, I'd stick them in the service packages values.yaml. like: | 18:36 |
kfox1111 | https://review.openstack.org/#/c/412239/22/helm/service/neutron/values.yaml | 18:36 |
kfox1111 | so they get set automatically when you use that package. | 18:36 |
*** prameswar has quit IRC | 18:36 | |
sbezverk | kfox1111: ok | 18:36 |
*** jmccarthy has quit IRC | 18:37 | |
kfox1111 | ok... looks like I got rid of all of the 2.0.2 containers. and it got further. | 18:39 |
kfox1111 | its complaining about lack of flavors now. | 18:40 |
kfox1111 | does newton not prebuild flavors anymore? | 18:40 |
sbezverk | kfox1111: yeah, it comes clean | 18:40 |
*** ipsecguy has quit IRC | 18:41 | |
*** mannamne has quit IRC | 18:42 | |
kfox1111 | ok. probably just need to add the mini flavor. | 18:43 |
satyar | Hi jeffrey4l | 18:45 |
*** openstackgerrit has joined #openstack-kolla | 18:45 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Set versions to 0.4.0 and 3.0.2 https://review.openstack.org/419082 | 18:45 |
satyar | regarding https://review.openstack.org/#/c/404944/22/ansible/library/rabbitmq_hostname.py@67 i am not suspecting we will get the error | 18:46 |
sbezverk | kfox1111: ceph related parameters still needs to be added at the run time, since monitor ip changes for multinode case | 18:46 |
satyar | as we will be getting the IPs from ansible facts | 18:46 |
satyar | jeffrey4l | 18:46 |
sbezverk | kfox1111: I mean for glance pv | 18:46 |
*** fguillot has joined #openstack-kolla | 18:47 | |
kfox1111 | sbezverk: yeah. I think thats covered in: | 18:47 |
kfox1111 | https://review.openstack.org/#/c/412239/22/tests/bin/setup_helm_entrypint_config.sh | 18:47 |
*** vinkman has joined #openstack-kolla | 18:47 | |
kfox1111 | can just add the appropriate variables to the list? | 18:47 |
kfox1111 | like, I think external_vip may need to change for multinode. don't quite remember. | 18:48 |
kfox1111 | can set the ceph bits in there too. | 18:48 |
*** cliles has joined #openstack-kolla | 18:49 | |
sbezverk | kfox1111: ok will try it | 18:49 |
*** rhallisey has quit IRC | 18:52 | |
*** MasterOfBugs has joined #openstack-kolla | 18:52 | |
*** PramodJ has joined #openstack-kolla | 18:52 | |
*** rhallisey has joined #openstack-kolla | 18:54 | |
openstackgerrit | sayantani proposed openstack/kolla: Improve Quick Start documentation https://review.openstack.org/404993 | 18:56 |
openstackgerrit | Edmund Rhudy proposed openstack/kolla: Add blocks for Ubuntu apt setup and package installation https://review.openstack.org/419610 | 19:00 |
erhudy | let's try that as a slightly lower-touch version of the last PR | 19:03 |
sayantan_ | kfox1111 inc0 sdake_ srwilkers_ portdirect I have addressed all your comments on the quickstart guide :) | 19:05 |
kfox1111 | sayantan_: cool. thanks. :) | 19:05 |
sdake_ | sayantan_ i hav ea dr appointment i need to head out for in about 5 mins so won't have time to do a review | 19:05 |
sdake_ | sayantan_ I will review once I return from that + picking up daughter from school | 19:05 |
sayantan_ | It will be great if you can review it again so that we can move forward | 19:06 |
sayantan_ | Sure sdake_ no rush | 19:06 |
sdake_ | sayantan_ agreed I will do that - in 1-2 hours - my word | 19:06 |
*** sdake_ is now known as sdake | 19:06 | |
sayantan_ | Thanks sdake_ Appreciate it :) | 19:06 |
sdake | np | 19:07 |
*** vinkman has left #openstack-kolla | 19:07 | |
Pavo | https://www.dropbox.com/s/m13annrmbmpja8f/Screenshot%202017-01-12%2014.12.34.png?dl=0 | 19:12 |
Pavo | not bad at all | 19:12 |
kfox1111 | weird... | 19:17 |
inc0 | it is not bad, 144 cores per server | 19:17 |
inc0 | I've seen worse;) | 19:17 |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Set versions to 0.4.0 and 3.0.2 https://review.openstack.org/419082 | 19:17 |
kfox1111 | http://logs.openstack.org/82/419082/8/check/gate-kolla-kubernetes-deploy-centos-binary-external-ovs-nv/f4afbae/console.html private shoudl have worked, as I didn't change tenants... | 19:18 |
inc0 | erhudy, you don't technically need to wrap packages into block | 19:18 |
satyar | Hi inc0 | 19:18 |
kfox1111 | I wonder if thats another memcached issue. :/ | 19:18 |
inc0 | stuff {% set=[] %} can be modified in a different way | 19:18 |
*** shardy has quit IRC | 19:18 | |
erhudy | okay, i'll revise it once the gates are done with it, in case something else needs fixed | 19:19 |
inc0 | reviewed | 19:20 |
inc0 | one more comment besides this one | 19:20 |
erhudy | backwards compatibility in what sense? adding another layer? | 19:22 |
inc0 | no, if someone has override file for that block | 19:22 |
inc0 | that will modify key installation | 19:22 |
erhudy | ah | 19:22 |
inc0 | it will stop working or work in a different way | 19:22 |
inc0 | not sure which one is worse;) | 19:22 |
inc0 | I'm not sure if you can't do nested blocks | 19:23 |
inc0 | I think you can | 19:23 |
inc0 | that would solve it | 19:23 |
erhudy | afaik jinja will allow it | 19:24 |
inc0 | check it out, then you can create 2 nested blocks the way you did and wrap it above | 19:25 |
inc0 | or, if you don't actually need this granularity | 19:25 |
inc0 | just leave it as it is, your call | 19:25 |
*** Pavo has quit IRC | 19:26 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Glance service orchestration with entrypoint https://review.openstack.org/410736 | 19:26 |
openstackgerrit | Edmund Rhudy proposed openstack/kolla: Add blocks for Ubuntu apt setup and package installation https://review.openstack.org/419610 | 19:27 |
*** jmontleon has joined #openstack-kolla | 19:27 | |
sbezverk | kfox1111: we do not checkin values.yaml, how do you push your neutron's values.yaml so it is get used at the gate? | 19:28 |
kfox1111 | git add -f helm/servicepackage/... | 19:29 |
kfox1111 | I always do two step commits. | 19:29 |
kfox1111 | git add | 19:29 |
kfox1111 | git commit | 19:29 |
kfox1111 | that way I know whats goign into the commit. | 19:29 |
kfox1111 | but I'm paranoid. :) | 19:29 |
*** rhallisey has quit IRC | 19:30 | |
inc0 | erhudy, looks good | 19:30 |
*** rhallisey has joined #openstack-kolla | 19:30 | |
inc0 | I'll fix the gates in the meantime | 19:30 |
erhudy | radical | 19:30 |
kfox1111 | sbezverk: commented on the review. | 19:31 |
*** srwilkers_aaway is now known as srwilkers | 19:36 | |
srwilkers | afternoon o/ | 19:36 |
kfox1111 | srwilkers: hey | 19:36 |
*** satyar has quit IRC | 19:37 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Glance service orchestration with entrypoint https://review.openstack.org/410736 | 19:43 |
sbezverk | kfox1111: thanks, addressed them now keeping fingers crossed cannot type ;-) | 19:44 |
kfox1111 | yeah. that looks better. should have a fair chance of working now I think. | 19:45 |
*** mgoddard has joined #openstack-kolla | 19:46 | |
openstackgerrit | Merged openstack/kolla: Fix ceilometer idempotent https://review.openstack.org/386314 | 19:46 |
kfox1111 | closer still.... http://logs.openstack.org/82/419082/9/check/gate-kolla-kubernetes-deploy-centos-binary-external-ovs-nv/366e0ed/console.html | 19:49 |
kfox1111 | got a meeting to go to. biab. | 19:50 |
*** tonanhngo has joined #openstack-kolla | 19:54 | |
*** salv-orl_ has quit IRC | 19:54 | |
*** mgoddard_ has joined #openstack-kolla | 19:56 | |
*** mgoddard has quit IRC | 19:57 | |
*** mgoddard has joined #openstack-kolla | 19:58 | |
*** tonanhngo has quit IRC | 19:59 | |
*** mgoddard_ has quit IRC | 20:01 | |
*** salv-orlando has joined #openstack-kolla | 20:01 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: WIP memcached operator https://review.openstack.org/418073 | 20:03 |
*** ipsecguy has joined #openstack-kolla | 20:06 | |
-openstackstatus- NOTICE: Gerrit will be offline between now and 20:30 for scheduled maintenance: http://lists.openstack.org/pipermail/openstack-dev/2017-January/109910.html | 20:08 | |
*** ChanServ changes topic to "Gerrit will be offline between now and 20:30 for scheduled maintenance: http://lists.openstack.org/pipermail/openstack-dev/2017-January/109910.html" | 20:08 | |
*** sayantan_ has quit IRC | 20:09 | |
*** adrian_otto1 has quit IRC | 20:17 | |
*** schwicht has quit IRC | 20:17 | |
*** tonanhngo has joined #openstack-kolla | 20:17 | |
*** aj_power has joined #openstack-kolla | 20:17 | |
*** schwicht has joined #openstack-kolla | 20:18 | |
*** awiddersheim has joined #openstack-kolla | 20:21 | |
*** awidders_ has quit IRC | 20:21 | |
*** awidders_ has joined #openstack-kolla | 20:23 | |
*** awiddersheim has quit IRC | 20:23 | |
*** awiddersheim has joined #openstack-kolla | 20:25 | |
*** awidders_ has quit IRC | 20:25 | |
*** sayantani01 has joined #openstack-kolla | 20:27 | |
*** aj_power has quit IRC | 20:29 | |
*** awiddersheim has quit IRC | 20:31 | |
*** awiddersheim has joined #openstack-kolla | 20:32 | |
*** mannamne has joined #openstack-kolla | 20:32 | |
-openstackstatus- NOTICE: Updated: Gerrit will be offline until 20:45 for scheduled maintenance (running longer than anticipated): http://lists.openstack.org/pipermail/openstack-dev/2017-January/109910.html | 20:33 | |
*** ChanServ changes topic to "Updated: Gerrit will be offline until 20:45 for scheduled maintenance (running longer than anticipated): http://lists.openstack.org/pipermail/openstack-dev/2017-January/109910.html" | 20:33 | |
*** jrist has quit IRC | 20:33 | |
*** goldyfruit has quit IRC | 20:42 | |
*** jrist has joined #openstack-kolla | 20:47 | |
*** ChanServ changes topic to "Have support questions? Please give ask.openstack.org a try. New to Kolla? Please read the documentation here: http://docs.openstack.org/developer/kolla/; Kolla IRC meetngs on Wednesdays @ 16:00 UTC - see agenda @ https://goo.gl/OXB0DL - IRC channel is *LOGGED* @ http://goo.gl/3mzZ7b" | 20:53 | |
sdake | sbezverk are we good to go with the release? | 20:53 |
*** jrobinson has joined #openstack-kolla | 20:54 | |
*** david-lyle has quit IRC | 21:01 | |
*** david-lyle has joined #openstack-kolla | 21:02 | |
sdake | inc0 about? | 21:05 |
inc0 | about to go to lunch | 21:08 |
inc0 | wassup? | 21:08 |
*** fallingsky has quit IRC | 21:09 | |
*** adrian_otto has joined #openstack-kolla | 21:10 | |
sean-k-mooney | is there a kolla ptg etherpad by the way. | 21:15 |
inc0 | indeed there is | 21:15 |
inc0 | https://etherpad.openstack.org/p/kolla-ocata-ptg | 21:16 |
*** mgoddard has quit IRC | 21:16 | |
sean-k-mooney | i see its not useing the same nameing convention as nova https://etherpad.openstack.org/p/kolla-ptg-pike | 21:16 |
sean-k-mooney | ah ok cool thanks | 21:16 |
inc0 | yeah it's technically pike | 21:16 |
inc0 | well, next one will be R | 21:16 |
inc0 | btw..do we have R release name? | 21:17 |
sean-k-mooney | isnt the next one Q not R | 21:17 |
inc0 | ahh | 21:17 |
inc0 | well, in Polish we don't use Q, so my bad | 21:17 |
sean-k-mooney | i think Queens was proposed. | 21:18 |
*** openstackgerrit has quit IRC | 21:18 | |
sean-k-mooney | inc0: no worries if i was going by the irish alphbet we only have 18 letters instad of 26 | 21:18 |
sean-k-mooney | its by everyone get confused when irish names with bh are pernouced like v | 21:19 |
kfox1111 | back. | 21:19 |
inc0 | in Polish we have 32 letter, but V,Q,X are not in it | 21:20 |
inc0 | we have couple polish only letter which taught me a lot about encoding in software | 21:20 |
sean-k-mooney | haha ya if you go outside of the ascii range it gets more diffucult utf-8 is interesting... | 21:23 |
kfox1111 | yeah, I like linux's general, just use utf8 solution over window's, lets support 30,000 different types of data types for unicode. ;) | 21:24 |
inc0 | my woe was web browsers | 21:25 |
inc0 | any python 2.x is all ascii | 21:25 |
*** dturner has joined #openstack-kolla | 21:26 | |
*** athomas has quit IRC | 21:26 | |
sean-k-mooney | kfox1111: well windows devs generally use utf8 now too | 21:28 |
sean-k-mooney | the issue is had was how to store utf8 perperly when each charater can be 1-4 bytes and may or may not contaien a byte order mark. | 21:29 |
*** l4yerffeJ has quit IRC | 21:34 | |
*** Jeffrey4l has quit IRC | 21:34 | |
*** JoseMello has quit IRC | 21:35 | |
sean-k-mooney | by the way anyone heard of or used ansible-lint https://github.com/willthames/ansible-lint | 21:36 |
sean-k-mooney | it might be worth looking into adding for kolla-ansible | 21:37 |
*** gfidente has quit IRC | 21:39 | |
*** srwilkers has quit IRC | 21:43 | |
*** fguillot has quit IRC | 21:44 | |
*** rhallisey has quit IRC | 21:44 | |
dmsimard | sean-k-mooney: huh, interesting bit about the 18 letters alphabet .. explains a couple pronounciations indeed | 21:47 |
dmsimard | like Eoghan | 21:48 |
inc0 | yeah, his name is crazy, when someone pronounced it to me, I had no idea that he's he | 21:49 |
sean-k-mooney | we have a foda over vouels too, áéíóú but they are not considered seperate leters | 21:49 |
kfox1111 | seeing some weird errors.... http://logs.openstack.org/82/419082/9/check/gate-kolla-kubernetes-deploy-centos-binary-external-ovs-nv/f8ff995/logs/openstack/kolla-neutron-server-1133147973-nxsz0.txt | 21:49 |
sean-k-mooney | owen is the english speeling of Éoghan | 21:50 |
*** openstackgerrit has joined #openstack-kolla | 21:50 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Set versions to 0.4.0 and 3.0.2 https://review.openstack.org/419082 | 21:50 |
inc0 | now I know that | 21:50 |
*** jrobinson is now known as jrobinson-afk | 21:50 | |
inc0 | but didn't at the time | 21:50 |
openstackgerrit | sayantani proposed openstack/kolla: Improve Quick Start documentation https://review.openstack.org/404993 | 21:50 |
sean-k-mooney | ya i used to use english phontic spelling in irish. it did not work very well | 21:51 |
sean-k-mooney | espcially when i used letters that did not exist | 21:52 |
dmsimard | lol | 21:52 |
kfox1111 | heh. you know you have a big company when you read "amazon" instead of the real word in the article, "arizona" over and over... :/ | 21:54 |
sean-k-mooney | kfox1111: i have dyslexia and if my brain half reads "arizona" and desides its "amazon" it will be so until i go do something else and then come back to it | 21:56 |
sbezverk | kfox1111: any word on that global api issue I hit? I think it is still might be out there.. | 21:56 |
kfox1111 | sbezverk: let me look again. | 21:56 |
sbezverk | kfox1111: I would appreciate, probably I will not see it again with new variables in values.yaml, but I am afraid it will hit us in different place | 21:57 |
sean-k-mooney | did that post in the channel or just set my away status locally | 21:59 |
*** jtriley_ has joined #openstack-kolla | 22:03 | |
*** inc0 has quit IRC | 22:04 | |
*** jtriley has quit IRC | 22:06 | |
*** absubram has quit IRC | 22:07 | |
*** jtriley_ has quit IRC | 22:08 | |
kfox1111 | arg... the docker hub is still messed up. :/ | 22:15 |
kfox1111 | 2017-01-12 22:11:51.128735 | Tag 3.0.1 not found in repository docker.io/kolla/ubuntu-binary-memcached | 22:15 |
*** unicell has quit IRC | 22:16 | |
*** salv-orl_ has joined #openstack-kolla | 22:23 | |
sbezverk | kfox1111: are you sure we did not use 2.0.2 before? maybe 3.0.1 has not been ever pushed for ubuntu | 22:24 |
*** mannamne has quit IRC | 22:25 | |
*** unicell has joined #openstack-kolla | 22:25 | |
*** salv-orlando has quit IRC | 22:25 | |
*** msimonin has joined #openstack-kolla | 22:27 | |
dmsimard | kfox1111: I had issues with docker hub yesterday, are they still having problems ? | 22:28 |
dmsimard | I would push and things would not show up. | 22:29 |
*** mannamne has joined #openstack-kolla | 22:29 | |
kfox1111 | sbezverk: possible. | 22:33 |
kfox1111 | dmsimard: yeah, I think i've still seen some flakeyness on pulls. | 22:33 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Glance service orchestration with entrypoint https://review.openstack.org/410736 | 22:38 |
*** lrensing has quit IRC | 22:39 | |
*** goldyfruit has joined #openstack-kolla | 22:40 | |
sbezverk | kfox1111: have a second for qq? | 22:51 |
*** mannamne has quit IRC | 22:56 | |
*** mannamne has joined #openstack-kolla | 22:56 | |
dmsimard | kfox1111: oh well there you go http://status.docker.com/ | 23:01 |
sdake | dmsimard thats pretty cool - thanks for the link :) | 23:04 |
*** pramodrj07 has joined #openstack-kolla | 23:04 | |
sdake | kfox1111 this is why we need to get away from the docker hub as part of our ci infra :) | 23:04 |
sdake | kfox1111 we need to be building images in the gate itself | 23:04 |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Glance service orchestration with entrypoint https://review.openstack.org/410736 | 23:06 |
kfox1111 | sdake: I don't disagree. but building images per run is not good either. | 23:06 |
sdake | this is how kolla-ansible works | 23:07 |
sdake | its ottally fine | 23:07 |
sdake | it sometimes fials | 23:07 |
kfox1111 | sdake: it takes 20 min per gate run already. doubling that by uilding containers would really cut down on productivity. we need a cache. | 23:07 |
sdake | recheck ftw | 23:07 |
kfox1111 | sbezverk: yeah. but, do you have a link to that previous running job of yours? | 23:07 |
*** PramodJ has quit IRC | 23:07 | |
sdake | ya - 40 min gate times are what we have currently for kolla-ansible | 23:07 |
*** MasterOfBugs has quit IRC | 23:07 | |
kfox1111 | I was trying to watch it to see, but you rechecked to quick. :/ | 23:07 |
kfox1111 | sdake: I'd be much less productive witha 40 min gate. :/ | 23:07 |
sdake | kfox1111 i'd challeng ethat productivity is busted when docker hub is down | 23:07 |
sdake | kfox1111 and docker hub is super unreliable | 23:08 |
sbezverk | kfox1111: ok we will deal then when it hits again | 23:08 |
kfox1111 | sdake: yeah, but the hub's been down much less then 20 extra min * dozens of ps's updates a day. | 23:08 |
*** MasterOfBugs has joined #openstack-kolla | 23:08 | |
sbezverk | kfox1111: by some reason glance manage db job did not have init contrainer at all? any particular reason why it was removed? | 23:08 |
sdake | right - the wya i deal with gate lag is to work on multiple things at once | 23:08 |
sdake | it requires some context switching but its not too terrible | 23:08 |
kfox1111 | sdake: that, I haven't seen. its been very reliable. probably down 3 days out of all of my time working on the kolla-kubernetes gate. | 23:09 |
kfox1111 | yeah. I do that too. but then you have merge conflict overhead to deal with. | 23:09 |
sdake | the other problem is we aren't testing the latest kolla master containers ;) | 23:09 |
kfox1111 | there's a lot of tradeoffs. | 23:09 |
kfox1111 | really we need to cache 3.0.2 on tarballs.o.o. | 23:09 |
sdake | ya - merge conflict overhead is a thing but not that major | 23:09 |
kfox1111 | or a registry. | 23:09 |
kfox1111 | then we can keep the 20 min gates. | 23:09 |
sdake | inc0 is working on that in some way to speed up the gate | 23:09 |
kfox1111 | yup. | 23:09 |
kfox1111 | he's really close too. | 23:10 |
sdake | however in the short term i think we need to transition to building images on each gate run | 23:10 |
kfox1111 | procrastination will solve this problem. ;) | 23:10 |
sdake | definately before we release 1.0.0 we need to be building againset master openstack | 23:10 |
kfox1111 | I think his trunk cache is close. much better to just wait for that. | 23:10 |
sdake | that way we can select the images we use | 23:10 |
kfox1111 | those images are tested too. so more likely to be reliable. | 23:10 |
sdake | we can wait on that sure - not sure where it is implementation wise | 23:10 |
*** Pavo has joined #openstack-kolla | 23:10 | |
kfox1111 | he's got a ps half done to make kolla-kubernetes use the cache. | 23:11 |
kfox1111 | I was going to help with that once 0.4.0's done. | 23:11 |
*** mannamne has quit IRC | 23:11 | |
sdake | just pointing out that testing with newton images is less then ideal when kolla-kube will implement ocata ;) | 23:11 |
kfox1111 | we'll always be chacing a newer version. ;) | 23:11 |
kfox1111 | ops usually only care about a stable version though. ;) | 23:11 |
kfox1111 | we need to ensure both work. | 23:12 |
kfox1111 | but if you want to make your customers happy, stable's got to work well. | 23:12 |
kfox1111 | chacing trunk just lets you get to the new stable sooner. | 23:12 |
kfox1111 | as you find issues quicker and fix them incrementally rather then all at once. | 23:13 |
*** Pavo has quit IRC | 23:13 | |
kfox1111 | so.... | 23:14 |
kfox1111 | 3.0.1 everything with 3.0.2 for a few things passed the gate for the things that didn't have hub issues. | 23:15 |
kfox1111 | but 3.0.2 everything was failing. | 23:15 |
kfox1111 | well, for centos.... ubuntu seems to be missing some 3.0.1 packages. :/ | 23:15 |
kfox1111 | what a mess. | 23:16 |
*** Pavo has joined #openstack-kolla | 23:17 | |
*** dave-mccowan has quit IRC | 23:17 | |
kfox1111 | so guys, what do you want to do? | 23:18 |
kfox1111 | 2.0.2 based gate is working well. | 23:18 |
sbezverk | kfox1111: staying with 2.0.2 is not an option | 23:19 |
kfox1111 | a 3.0.1 mostly version coudl be made to work too. | 23:19 |
kfox1111 | a 3.0.2 could be made to work with more debugging and maybe some new containers. | 23:19 |
*** saneax-_-|AFK is now known as saneax | 23:19 | |
kfox1111 | its a snapshot in time release, so we can fix things in a 0.5.0 whenever we want to do that. | 23:20 |
kfox1111 | so, do we delay 0.4.0 further? | 23:20 |
sbezverk | kfox1111: personnaly I could not care less about 0.4.0, but I do care about 3.0.2 | 23:23 |
sbezverk | maybe we could get gate stable on 3.0.2 and then release 0.4.0? | 23:24 |
kfox1111 | I need 0.4.0 soonish but can probaby wait till next week. I need to start deploying it and would rather not deploy trunk. | 23:24 |
kfox1111 | I think thats probably doable in the next few days. | 23:25 |
kfox1111 | just gota figure out whats different between 3.0.1 and 3.0.2. | 23:25 |
kfox1111 | I can try pinning neutron back down to 3.0.1 and leave the rest at 3.0.2 and see if that fixes things. it only seemed to be neutron being affected. | 23:25 |
*** Pavo has quit IRC | 23:26 | |
openstackgerrit | Kevin Fox proposed openstack/kolla-kubernetes: Set versions to 0.4.0 and 3.0.2 https://review.openstack.org/419082 | 23:27 |
sbezverk | sure, then instead of fighting with everything we can deal with just neutron | 23:27 |
sbezverk | releasing 0.4.0 now I think is not right because of some instability we see with newer images/cintainers | 23:28 |
kfox1111 | yeah, thats kind of the problem with using untested containers though. you can never really know if it will work. :/ | 23:29 |
kfox1111 | thats the real power of docker. | 23:29 |
kfox1111 | you can use the exact container that you gate tested against. | 23:29 |
kfox1111 | one less thing that can go wrong. | 23:29 |
sbezverk | sure if docker hub is still alive ;-) | 23:30 |
kfox1111 | you dont' run into stuff like, "building ubuntu container on centos 7.3.2 builds broken containers" | 23:30 |
kfox1111 | true. but there are other ways around that. :) | 23:30 |
*** aj_power has joined #openstack-kolla | 23:30 | |
kfox1111 | you just make sure you have containers on the hub and quay or wherver. | 23:31 |
*** l4yerffeJ has joined #openstack-kolla | 23:31 | |
kfox1111 | replication is relatively easy. :) | 23:31 |
kfox1111 | building binary identicle containers is tough. | 23:31 |
*** Pavo has joined #openstack-kolla | 23:34 | |
*** haplo37_ has quit IRC | 23:36 | |
*** Jeffrey4l has joined #openstack-kolla | 23:37 | |
kfox1111 | sbezverk: don't push antyhing yet. | 23:38 |
kfox1111 | looking at your ps. | 23:38 |
*** haplo37_ has joined #openstack-kolla | 23:39 | |
kfox1111 | sbezverk: there is a one character misalignment in values.yaml. | 23:40 |
*** Pavo has quit IRC | 23:40 | |
kfox1111 | sbezverk: but, since all your values are set to true, | 23:41 |
kfox1111 | you can just make that values file: | 23:41 |
kfox1111 | globals: | 23:41 |
kfox1111 | kolla: | 23:41 |
kfox1111 | all: | 23:41 |
kfox1111 | kubernetes_entrypoint: true | 23:41 |
kfox1111 | the rest of the stuff will be automatic. | 23:41 |
*** Pavo has joined #openstack-kolla | 23:42 | |
*** msimonin has quit IRC | 23:42 | |
*** msimonin has joined #openstack-kolla | 23:42 | |
openstackgerrit | sayantani proposed openstack/kolla-ansible: Modify config.yml playbook for bifrost role https://review.openstack.org/419196 | 23:43 |
*** msimonin has quit IRC | 23:43 | |
*** msimonin has joined #openstack-kolla | 23:43 | |
*** msimonin has quit IRC | 23:44 | |
*** msimonin has joined #openstack-kolla | 23:44 | |
*** schwicht has quit IRC | 23:44 | |
*** msimonin has quit IRC | 23:44 | |
*** msimonin has joined #openstack-kolla | 23:45 | |
*** msimonin has quit IRC | 23:45 | |
*** msimonin has joined #openstack-kolla | 23:45 | |
sayantani01 | sdake_ I made changes based on your comment. It will be great if you can take a last look (hopefully last :P) | 23:46 |
*** msimonin has quit IRC | 23:46 | |
*** msimonin has joined #openstack-kolla | 23:46 | |
openstackgerrit | Serguei Bezverkhi proposed openstack/kolla-kubernetes: Glance service orchestration with entrypoint https://review.openstack.org/410736 | 23:47 |
*** msimonin has quit IRC | 23:47 | |
sbezverk | kfox1111: I will change it as last step | 23:48 |
sbezverk | it is still failing.. but it was my error | 23:48 |
*** salv-orl_ has quit IRC | 23:48 | |
sdake | sayantani01 sounds good | 23:48 |
sayantani01 | Thanks sdake :) | 23:50 |
kfox1111 | Tag 3.0.1 not found in repository docker.io/kolla/ubuntu-binary-neutron-server really? | 23:50 |
kfox1111 | did ubuntu just not get a 3.0.1 release of a bunch of things? | 23:50 |
*** sayantani01 has quit IRC | 23:50 | |
sbezverk | kfox1111: that what I was trying to tell you :-) 3.0.1 ubuntu binary has not been pushed | 23:52 |
sbezverk | just some bits anf pieces | 23:52 |
sbezverk | what if you use 3.0.2 tag? instead of 3.0.1 | 23:53 |
kfox1111 | sorry. I didn't quite get that. :/ | 23:53 |
* kfox1111 sighs. | 23:53 | |
kfox1111 | well, at least for centos, 3.0.2 doesn't seem to work. | 23:53 |
kfox1111 | still digging into why. | 23:53 |
kfox1111 | I was suspecting neutron, so I pinned that down to 3.0.1, but ubuntu's choking on it. | 23:53 |
sbezverk | kfox1111: how about pinning neutron back to 2.0.2? | 23:54 |
sbezverk | maybe doing that we could identify all failing cases and bring at least majority to 3.0.2 | 23:55 |
Pavo | need me to build 3.0.2 and try it for centos kfox1111? | 23:55 |
*** lamt has quit IRC | 23:55 | |
kfox1111 | may be an option. going to see what falls out of 3.0.1 first. maybe it will tell us whats wrong with the 3.0.2 packages if anything. | 23:55 |
*** schwicht has joined #openstack-kolla | 23:56 | |
Pavo | how are you see what fails? | 23:56 |
kfox1111 | Pavo: I'm watching zuul. | 23:57 |
kfox1111 | status.openstack.org/zuul/ | 23:57 |
Pavo | because my 3.0.1 works pretty good minus murano and magnum horizon plugins | 23:57 |
Pavo | ah | 23:57 |
kfox1111 | ps 419082 | 23:57 |
Pavo | its this k8s? | 23:58 |
kfox1111 | yeah. | 23:58 |
kfox1111 | k8s gate. | 23:58 |
Pavo | nice | 23:58 |
kfox1111 | the problem is, its hard to know if your 3.0.1 is closer to the hub's 3.0.1 or the hub's 3.0.2, because of when it was built. :/ | 23:58 |
Pavo | makes sense | 23:59 |
kfox1111 | ok. the good news, the global api is working very nicely. :) | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!