*** daneyon has quit IRC | 00:03 | |
*** daneyon_ has quit IRC | 00:09 | |
*** dave-mccowan has joined #kolla | 00:11 | |
sdake | SamYaple is there some reason /dev is bindmounted in ansible? | 00:29 |
---|---|---|
sdake | in kolla_ansible | 00:29 |
*** neilus has joined #kolla | 00:32 | |
SamYaple | sdake: yup | 00:36 |
sdake | well with master ansible it doesn't work | 00:36 |
SamYaple | youll have to be more specificate | 00:36 |
sdake | see review - there is some debugging in it | 00:36 |
sdake | you will understand when you look in review | 00:37 |
sdake | https://review.openstack.org/#/c/276328/1 | 00:37 |
SamYaple | I don't understand your question | 00:37 |
sdake | read the pastes please | 00:37 |
sdake | it will take you 5 minutes | 00:37 |
sdake | one with dev bindmounted | 00:38 |
sdake | one without | 00:38 |
sdake | with that patch applied | 00:38 |
SamYaple | I have already read the review | 00:38 |
sdake | ok, so you see how ansible fails with dev bindmounted in the first comment? | 00:38 |
SamYaple | i see your paste, i dont see that behaviour, no | 00:39 |
sdake | the err 13 | 00:39 |
*** dims_ has joined #kolla | 00:39 | |
sdake | permission denied comes from multiprocessing library | 00:39 |
sdake | because of some wierd interaction with not being able to read stuff in /dev/shm | 00:39 |
sdake | so the patch as is doesn't work with a dev bindmount | 00:40 |
sdake | that means new versions of ansible wont work with a dev bindmount, atleast not docker 1.8.2 | 00:40 |
*** dims has quit IRC | 00:40 | |
SamYaple | no it means you need to fix your host /dev/shm | 00:40 |
sdake | the second paste removes the dev bindmount | 00:40 |
sdake | and the original problem still exists | 00:40 |
SamYaple | dev is required | 00:40 |
SamYaple | and i cannot reproduce the stack in the bug | 00:40 |
sdake | cool i am up to fixing my /dev/shm | 00:40 |
sdake | we can document that | 00:41 |
SamYaple | so youll have to take over this bug as it is centos only | 00:41 |
sdake | what precisely needs fixing | 00:41 |
*** neilus has quit IRC | 00:41 | |
SamYaple | youll have to track that down sdake | 00:41 |
SamYaple | its on you | 00:41 |
sdake | second paste | 00:41 |
sdake | shows the ordering doesn't work | 00:41 |
SamYaple | doesnt happen on ubuntu | 00:41 |
sdake | on stable/liberty? | 00:41 |
sdake | did you try it | 00:41 |
SamYaple | yup | 00:41 |
SamYaple | yup | 00:41 |
SamYaple | so did inc0 | 00:42 |
SamYaple | no issues | 00:42 |
sdake | did he report that in channel today? | 00:42 |
SamYaple | yup | 00:43 |
sdake | please educate why /dev needs bindmounting | 00:43 |
SamYaple | because we containerized the find_disk module | 00:43 |
sdake | thanks | 00:44 |
*** aginwala has quit IRC | 00:44 | |
sdake | samyaple http://eavesdrop.openstack.org/irclogs/%23kolla/%23kolla.2016-02-04.log.html#t2016-02-04T21:43:17 | 00:52 |
*** farid has joined #kolla | 00:52 | |
sdake | my intpretation of that log is that inc0 did not test with haproxy, and as a result was not testing the kolla-ansible change you made | 00:52 |
sdake | and hence, didn't hit the ordering error | 00:52 |
SamYaple | but thats not how kolla works | 00:53 |
sdake | i understand kolla ansible gets calleda a million times | 00:53 |
sdake | but he didn't say your patch worked | 00:53 |
sdake | he said it worked dwithout your patch | 00:53 |
SamYaple | im not claiming we need my patch | 00:53 |
sdake | your patch doesn't work | 00:54 |
SamYaple | it does on ubuntu | 00:54 |
SamYaple | but ubuntu isnt the one with either of these issues | 00:54 |
sdake | with ubuntu you build everything from source | 00:54 |
sdake | so thats a good data point | 00:54 |
SamYaple | so ignore my patch and figure out why centos is busted would be my suggestion | 00:54 |
sdake | yup | 00:54 |
sdake | i am going to build source centos | 00:54 |
*** salv-orl_ has quit IRC | 00:54 | |
sdake | if that works, then my speculation would be two packages installed at once | 00:55 |
sdake | pip and rpms | 00:55 |
* sdake hates computers | 00:55 | |
SamYaple | which is what mandre stated he just never finished his patch | 00:55 |
sdake | did he write that in the bug log? | 00:56 |
sdake | or irc | 00:56 |
mandre | hey | 00:56 |
sdake | hey mandre | 00:56 |
SamYaple | both | 00:56 |
*** farid has quit IRC | 00:57 | |
mandre | i see you're talking about mysql failing in kolla-ansible container | 00:57 |
sdake | mandre working on https://bugs.launchpad.net/kolla/liberty/+bug/1530962 | 00:57 |
openstack | Launchpad bug 1530962 in kolla liberty "six version dependency in kolla_ansible appears to be incorrect." [Critical,Confirmed] - Assigned to Sam Yaple (s8m) | 00:57 |
sdake | well i dont have my lab gear so i had to build a vm which took all damn day | 00:58 |
sdake | but yup i'm good to go now :) | 00:58 |
SamYaple | mandre: yea i cant reproduce anywhere so im going to step out of this conversation | 00:58 |
sdake | mandre what do you think the problem is | 00:58 |
sdake | thanks for tyring SamYaple | 00:58 |
sdake | trying | 00:58 |
mandre | yeah, it's been on my todo list for some time, just never got around to finish it | 00:58 |
mandre | I can reproduce, I was planning to tackle the issue today actually | 00:59 |
*** farid has joined #kolla | 00:59 | |
sdake | good lets both work on it | 00:59 |
sdake | because stable is broken | 00:59 |
sdake | and that needs to be fixed asap | 00:59 |
sdake | sound like a plan? | 01:00 |
mandre | not sure, what the issue is sdake, but I suspect version mismatch] | 01:00 |
sdake | mandre i think its pip and rpm package installed at same time | 01:00 |
sdake | speculation | 01:00 |
sdake | i am goign to build centos source now | 01:00 |
mandre | sdake, it's only broken for centos-binary | 01:00 |
mandre | centos source works fine | 01:00 |
sdake | well there ya go | 01:00 |
sdake | thanks for that info that saved a couple hours | 01:01 |
mandre | i'll submit a proper fix today hopefully | 01:01 |
*** Allen_Gao has joined #kolla | 01:02 | |
sdake | how are you thinking of fixing it | 01:02 |
mandre | not sure yet | 01:03 |
mandre | i'll find something | 01:03 |
mandre | actually, ansible is totally busted in that container | 01:03 |
mandre | 'ansible --version' also generates a traceback | 01:04 |
sdake | lets work on this together mandre | 01:04 |
sdake | i need this quickly as well | 01:04 |
sdake | what other data do you have | 01:04 |
mandre | pip shows it's ansible 2.0.0.2, and six 1.9.0 | 01:05 |
sdake | my experience with six errors is its actually a different dependency | 01:06 |
sdake | or two versions of six installed at once | 01:06 |
sdake | 2900 commits in kolla | 01:08 |
sdake | crazy growth | 01:08 |
mandre | the "AttributeError: 'module' object has no attribute 'add_metaclass'" for six makies me think of an old (< 1.4) version of six, but I can't find where it's coming from | 01:08 |
*** daneyon has joined #kolla | 01:08 | |
*** daneyon_ has joined #kolla | 01:09 | |
sdake | yup i hear you | 01:10 |
*** daneyon has quit IRC | 01:13 | |
openstackgerrit | Sam Yaple proposed openstack/kolla: Convert Neutron to thin containers https://review.openstack.org/272715 | 01:13 |
sdake | i have found the problem | 01:13 |
sdake | rpm installs six 1.9.0 | 01:13 |
sdake | ansible installs six 1.10.0 | 01:14 |
sdake | we dont need 1.9.0 | 01:14 |
mandre | sdake, have you tested? | 01:15 |
sdake | mandre i looked at the typescript | 01:15 |
sdake | i am testing now | 01:15 |
sdake | wtb backward compatible libraries in python | 01:16 |
*** iceyao has joined #kolla | 01:16 | |
*** Jeffrey4l has joined #kolla | 01:18 | |
*** daneyon_ has quit IRC | 01:18 | |
sdake | mandre do you get a problem with the ansible container about /dev semlock? | 01:19 |
sdake | ok that works | 01:19 |
sdake | now I get can't connec tt o my sql server at 10.0.3.15 | 01:20 |
sdake | which is kind of what i expect with the fact that my vm networking is fucked | 01:20 |
sdake | mandre when does your work day begin? | 01:20 |
sdake | mandre http://paste.fedoraproject.org/318707/54635261/ | 01:21 |
sdake | i'm a bit jammed up with this virtualbox networking setup | 01:21 |
sdake | I have it set to nat | 01:22 |
sdake | but I think it needs to be set to something else | 01:22 |
mandre | sdake, incredibly well... you're fixing the bugs assigned to me :) | 01:22 |
*** neilus has joined #kolla | 01:22 | |
sdake | well i assigned it to me but you can assign it to you if you finis htthe job | 01:22 |
sdake | i just want it fixed :) | 01:22 |
mandre | the yum clean all on line 20 is redundant | 01:23 |
*** Allen_Gao has quit IRC | 01:24 | |
*** Allen_Gao has joined #kolla | 01:28 | |
*** neilus has quit IRC | 01:28 | |
mandre | sdake, just like you did in kolla-ansible removing python-six worked for me | 01:32 |
*** sdake has quit IRC | 01:38 | |
openstackgerrit | Martin André proposed openstack/kolla: Uninstall conflicting python-six in kolla-ansible https://review.openstack.org/268863 | 01:41 |
mandre | sdake, let's see if the gate agrees with us | 01:41 |
mandre | inc0: if you're reading this, I'm in Japan until end of March | 01:45 |
*** alisonh has quit IRC | 01:45 | |
*** jtriley has joined #kolla | 01:47 | |
*** aginwala has joined #kolla | 01:47 | |
*** tfukushima has joined #kolla | 01:50 | |
*** iotapi322 has joined #kolla | 01:52 | |
iotapi322 | Hi All, looking for a little help. I'm trying to run a build of kolla. I've tried with the tagged release of 2.000b from github, but I can't seem to get all the containers to build. | 01:54 |
iotapi322 | Is there any wiki pages or blog pages that point to a known good instructions on how to get a good build up and running, maybe a CI loop i can look at? | 01:54 |
*** sdake has joined #kolla | 01:59 | |
sdake | mandre sorry i had to take a minor emergency in the household | 01:59 |
sdake | so I may hae missed what you said | 01:59 |
sdake | mandre does that deploy for you, and my virtualbox networking is just busted? | 02:00 |
sdake | the patch isn't correct, it needs a conditional around the rpm removal | 02:00 |
sdake | and /dev needs to be sorted out - probably on centos in some way | 02:00 |
sdake | SamYaple did the find_disks.py get backported into stable after it was containerized? | 02:01 |
sdake | f | 02:01 |
sdake | 11/20/2015 | 02:03 |
sdake | was when that change went in | 02:03 |
sdake | we tagged 10/21/2015 the release | 02:03 |
sdake | that change may need to be reverted in stable | 02:03 |
*** alisonh has joined #kolla | 02:04 | |
*** jasonsb has joined #kolla | 02:10 | |
*** jasonsb has quit IRC | 02:12 | |
*** tfukushima has quit IRC | 02:13 | |
*** jasonsb has joined #kolla | 02:13 | |
sdake | f | 02:16 |
*** neilus has joined #kolla | 02:18 | |
*** aginwala has quit IRC | 02:18 | |
iotapi322 | anyone have a blog post with instructions or maybe even a bash script to share that start from scratch and build out a kolla liberty release? | 02:19 |
iotapi322 | or maybe someone can point me to the RDO CI loop? | 02:19 |
sdake | iotapi322 | 02:19 |
sdake | docs.openstack.org/developer/kolla | 02:19 |
sdake | actually | 02:20 |
sdake | those are master docs | 02:20 |
sdake | you will need to read the docs for the liberty release specifically which are here: | 02:20 |
iotapi322 | Yes I' have followed those directions and keep running into different errors such as " he command '/bin/sh -c echo 'binary not yet available for centos' && /bin/false' returned a non-zero code: 1 ERROR:__main__:mistral-engine:Parent image error'd with message "error"" | 02:21 |
iotapi322 | on numerous containers after several build retries | 02:21 |
iotapi322 | I'm just looking for someone to share a known good version... | 02:21 |
iotapi322 | like release tree tag 2.0.0.b with centos as the base | 02:22 |
*** aginwala has joined #kolla | 02:22 | |
sdake | https://github.com/openstack/kolla/tree/1a1958f209d69abfdc6634d687824f1ace62a626/doc | 02:22 |
*** alisonh has quit IRC | 02:22 | |
sdake | iotap some containers are not available for centos in binary form because rdo has not yet released them | 02:23 |
sdake | mistral is one of those | 02:23 |
sdake | those are just warnings, as long as you dont actually deploy mistral you will be good to go | 02:23 |
iotapi322 | i get those for murano, mistral, and gnocchi, | 02:24 |
sdake | I would recommend stable/liberty but we are working on 1 bug that was broken in an upstream ansible change that broke all of stable/liberty for kolla | 02:24 |
sdake | if mandre appears we should have a fix in 1-2 hours | 02:24 |
sdake | it has been root caused | 02:24 |
sdake | my house is in a remodel, so my lab is busted, so I am running on vms | 02:25 |
sdake | and I can't test the change | 02:25 |
sdake | until I figure out how to operate virtual box ;) | 02:25 |
iotapi322 | been there with the house thing | 02:25 |
sdake | yup we are on room 5 of 6 for 1100 square feet of tile | 02:25 |
sdake | and then done and back in action yay :) | 02:26 |
sdake | but I am doing all the furniture moving and floor demo | 02:26 |
sdake | which sucks ass | 02:26 |
sdake | iot is your goal to deploy kolla for eval purpose? | 02:26 |
sdake | iotapi322 | 02:27 |
mandre | sdake, the one liner did fix kolla-ansible but we seem to have different issue now: https://jenkins05.openstack.org/job/gate-kolla-dsvm-deploy-centos-binary/244/console | 02:27 |
iotapi322 | been there we moved into a fixer up a few years ago and lived up stairs while the kitchen, and master was a complete demo reno... I would go to work, then work on the house at night. I lost a few years of my life on that project | 02:27 |
sdake | iotapi322 always have to have a project :) | 02:27 |
iotapi322 | yes eval purposes, and home lab. I work for Cisco, and we have mirantis doing a deployment and I'm evaluating another way to do openstack, because deploying openstack is HARD. | 02:28 |
iotapi322 | oh wait sdrake, you must know Victor | 02:28 |
sdake | mandre that is an oslo bug | 02:28 |
mandre | you know the fix? | 02:28 |
sdake | mandre there is a patch in master for it | 02:28 |
sdake | second | 02:28 |
*** farid has quit IRC | 02:29 | |
*** farid has joined #kolla | 02:30 | |
*** neilus has quit IRC | 02:30 | |
sdake | mandre https://github.com/openstack/kolla/blob/master/docker/openstack-base/oslo-log-bug1499620.patch | 02:31 |
sdake | that is the same patch i wrote that got -2 | 02:31 |
sdake | not sure how it made it in the repo | 02:31 |
sdake | oh well ;) | 02:31 |
sdake | https://github.com/openstack/kolla/blob/master/docker/openstack-base/Dockerfile.j2#L42 | 02:32 |
sdake | https://github.com/openstack/kolla/blob/master/docker/openstack-base/Dockerfile.j2#L69 | 02:32 |
mandre | that patch is only applied on ubuntu binary | 02:32 |
sdake | ya mine was global | 02:33 |
sdake | sam didn't like my fix and -2 it | 02:33 |
sdake | let me see if i can find his solution moment | 02:33 |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Dependencies clean up / simplification https://review.openstack.org/272912 | 02:33 |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Add some unit tests for show_tasks https://review.openstack.org/273842 | 02:33 |
*** tfukushima has joined #kolla | 02:33 | |
sdake | https://github.com/openstack/kolla/blob/stable/liberty/docker/openstack-base/Dockerfile.j2#L80 | 02:35 |
sdake | that is in stable already | 02:36 |
sdake | apparently it does not work | 02:36 |
sdake | because it is source only | 02:36 |
sdake | mandre what needs to happen is that ubuntu binary patch needs to be made to work on centos as well | 02:36 |
*** alisonh has joined #kolla | 02:36 | |
*** daneyon has joined #kolla | 02:37 | |
mandre | or something requires the broken version of oslo.log | 02:38 |
sdake | nothign requires oslo log 1.11 | 02:38 |
sdake | rdo is just lazy on updating dependencies | 02:38 |
mandre | well, we might have the same type of issue with install oslo-log rpm | 02:39 |
sdake | not if you apply teh patch | 02:40 |
sdake | at some point if they udpate the repo we will have to revert the patch | 02:40 |
sdake | i have already filed a bug in red hat's bug tracker for this issue months ago | 02:40 |
iotapi322 | anyway: so the you would recommend stable/liberty, with base_distro = centos? This is a known good configuration? | 02:40 |
sdake | i'm surprised they haven't moved the version yet | 02:40 |
mandre | i don't like that solution, I prefer to get a working package | 02:40 |
sdake | mandre lets do both :) | 02:41 |
*** daneyon_ has joined #kolla | 02:41 | |
sdake | i will find the bugzilla for you so you can push appropriate buttons | 02:41 |
mandre | centos-source deploys fine in liberty btw | 02:42 |
*** achanda has joined #kolla | 02:42 | |
*** daneyon has quit IRC | 02:43 | |
iotapi322 | mandre, thanks, | 02:46 |
iotapi322 | alright, I'll give it a try and give some feedback | 02:59 |
*** unicell has quit IRC | 03:01 | |
iotapi322 | Yeah tried to do a deploy with what I had: TASK: [nova | Creating nova-compute volume] *********************************** failed: [localhost] => {"changed": true, "failed": true} msg: InvalidVersion('volumes is not available for version < 1.21',) FATAL: all hosts have already failed -- aborting | 03:02 |
*** yuanying_ has joined #kolla | 03:18 | |
*** yuanying has quit IRC | 03:21 | |
*** aginwala has quit IRC | 03:22 | |
sdake | iotapi322 if your using master ou need docker 1.10 | 03:28 |
sdake | it was tagged about 4 hours ago :) | 03:28 |
sdake | enjoy | 03:28 |
sdake | the docs are out of date | 03:28 |
sdake | let me fix that | 03:28 |
iotapi322 | ok, i was using the tag 2.0.0.b | 03:29 |
iotapi322 | off of the stable release | 03:29 |
sdake | 1.9 needed for that | 03:29 |
iotapi322 | ah ok | 03:29 |
sdake | it should be in the docs | 03:29 |
iotapi322 | i was doing the downgrade to 1.8 from the webpage | 03:29 |
sdake | we do a pretty good job of keepign the dependency requireements up to date | 03:29 |
iotapi322 | ok, i'm out for the night. THanks | 03:30 |
sdake | come back tomrorow | 03:30 |
sdake | we will be happy to help ;) | 03:30 |
iotapi322 | will do! | 03:31 |
sdake | this channel runs 24/7 | 03:31 |
*** dims_ has quit IRC | 03:33 | |
*** sdake has quit IRC | 03:37 | |
*** sdake has joined #kolla | 03:38 | |
*** yuanying has joined #kolla | 03:42 | |
mdnadeem | ajafo, I will do, Thanks :) | 03:43 |
*** yuanying_ has quit IRC | 03:44 | |
openstackgerrit | Martin André proposed openstack/kolla: Workaround for oslo.log 1.10.0 in centos-binary https://review.openstack.org/276555 | 03:46 |
*** daneyon_ has quit IRC | 03:57 | |
*** achanda has quit IRC | 03:59 | |
*** dave-mccowan has quit IRC | 04:00 | |
sdake | mandre https://review.openstack.org/#/c/268863/3/docker/kolla-ansible/Dockerfile.j2 | 04:02 |
mandre | sdake, yeah I've seen your comment, I want to get the gate results for the oslo.log patch before updating this one | 04:04 |
sdake | your patches are in the wrong order for that mandre | 04:04 |
sdake | unless that second one is really the first one | 04:04 |
sdake | learning my way around the new gerrit interface :) | 04:04 |
*** sdake has quit IRC | 04:05 | |
mandre | first one is removing python-six, which exposed the oslo.log bug | 04:06 |
*** yuanying has quit IRC | 04:06 | |
mandre | sdake, use gertty, seriously | 04:07 |
*** sdake has joined #kolla | 04:07 | |
*** yuanying has joined #kolla | 04:07 | |
sdake | back | 04:08 |
*** alisonh has quit IRC | 04:12 | |
mandre | i'll generate the proper diff once i have the container building | 04:14 |
*** sdake_ has joined #kolla | 04:16 | |
*** sdake has quit IRC | 04:17 | |
*** ssurana has joined #kolla | 04:27 | |
openstackgerrit | Martin André proposed openstack/kolla: WIP Workaround for oslo.log 1.10.0 in centos-binary https://review.openstack.org/276555 | 04:36 |
openstackgerrit | Martin André proposed openstack/kolla: Uninstall conflicting python-six in kolla-ansible https://review.openstack.org/268863 | 04:36 |
*** tfukushima has quit IRC | 04:39 | |
*** alisonh has joined #kolla | 04:49 | |
*** achanda has joined #kolla | 04:58 | |
*** aginwala has joined #kolla | 04:59 | |
*** sdake_ has quit IRC | 05:03 | |
*** salv-orlando has joined #kolla | 05:08 | |
*** sdake has joined #kolla | 05:16 | |
*** salv-orlando has quit IRC | 05:18 | |
*** tfukushima has joined #kolla | 05:18 | |
*** Marga_ has quit IRC | 05:19 | |
*** iceyao has quit IRC | 05:20 | |
*** sdake_ has joined #kolla | 05:31 | |
*** v1k0d3n has joined #kolla | 05:32 | |
*** sdake has quit IRC | 05:32 | |
openstackgerrit | Martin André proposed openstack/kolla: Workaround for oslo.log 1.10.0 in centos-binary https://review.openstack.org/276555 | 05:35 |
mandre | ^ please review, this fixes centos-binary in liberty | 05:36 |
*** sdake_ has quit IRC | 05:44 | |
mdnadeem | ajafo, ping | 05:45 |
nihilifer | ajafo is probably sleeping now | 05:46 |
mdnadeem | nihilifer, I face an issue while writing code for Ubuntu_binary for designate | 05:49 |
mdnadeem | Just for testing purpose I try to install designate api on my Ubuntu server (14.04)using command | 05:50 |
mdnadeem | $ sudo apt-get install -y --no-install-recommends designate-api | 05:50 |
*** Marga_ has joined #kolla | 05:50 | |
mdnadeem | >> it report E: Unable to locate package | 05:51 |
openstackgerrit | Michal Rostecki proposed openstack/kolla-mesos: Add nova config https://review.openstack.org/265021 | 05:51 |
nihilifer | mdnadeem: and you're using ubuntu cloud archive as a repo? | 05:51 |
mdnadeem | nihilifer, I have not create any repo, may be it use default repo | 05:52 |
*** Marga_ has quit IRC | 05:53 | |
nihilifer | mdnadeem: i don't mean creating any repos... i'm asking whether you're using https://wiki.ubuntu.com/ServerTeam/CloudArchive or some other apt repo | 05:53 |
*** Marga_ has joined #kolla | 05:54 | |
*** neilus has joined #kolla | 05:54 | |
*** Allen_Gao has quit IRC | 05:58 | |
mdnadeem | nihilifer, I guess my server use cloud archive | 05:59 |
*** Allen_Gao has joined #kolla | 05:59 | |
*** neilus has quit IRC | 05:59 | |
nihilifer | mdnadeem: just check your /etc/apt/sources.list.d | 06:05 |
nihilifer | but nvm | 06:05 |
nihilifer | i see that there is only one "designate" package | 06:05 |
nihilifer | sudo add-apt-repository cloud-archive:liberty | 06:05 |
nihilifer | sorry, bad link | 06:05 |
nihilifer | https://launchpad.net/~ubuntu-cloud-archive/+archive/ubuntu/liberty-staging | 06:05 |
nihilifer | so maybe just check whether this "designate" package has everything you need? (it means - central, api, sink, dns components of designate) | 06:06 |
nihilifer | if not, then well, we should ping ubuntu packagers then | 06:07 |
mdnadeem | nihilifer, ok, i will check | 06:07 |
*** yuanying_ has joined #kolla | 06:08 | |
*** yuanying has quit IRC | 06:11 | |
*** yuanying_ has quit IRC | 06:13 | |
*** yuanying has joined #kolla | 06:14 | |
*** yuanying has quit IRC | 06:15 | |
*** achanda has quit IRC | 06:19 | |
*** achanda has joined #kolla | 06:20 | |
*** yuanying has joined #kolla | 06:28 | |
*** iotapi322 has quit IRC | 06:29 | |
*** weiyu has joined #kolla | 06:33 | |
*** sdake has joined #kolla | 06:36 | |
openstackgerrit | Merged openstack/kolla: Update kernel version in ubuntu-bootstrap.sh https://review.openstack.org/276178 | 06:39 |
*** tfukushi_ has joined #kolla | 06:42 | |
*** tfukushima has quit IRC | 06:45 | |
elemoine- | hi | 06:45 |
*** ssurana has quit IRC | 06:49 | |
*** weiyu has quit IRC | 06:49 | |
sdake | yo elemoine- | 06:49 |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Use a new per-service format https://review.openstack.org/273843 | 06:49 |
*** ssurana has joined #kolla | 06:49 | |
sdake | mandre your 2nd patch has athe oslo log syslog bug in it for some odd reason | 06:49 |
sdake | I think you need to rebase it | 06:49 |
sdake | i have to ptfo | 06:50 |
sdake | later | 06:50 |
*** ssurana has quit IRC | 06:53 | |
*** aginwala has quit IRC | 06:53 | |
*** weiyu has joined #kolla | 06:54 | |
*** sdake has quit IRC | 06:55 | |
openstackgerrit | Martin André proposed openstack/kolla: Workaround for oslo.log 1.10.0 in centos-binary https://review.openstack.org/276555 | 06:55 |
openstackgerrit | Martin André proposed openstack/kolla: Uninstall conflicting python-six in kolla-ansible https://review.openstack.org/268863 | 06:55 |
*** sdake has joined #kolla | 06:55 | |
*** weiyu has quit IRC | 06:57 | |
mandre | sdake, that was my first patch you were looking at | 06:58 |
sdake | oslo.log works | 06:58 |
sdake | the second patch fails the gate | 06:58 |
mandre | oslo.log workaround is the second patch, so all good | 06:58 |
*** aginwala has joined #kolla | 06:58 | |
sdake | ok so can yu reorder them so the gate passes :) | 06:58 |
sdake | i guess the gate is busted either way | 06:59 |
sdake | ok bedtime | 06:59 |
mandre | you'll always have the 1rt patch fail the gate no matter what order I put them in | 06:59 |
mandre | they both fixes different issues | 06:59 |
mandre | good night sdake | 07:00 |
*** sdake has quit IRC | 07:15 | |
*** reddy has joined #kolla | 07:16 | |
*** unicell has joined #kolla | 07:22 | |
*** weiyu has joined #kolla | 07:25 | |
*** weiyu has quit IRC | 07:28 | |
*** weiyu has joined #kolla | 07:29 | |
*** weiyu has quit IRC | 07:32 | |
*** weiyu has joined #kolla | 07:33 | |
*** farid has quit IRC | 07:35 | |
*** weiyu has quit IRC | 07:36 | |
*** ppetit has joined #kolla | 07:37 | |
*** tfukushi_ has quit IRC | 07:42 | |
*** neilus has joined #kolla | 07:43 | |
*** tfukushima has joined #kolla | 07:44 | |
*** weiyu has joined #kolla | 07:47 | |
*** neilus has quit IRC | 07:47 | |
*** Jeffrey4l has quit IRC | 07:48 | |
*** dwalsh has joined #kolla | 07:51 | |
*** esikachev has joined #kolla | 07:54 | |
openstackgerrit | Merged openstack/kolla: Remove unexisting data_files in setup.cfg https://review.openstack.org/275238 | 07:54 |
*** weiyu has quit IRC | 07:55 | |
*** fgimenez has joined #kolla | 07:58 | |
*** Jeffrey4l has joined #kolla | 07:58 | |
*** weiyu has joined #kolla | 07:58 | |
*** tzn has joined #kolla | 08:01 | |
*** weiyu has quit IRC | 08:01 | |
*** weiyu has joined #kolla | 08:03 | |
openstackgerrit | Evgeny Sikachev proposed openstack/kolla-mesos: Add unit-tests for chronos.py https://review.openstack.org/276298 | 08:03 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: Fix image plugin functionality for oslo.config https://review.openstack.org/268211 | 08:06 |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: fix the custome profile raise exception issue https://review.openstack.org/272022 | 08:07 |
*** weiyu has quit IRC | 08:08 | |
*** aginwala has quit IRC | 08:10 | |
*** aginwala has joined #kolla | 08:12 | |
*** weiyu has joined #kolla | 08:12 | |
*** akwasnie has joined #kolla | 08:16 | |
*** weiyu has quit IRC | 08:16 | |
*** weiyu has joined #kolla | 08:17 | |
*** weiyu has quit IRC | 08:21 | |
*** weiyu has joined #kolla | 08:22 | |
*** weiyu has quit IRC | 08:22 | |
*** jmccarthy has quit IRC | 08:27 | |
*** jmccarthy has joined #kolla | 08:28 | |
openstackgerrit | Merged openstack/kolla: Ubuntu custom apt repos https://review.openstack.org/275463 | 08:28 |
*** athomas has joined #kolla | 08:37 | |
*** kproskurin has joined #kolla | 08:42 | |
*** Allen_Gao has quit IRC | 08:43 | |
*** salv-orlando has joined #kolla | 08:45 | |
*** mbound has joined #kolla | 08:47 | |
*** akwasnie has quit IRC | 08:47 | |
*** Allen_Gao has joined #kolla | 08:48 | |
*** salv-orlando has quit IRC | 08:49 | |
openstackgerrit | Allen Gao proposed openstack/kolla: Fix the ugly base directory path https://review.openstack.org/276628 | 08:59 |
ajafo | o/ | 09:00 |
*** salv-orlando has joined #kolla | 09:00 | |
*** aginwala has quit IRC | 09:01 | |
*** ppetit has quit IRC | 09:05 | |
*** akwasnie has joined #kolla | 09:07 | |
openstackgerrit | Martin André proposed openstack/kolla: Add a new tool to help find the latest versions https://review.openstack.org/276630 | 09:08 |
*** tzn has quit IRC | 09:10 | |
*** ppetit has joined #kolla | 09:11 | |
*** openstackgerrit has quit IRC | 09:17 | |
*** openstackgerrit has joined #kolla | 09:17 | |
*** dwalsh has quit IRC | 09:22 | |
*** fgimenez has quit IRC | 09:23 | |
openstackgerrit | Evgeny Sikachev proposed openstack/kolla-mesos: Add unit-tests for chronos.py https://review.openstack.org/276298 | 09:25 |
*** fgimenez has joined #kolla | 09:25 | |
*** achanda has quit IRC | 09:31 | |
*** neilus has joined #kolla | 09:31 | |
*** tfukushima has quit IRC | 09:31 | |
*** ppetit_ has joined #kolla | 09:32 | |
*** tfukushima has joined #kolla | 09:33 | |
*** ppetit has quit IRC | 09:33 | |
*** ppetit_ is now known as ppetit | 09:33 | |
*** tfukushima has quit IRC | 09:34 | |
openstackgerrit | Jeffrey Zhang proposed openstack/kolla: fix the custome profile raise exception issue https://review.openstack.org/272022 | 09:34 |
*** neilus has quit IRC | 09:36 | |
*** esikachev has quit IRC | 09:37 | |
*** aginwala has joined #kolla | 09:41 | |
*** aginwala has quit IRC | 09:46 | |
*** aginwala has joined #kolla | 09:47 | |
*** esikachev has joined #kolla | 09:48 | |
*** aginwala has quit IRC | 09:52 | |
*** tzn has joined #kolla | 09:56 | |
*** pbourke has quit IRC | 10:00 | |
*** tzn has quit IRC | 10:00 | |
*** pbourke has joined #kolla | 10:04 | |
*** akwasnie has quit IRC | 10:16 | |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect RabbitMQ logs https://review.openstack.org/275844 | 10:21 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder for OpenStack https://review.openstack.org/275845 | 10:21 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect Nova logs https://review.openstack.org/275846 | 10:21 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder for MariaDB https://review.openstack.org/275847 | 10:21 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add a Dockerfile for Heka https://review.openstack.org/275840 | 10:21 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka to common role https://review.openstack.org/275841 | 10:21 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect HAProxy logs https://review.openstack.org/275842 | 10:21 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect Keepalived logs https://review.openstack.org/275843 | 10:21 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect MariaDB logs https://review.openstack.org/275848 | 10:21 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder infrastructure https://review.openstack.org/276667 | 10:21 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder for RabbitMQ https://review.openstack.org/276668 | 10:21 |
*** tfukushima has joined #kolla | 10:24 | |
openstackgerrit | MD NADEEM proposed openstack/kolla: Designate ubuntu binary container https://review.openstack.org/276676 | 10:31 |
*** akwasnie has joined #kolla | 10:31 | |
*** macsz1 has quit IRC | 10:35 | |
*** macsz has joined #kolla | 10:35 | |
*** dave-mccowan has joined #kolla | 10:36 | |
*** salv-orl_ has joined #kolla | 10:39 | |
*** salv-orlando has quit IRC | 10:42 | |
*** macsz has quit IRC | 10:51 | |
*** tzn has joined #kolla | 10:54 | |
*** macsz has joined #kolla | 10:56 | |
mdnadeem | ajafo, Hi | 10:57 |
*** dave-mccowan has quit IRC | 10:58 | |
ajafo | mdnadeem: hi | 10:59 |
mdnadeem | >>regarding your comment on https://review.openstack.org/#/c/276676/1/docker/designate/designate-api/Dockerfile.j2 | 10:59 |
*** tzn has quit IRC | 10:59 | |
mdnadeem | >> You ask me to keep 8 space indentation, So should i change indentaion of my code only or keep 8 indentaion through out the file | 11:00 |
mdnadeem | As line 8 and 9 also have 4 space indent only | 11:01 |
mdnadeem | ajafo, Please confirm | 11:01 |
ajafo | mdnadeem: 8 yes, and 9 not | 11:04 |
ajafo | mdnadeem: please look as an example here https://github.com/openstack/kolla/blob/master/docker/openstack-base/Dockerfile.j2 | 11:04 |
mdnadeem | ajafo, ok, got it . Thanks a lot :) | 11:05 |
ajafo | mdnadeem: n/p btw. at this moment I've no possibility to check this patch working :( hope you did it | 11:07 |
mdnadeem | ajafo, yah, i will take care of it. | 11:09 |
ajafo | :) | 11:10 |
ajafo | thx :) | 11:10 |
*** jmccarthy has quit IRC | 11:12 | |
*** reddy has quit IRC | 11:12 | |
*** dmsimard has quit IRC | 11:12 | |
*** ccesario has quit IRC | 11:12 | |
*** neilus has joined #kolla | 11:19 | |
openstackgerrit | Merged openstack/kolla-mesos: Add some unit tests for show_tasks https://review.openstack.org/273842 | 11:21 |
*** jmccarthy has joined #kolla | 11:23 | |
*** neilus has quit IRC | 11:24 | |
*** ccesario has joined #kolla | 11:25 | |
*** reddy has joined #kolla | 11:26 | |
*** dmsimard has joined #kolla | 11:26 | |
*** gfidente has joined #kolla | 11:28 | |
*** gfidente has quit IRC | 11:31 | |
*** ppetit has quit IRC | 11:35 | |
*** macsz has quit IRC | 11:38 | |
*** macsz1 has joined #kolla | 11:38 | |
*** macsz has joined #kolla | 11:39 | |
*** dims has joined #kolla | 11:46 | |
*** esikachev has quit IRC | 11:53 | |
*** tzn has joined #kolla | 11:56 | |
*** esikachev has joined #kolla | 11:58 | |
*** salv-orl_ has quit IRC | 12:01 | |
*** tzn has quit IRC | 12:01 | |
*** salv-orlando has joined #kolla | 12:29 | |
*** fgimenez has quit IRC | 12:29 | |
*** fgimenez has joined #kolla | 12:31 | |
*** fgimenez has quit IRC | 12:31 | |
*** fgimenez has joined #kolla | 12:31 | |
*** alyson_ has quit IRC | 12:32 | |
*** tfukushima has quit IRC | 12:36 | |
*** tzn has joined #kolla | 12:36 | |
*** akwasnie has quit IRC | 12:40 | |
*** Allen_Gao has quit IRC | 12:41 | |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect RabbitMQ logs https://review.openstack.org/275844 | 12:59 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder for OpenStack https://review.openstack.org/275845 | 12:59 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder for RabbitMQ https://review.openstack.org/276668 | 12:59 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect Nova logs https://review.openstack.org/275846 | 12:59 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder for MariaDB https://review.openstack.org/275847 | 12:59 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect HAProxy and Keepalived logs https://review.openstack.org/275842 | 12:59 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect MariaDB logs https://review.openstack.org/275848 | 12:59 |
*** esikachev has quit IRC | 13:04 | |
*** neilus has joined #kolla | 13:08 | |
*** esikachev has joined #kolla | 13:09 | |
*** Jeffrey4l has quit IRC | 13:09 | |
*** Jeffrey4l has joined #kolla | 13:09 | |
*** neilus has quit IRC | 13:12 | |
*** salv-orlando has quit IRC | 13:13 | |
*** aginwala has joined #kolla | 13:18 | |
*** aginwala has quit IRC | 13:23 | |
*** jtriley has quit IRC | 13:27 | |
*** jtriley has joined #kolla | 13:29 | |
*** ppetit has joined #kolla | 13:31 | |
*** fthiagogv_ has quit IRC | 13:38 | |
*** fthiagogv has joined #kolla | 13:39 | |
*** sdake has joined #kolla | 13:40 | |
*** sdake_ has joined #kolla | 13:53 | |
*** sdake has quit IRC | 13:55 | |
*** rmart04 has joined #kolla | 14:00 | |
*** rmart04 has quit IRC | 14:01 | |
*** jtriley has quit IRC | 14:01 | |
*** dave-mccowan has joined #kolla | 14:06 | |
limamauricio | why kolla not works with the generic driver cinder? | 14:08 |
*** salv-orlando has joined #kolla | 14:13 | |
*** akwasnie has joined #kolla | 14:17 | |
sdake_ | limamauricio you mean lvm? | 14:18 |
sdake_ | limamauricio because its too hard(tm) | 14:18 |
sdake_ | limamauricio we have pretty much made a decision to abandon iscsi+lvm in favor of ceph | 14:19 |
sdake_ | limamauricio at some point we will tackle plugins for third party hardware gear, the community agrees it needs to happen just a matter of priorities | 14:19 |
sdake_ | wont happen this cycle, not a priority | 14:20 |
limamauricio | make sense | 14:20 |
sdake_ | ceph works fantasticially | 14:20 |
sdake_ | although I know your having problems with it | 14:20 |
limamauricio | :/ | 14:21 |
limamauricio | I have no knowledge of the ceph, I follow all the guides but does not work, at least the aio | 14:23 |
limamauricio | I have always the same error, [ceph | Bootstrapping Ceph OSDs] | 14:23 |
limamauricio | and the logs show me a timeout error. :/ | 14:24 |
limamauricio | I have to disable the firewalld? | 14:24 |
sdake_ | oh ya | 14:24 |
sdake_ | your firewall is on? | 14:24 |
sdake_ | openstack requires iptables not firewalld | 14:25 |
limamauricio | ok | 14:25 |
limamauricio | :) | 14:25 |
limamauricio | Error connecting to cluster: TimedOut | 14:25 |
limamauricio | i have always this error | 14:26 |
limamauricio | Error connecting to cluster: TimedOut | 14:26 |
limamauricio | :/ | 14:26 |
ccesario | limamauricio, are you from Brazil!? O_o | 14:26 |
*** fgimenez has quit IRC | 14:29 | |
*** salv-orlando has quit IRC | 14:30 | |
*** fgimenez has joined #kolla | 14:31 | |
*** fgimenez has quit IRC | 14:31 | |
*** fgimenez has joined #kolla | 14:31 | |
sdake_ | pbourke around | 14:35 |
pbourke | sdake_: yup | 14:35 |
sdake_ | need two reviews please | 14:36 |
sdake_ | stable is broken | 14:36 |
sdake_ | https://review.openstack.org/#/c/268863/ | 14:36 |
sdake_ | the gate is expected to fail on one of those | 14:36 |
sdake_ | because two things are broken | 14:36 |
sdake_ | the last patch fixes the gate | 14:36 |
sdake_ | assuming i am reading gerrit correctly | 14:36 |
sdake_ | pbourke can you confirm on that point? | 14:37 |
sdake_ | basically oslo..log fails the first patch ( the bottom of the list) | 14:37 |
sdake_ | becuase it hasn't been patched yet | 14:37 |
*** inc0 has joined #kolla | 14:37 | |
inc0 | good morning | 14:37 |
sdake_ | inc0 need revies | 14:37 |
sdake_ | https://review.openstack.org/#/c/268863/ | 14:37 |
sdake_ | stable is now working witht hese patches | 14:37 |
inc0 | wow that was fast | 14:38 |
inc0 | ok, I couldn't reproduce bug you mentioned yesterday | 14:38 |
sdake_ | the gate is expected to be broken on one of those patches | 14:38 |
sdake_ | inc0 its centos binary only | 14:38 |
sdake_ | you were running non on centos binary i assume | 14:38 |
inc0 | I see, no I run ubuntu-source | 14:38 |
inc0 | normally | 14:38 |
inc0 | yeah that makes sense | 14:39 |
inc0 | the review | 14:39 |
*** aginwala has joined #kolla | 14:39 | |
inc0 | I think venvs will prevent such things from happening in the future | 14:40 |
sdake_ | venv is not in stable | 14:40 |
sdake_ | but i agree | 14:40 |
inc0 | yeah I know | 14:40 |
sdake_ | also we are not backportign venv to stable ;) | 14:40 |
inc0 | in the future;) | 14:40 |
inc0 | sure, however what we should backport is version pin to everything | 14:41 |
inc0 | write and backport | 14:41 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect RabbitMQ logs https://review.openstack.org/275844 | 14:41 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder for OpenStack https://review.openstack.org/275845 | 14:41 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder for RabbitMQ https://review.openstack.org/276668 | 14:41 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect Nova logs https://review.openstack.org/275846 | 14:41 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder for MariaDB https://review.openstack.org/275847 | 14:41 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka to common role https://review.openstack.org/275841 | 14:41 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect HAProxy and Keepalived logs https://review.openstack.org/275842 | 14:41 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect MariaDB logs https://review.openstack.org/275848 | 14:41 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect Glance logs https://review.openstack.org/276760 | 14:41 |
* elemoine- is sorry for the patches wave | 14:41 | |
pbourke | inc0: have you come across an issue where kolla-ansible wont upgrade | 14:42 |
pbourke | inc0: I think its due to the common_run attrs that are set around it | 14:43 |
openstackgerrit | Merged openstack/kolla: Uninstall conflicting python-six in kolla-ansible https://review.openstack.org/268863 | 14:43 |
ajafo | have you notice some problems with docker 1.10? | 14:44 |
inc0 | pbourke, not yet, it just rebuilds for me | 14:44 |
openstackgerrit | Merged openstack/kolla: Workaround for oslo.log 1.10.0 in centos-binary https://review.openstack.org/276555 | 14:44 |
pbourke | inc0: hmm ok, thanks | 14:45 |
*** aginwala has quit IRC | 14:45 | |
inc0 | pbourke, you're reffering to docker upgrade or service upgrade? | 14:46 |
pbourke | inc0: service upgrade. I think I see the problem. basically if you do a liberty deploy, then try to upgrade a single service (-e action=upgrade), it wont run the common role so no new kolla_ansible gets deployed | 14:46 |
pbourke | inc0: which is a problem if the service you're upgrading needs a new kolla_ansible | 14:47 |
inc0 | well we didn't write kolla_ansible upgrade yet | 14:47 |
inc0 | we should | 14:47 |
pbourke | inc0: also it seems swift also recommend sighup when upgrading :) | 14:47 |
pbourke | for graceful shutdown | 14:48 |
pbourke | you would think there'd be a way of telling docker to send that signal though (other than exec) | 14:48 |
inc0 | afair there was issue open for that | 14:49 |
sdake_ | guys i'll be out for 2-3 hours this morning, then put together the agenda | 14:49 |
sdake_ | that should give people a chance to erview it | 14:49 |
sdake_ | gotta go right now | 14:49 |
*** sdake_ has quit IRC | 14:49 | |
inc0 | cyas | 14:49 |
*** ftcpops has joined #kolla | 14:50 | |
pbourke | inc0: actually question on that area. in your upgrade.yml you run docker start tasks on each nova container before reload.yml at the end | 14:52 |
pbourke | inc0: but wont the start tasks stop/start the containers? | 14:52 |
inc0 | yeah | 14:52 |
pbourke | inc0: making the reload pointless | 14:53 |
inc0 | it will | 14:53 |
inc0 | it does not | 14:53 |
inc0 | so I run controller nodes upgrade first | 14:53 |
inc0 | so you will have new containers in controller | 14:53 |
inc0 | then run compute one at the time | 14:53 |
inc0 | so potentially there will be hundreds of them | 14:53 |
inc0 | and it will take time | 14:53 |
inc0 | for careful people, LMing VMs out of compute just in case, weeks even | 14:54 |
inc0 | upgrade_level=auto binds nova communication to oldest version it knows in cluster | 14:54 |
pbourke | right but hold on | 14:54 |
inc0 | so if you run them initially, that still be Liberty | 14:54 |
inc0 | we need to tell it "check again, now it's all Mitaka" | 14:54 |
pbourke | ok maybe you're reloading for a different reason that I need to | 14:55 |
pbourke | for swift its to allow them to finish up all communication before stopping | 14:55 |
inc0 | that's only reason I'm reloading:) | 14:55 |
pbourke | maybe you should do the same with nova? | 14:56 |
inc0 | no, in nova I don't really care about graceful shutdown | 14:56 |
inc0 | most of operations are atomic | 14:56 |
pbourke | why not | 14:56 |
pbourke | ok | 14:56 |
*** akwasnie has quit IRC | 14:56 | |
inc0 | there are certain cases, like Live Migration | 14:56 |
*** neilus has joined #kolla | 14:56 | |
inc0 | but well, it's ongonig problem for nova anyway | 14:56 |
pbourke | what if say a request to boot has made it through the api and then the scheduler gets upgraded during that request | 14:56 |
inc0 | however that might be good thing to do later on | 14:57 |
inc0 | pbourke, if api finished to post it to rabbit, msg will stay in rabbit | 14:57 |
inc0 | and another scheduler will take it on | 14:57 |
pbourke | ok sounds good | 14:57 |
inc0 | if you happen to restart in EXACT moment this is processed, request might be lost | 14:57 |
inc0 | you can fiddle around haproxy to prevent that | 14:58 |
inc0 | for scheduler, it might be lost | 14:58 |
inc0 | but yeah, room for improvement there | 14:58 |
*** jtriley has joined #kolla | 14:58 | |
pbourke | im assuming docker stop sends sigkill? | 14:58 |
pbourke | without googling | 14:58 |
inc0 | yeah | 14:59 |
inc0 | but you can specify this in config | 14:59 |
inc0 | so we can send sigterm on docker rm | 14:59 |
SamYaple | morning | 14:59 |
SamYaple | hey inc0 pbourke | 14:59 |
inc0 | hey SamYaple | 14:59 |
pbourke | SamYaple: morning | 14:59 |
inc0 | room for improvement;) | 14:59 |
SamYaple | inc0: would you do the +W honors on the kolla-toolbox rename since you are the only one that has voiced any concern at all? | 15:00 |
*** neilus has quit IRC | 15:01 | |
inc0 | SamYaple, ok | 15:01 |
inc0 | there you go man | 15:01 |
*** jtriley has quit IRC | 15:03 | |
*** aginwala has joined #kolla | 15:04 | |
openstackgerrit | Merged openstack/kolla: Rename kolla_ansible to kolla_toolbox https://review.openstack.org/274154 | 15:04 |
openstackgerrit | Sam Yaple proposed openstack/kolla-mesos: Convert kolla_ansible container to kolla_toolbox https://review.openstack.org/276772 | 15:05 |
SamYaple | inc0: thanks! | 15:05 |
SamYaple | hey everyone, just FYI | 15:06 |
SamYaple | gates are breaking because docker 1.10 was released | 15:06 |
SamYaple | im working on a patch to fix this | 15:06 |
SamYaple | please don't recheck | 15:06 |
SamYaple | its pointless | 15:06 |
elemoine- | SamYaple, thanks | 15:06 |
*** athomas has quit IRC | 15:07 | |
elemoine- | I'm seeing errors when deploying Rsyslog | 15:07 |
SamYaple | mandre inc0 pbourke Jeffrey4l ^^ | 15:07 |
pbourke | SamYaple: thanks | 15:07 |
Jeffrey4l | roger that. thanks | 15:07 |
*** aginwala has quit IRC | 15:09 | |
*** akwasnie has joined #kolla | 15:11 | |
*** aginwala has joined #kolla | 15:16 | |
*** jtriley has joined #kolla | 15:17 | |
*** dave-mccowan has quit IRC | 15:17 | |
*** jpeeler has quit IRC | 15:20 | |
*** blahRus has joined #kolla | 15:21 | |
*** esikachev has quit IRC | 15:26 | |
*** aginwala has quit IRC | 15:27 | |
ajafo | elemoine-: it's docker 1.10 problem | 15:28 |
ajafo | probably | 15:28 |
ajafo | i've got the same | 15:28 |
elemoine- | yeah | 15:28 |
elemoine- | so let's way for SamYaple's patch | 15:28 |
elemoine- | s/way/wait | 15:28 |
ajafo | elemoine-: sthg like this http://paste.openstack.org/show/486098/ ? | 15:29 |
elemoine- | ajafo, yep exactly that | 15:29 |
*** dave-mccowan has joined #kolla | 15:32 | |
ajafo | hmm sudo yum downgrade docker-engine-1.9.1 docker-engine-selinux-1.9.1 and service restart didn't help | 15:34 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Convert Neutron to thin containers https://review.openstack.org/272715 | 15:35 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Correct issue with virtualenv rootwrap https://review.openstack.org/275504 | 15:35 |
openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Fix gate for docker 1.10 https://review.openstack.org/276788 | 15:35 |
SamYaple | ajafo: check your docker-py version | 15:35 |
SamYaple | ajafo: if you have docker-py 1.7.0 is probably causing your issue | 15:36 |
*** jtriley has quit IRC | 15:39 | |
*** jtriley has joined #kolla | 15:42 | |
*** reddy has quit IRC | 15:44 | |
sean-k-mooney | o/ | 15:45 |
sean-k-mooney | SamYaple: i was wondering if i could ask you a question about how kolla currently deploy neutron with ansible somewhat related to the neutron thin contianers | 15:46 |
*** kproskurin has quit IRC | 15:46 | |
elemoine- | inc0, I just sent a "make central logging proposal" on the mailing list, to follow up on our discussion | 15:47 |
inc0 | ok, thanks, will look it up later today | 15:48 |
inc0 | meetings meetings :( | 15:48 |
sean-k-mooney | SamYaple: currently kolla starts the network backed as part of configuring and starting the neutron agents. chould we seperate the two? | 15:48 |
elemoine- | inc0, it's WE here so no rush :) | 15:48 |
inc0 | thanks for all the cool stuff elemoine- ! | 15:48 |
elemoine- | inc0, no worries, I enjoy working in Kolla | 15:49 |
inc0 | sean-k-mooney, with docker 1.10 we potentially could | 15:49 |
ajafo | SamYaple: thx, you're right I had docker 1.7.0 | 15:49 |
inc0 | before that we had to have all the neutron agents in same container as namespave | 15:50 |
inc0 | as for ovs, well, I see no reason why not | 15:50 |
sean-k-mooney | inc0: so looking at https://review.openstack.org/#/c/272715/10/ansible/roles/neutron/tasks/start.yml i just find all the openvswitch vs linux bridge branching messey | 15:51 |
SamYaple | sean-k-mooney: can you expand on what your request is? | 15:52 |
sean-k-mooney | it would be nice if we could first start and configure the network backend(ovs/linux bridge) then handel deploying the neutron agents independently | 15:52 |
SamYaple | define independantly? | 15:52 |
SamYaple | and the agents have to start first | 15:52 |
SamYaple | neutron-server needs to start first* | 15:53 |
inc0 | SamYaple, why neutron server has to start before ovs? | 15:53 |
inc0 | not agents, ovs itself | 15:53 |
sean-k-mooney | SamYaple: i was thinking of a simple refactoring to have a start ovs and start linux bridge that would be called first the the start.yml would start the agents/server | 15:54 |
SamYaple | inc0: ovs _itself_ can start whenever | 15:54 |
inc0 | I think that's the point here | 15:54 |
SamYaple | but there is no linux-bridge continer like there is an ovs one | 15:54 |
ajafo | downgrade docker-py to 1.6 didn't help in my case :/ now i've docker 1.9.1 and docker-py 1.6.0 and still have the same issue :/ | 15:54 |
SamYaple | so im not sure what "start linux bridge" means there | 15:55 |
SamYaple | ajafo: the issue is with mounts | 15:55 |
inc0 | I actually think that would be cleaner to separate these two | 15:55 |
SamYaple | ajafo: check `mount` and unmount the docker container mounts | 15:55 |
SamYaple | inc0: define seperate, thats been my question | 15:55 |
sean-k-mooney | on sec ill share a link to the linux bridge peice | 15:55 |
SamYaple | as long as everything starts and we don't duplicate _any_ tasks, then i dont care if they are in different playbooks | 15:56 |
ajafo | SamYaple: thx for hints | 15:56 |
inc0 | SamYaple, for cleanliness purpose, at separate file in tasks | 15:56 |
SamYaple | ajafo: stop the daemon first though | 15:56 |
SamYaple | ajafo: there were some changes around all this stuff and im still working through it as well | 15:56 |
SamYaple | inc0: in fact we are going to have to break them into different playbooks anyway so as not to dup them for upgrades | 15:57 |
SamYaple | the tasks I mean | 15:57 |
sean-k-mooney | so basicaly i am just suggesting have a #include openvswitch_start.yml when neutron_plugin_agent == "openvswitch" | 15:57 |
sean-k-mooney | and remove all the direct check for openvswitch in the start.yml | 15:57 |
inc0 | yeah I'm ok with this, that's good change | 15:58 |
*** aginwala has joined #kolla | 15:58 | |
SamYaple | yea as long as funcationality doesnt change we can do includes all day long | 15:58 |
ajafo | SamYaple: yes I remember to stop deamon | 15:58 |
SamYaple | if its easier to read | 15:58 |
SamYaple | but sean-k-mooney dont worry about that now | 15:58 |
SamYaple | ive got to break all that stuff up | 15:58 |
SamYaple | for the neutron upgrades | 15:58 |
SamYaple | when the dust settles see if you still need a change | 15:59 |
sean-k-mooney | ok i might submits a patch ontop of your neutron slim containers patch to do that then. i wont change any of the task just move them so we dont have so many neutron_plugin_agent == x checks | 15:59 |
inc0 | afk for now | 15:59 |
SamYaple | sean-k-mooney: well have conflicting patches then, because ill have upgrades too | 15:59 |
*** logan_ has joined #kolla | 16:00 | |
*** logan_ is now known as Guest78654 | 16:01 | |
*** logan- has quit IRC | 16:01 | |
*** Guest78654 is now known as logan- | 16:02 | |
*** aginwala has quit IRC | 16:03 | |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect RabbitMQ logs https://review.openstack.org/275844 | 16:06 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder for OpenStack https://review.openstack.org/275845 | 16:06 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder for RabbitMQ https://review.openstack.org/276668 | 16:06 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect Nova logs https://review.openstack.org/275846 | 16:06 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder for MariaDB https://review.openstack.org/275847 | 16:06 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add a Dockerfile for Heka https://review.openstack.org/275840 | 16:06 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka to common role https://review.openstack.org/275841 | 16:06 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect HAProxy and Keepalived logs https://review.openstack.org/275842 | 16:06 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Add Heka log decoder infrastructure https://review.openstack.org/276667 | 16:06 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect Glance logs https://review.openstack.org/276760 | 16:06 |
openstackgerrit | Eric Lemoine proposed openstack/kolla: Make Heka collect MariaDB logs https://review.openstack.org/275848 | 16:06 |
*** sean-k-m1 has joined #kolla | 16:07 | |
sean-k-m1 | sorry i had a disconnect so i was missing lines. sure ill wait for your upgrade stuff to land and look at it later if it still need to be done | 16:13 |
*** sean-k-mooney has quit IRC | 16:14 | |
*** sean-k-m1 has quit IRC | 16:14 | |
SamYaple | sean-k-m1: i think what you are asking for may just happen just because, unrelated to why you are asking for it | 16:14 |
*** Jeffrey4l has quit IRC | 16:15 | |
SamYaple | since we have to start sharing the tasks between deploy and upgrade, they just get split up anyway | 16:15 |
*** sean-k-mooney has joined #kolla | 16:15 | |
*** jpeeler has joined #kolla | 16:15 | |
*** jpeeler has joined #kolla | 16:15 | |
*** jtriley has quit IRC | 16:16 | |
sean-k-mooney | SamYaple: ya it likely will be covered by the deploy/upgrade refactoring. cool it just jumped out at me as something that would simply the start.yml when i was looking at your review | 16:18 |
*** esikachev has joined #kolla | 16:19 | |
*** jtriley has joined #kolla | 16:28 | |
*** esikachev has quit IRC | 16:30 | |
*** esikachev has joined #kolla | 16:32 | |
*** shardy has joined #kolla | 16:43 | |
*** athomas has joined #kolla | 16:43 | |
*** neilus has joined #kolla | 16:45 | |
*** neilus has quit IRC | 16:49 | |
*** esikachev has quit IRC | 16:49 | |
*** esikachev has joined #kolla | 16:52 | |
*** jtriley has quit IRC | 16:53 | |
*** esikachev has quit IRC | 16:58 | |
*** macsz has quit IRC | 17:01 | |
*** aginwala has joined #kolla | 17:01 | |
*** jtriley has joined #kolla | 17:01 | |
*** aginwala has quit IRC | 17:02 | |
*** akwasnie has quit IRC | 17:02 | |
*** ppetit has quit IRC | 17:06 | |
*** achanda has joined #kolla | 17:15 | |
*** fgimenez has quit IRC | 17:19 | |
*** farid has joined #kolla | 17:20 | |
openstackgerrit | Sam Yaple proposed openstack/kolla: Fix two missing toolbox references https://review.openstack.org/276841 | 17:25 |
SamYaple | inc0: pbourke https://review.openstack.org/276841 | 17:25 |
SamYaple | out-of-date patch merged | 17:25 |
SamYaple | our gates need to be voting :( | 17:25 |
*** ftcpops has quit IRC | 17:27 | |
*** unicell has quit IRC | 17:37 | |
SamYaple | nihilifer: https://review.openstack.org/#/c/276772/ | 17:39 |
*** ssurana has joined #kolla | 17:45 | |
openstackgerrit | Sam Yaple proposed openstack/kolla: Make a final check to validate mysql works https://review.openstack.org/276851 | 17:49 |
SamYaple | git branch | 17:49 |
openstackgerrit | Merged openstack/kolla: Fix the ugly base directory path https://review.openstack.org/276628 | 17:52 |
openstackgerrit | Sam Yaple proposed openstack/kolla: [WIP] Fix gate for docker 1.10 https://review.openstack.org/276788 | 17:53 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Convert Neutron to thin containers https://review.openstack.org/272715 | 17:53 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Correct issue with virtualenv rootwrap https://review.openstack.org/275504 | 17:53 |
openstackgerrit | Merged openstack/kolla: Fix manila perms https://review.openstack.org/275359 | 17:53 |
*** salv-orlando has joined #kolla | 17:55 | |
*** sdake has joined #kolla | 17:56 | |
*** esikachev has joined #kolla | 17:57 | |
openstackgerrit | Merged openstack/kolla-mesos: Convert kolla_ansible container to kolla_toolbox https://review.openstack.org/276772 | 17:58 |
*** sdake_ has joined #kolla | 17:59 | |
openstackgerrit | Paul Bourke proposed openstack/kolla: Upgrade tasks for Swift https://review.openstack.org/276855 | 17:59 |
openstackgerrit | Paul Bourke proposed openstack/kolla: Upgrade tasks for Swift https://review.openstack.org/276855 | 18:00 |
*** sdake has quit IRC | 18:01 | |
nihilifer | SamYaple: can you look at this named volumes workaround? https://review.openstack.org/#/c/275237 | 18:02 |
nihilifer | it's dirty as hell and only for AIO | 18:03 |
*** esikachev has quit IRC | 18:03 | |
nihilifer | but well, i'd like to wait for mesos to deal with named volumes | 18:03 |
nihilifer | before trying to do such a workaroud for multinode, which would be even more ugly | 18:03 |
*** achanda has quit IRC | 18:06 | |
*** pbourke has quit IRC | 18:06 | |
nihilifer | oh, nvm, i have over 9k merge conflicts now | 18:09 |
*** unicell has joined #kolla | 18:10 | |
limamauricio | SamYaple i have to create two partitions sdb1 and sdb2 to deploy ceph? | 18:10 |
*** athomas has quit IRC | 18:11 | |
limamauricio | docker logs bootstrap_osd_0 -> http://pastebin.com/2fMqiKqn | 18:12 |
*** blahRus has quit IRC | 18:12 | |
*** achanda has joined #kolla | 18:18 | |
*** dims is now known as dimsum__ | 18:30 | |
*** neilus has joined #kolla | 18:33 | |
*** sdake_ has quit IRC | 18:36 | |
*** esikachev has joined #kolla | 18:36 | |
*** neilus has quit IRC | 18:38 | |
*** mbound has quit IRC | 18:38 | |
*** sdake has joined #kolla | 18:41 | |
*** aginwala has joined #kolla | 18:46 | |
*** ayoung has joined #kolla | 18:49 | |
ayoung | Can anyone give me the simplest way to deploy just Keystone and MySQL on a local host using Kolla? | 18:50 |
*** esikachev has quit IRC | 18:50 | |
ayoung | I'm sure it is trivial, but I'm having trouble knowing where to start. Do I need an inventory file> | 18:51 |
ayoung | and then use Ansible? | 18:51 |
inc0 | well | 18:51 |
inc0 | you always need inventory | 18:52 |
inc0 | it tells ansible where to deploy | 18:52 |
inc0 | bot if you want to do local deploy, there is predefined inventory for that | 18:52 |
inc0 | aio | 18:52 |
inc0 | ayoung, so https://github.com/openstack/kolla/blob/master/doc/quickstart.rst this is quickstart guide | 18:53 |
SamYaple | nihilifer: ok ill look | 18:53 |
ayoung | inc0, OK..I've been reading that. Still going through all the "here;'s how you fetch docker" type lines | 18:54 |
SamYaple | limamauricio: no, check the guide | 18:54 |
inc0 | to deploy just keystone you can run tools/kolla-ansible deploy -t keystone | 18:54 |
*** esikachev has joined #kolla | 18:54 | |
SamYaple | limamauricio: you should onlt be creating one partition named KOLLA_BOOSTRAP_OSD | 18:54 |
ayoung | inc0, let me try that... | 18:55 |
inc0 | ayoung, so tell me what do you have now and what is problem you're having | 18:55 |
inc0 | also | 18:55 |
inc0 | correct | 18:55 |
inc0 | correction | 18:55 |
ayoung | inc0, ok so I need an inventory, ...and I guess if I am going to do this on my local machine, I need sudo set up to not require a password from whatever user I connect to... | 18:57 |
ayoung | inc0, I'm on F23, my laptop, an I'd like to not open and glaring security holes if I can help it | 18:58 |
ayoung | I'm OK with having to type my sudo password to run this...prefer it to disabling passwords, or making an key'ed log in to a trusted user. | 18:58 |
*** shardy has quit IRC | 18:59 | |
SamYaple | ayoung: F23 wont work | 18:59 |
ayoung | SamYaple, I only want Keystone | 18:59 |
ayoung | SamYaple, I don't care about Nova for this | 18:59 |
ayoung | I might pull in glance to have a remote app, but even that is optional | 19:00 |
SamYaple | ayoung: oh hey you are ayoung! | 19:00 |
SamYaple | i know you | 19:00 |
ayoung | Yes you do | 19:00 |
SamYaple | give me a minute ill see what you need | 19:00 |
limamauricio | KOLLA_CEPH_OSD_BOOTSTRAP or KOLLA_BOOSTRAP_OSD? | 19:00 |
SamYaple | limamauricio: the first one, apologies | 19:00 |
SamYaple | the guide does work, i promise | 19:01 |
limamauricio | ok | 19:01 |
*** aginwala has quit IRC | 19:01 | |
limamauricio | :) | 19:01 |
SamYaple | so ayoung just keystone not mariadb or any other services? | 19:02 |
ayoung | SamYaple, I need maria too | 19:04 |
ayoung | SamYaple, I'm trying to mimic a live setup for Functional testing, but also want to, in the future, replace the Keystone code with -v mounted file from the underlying FS (my git repo) | 19:05 |
SamYaple | ayoung: cool. ill patsebin the config you need in globals.yml for just keystone and mariadb | 19:05 |
ayoung | thanks | 19:05 |
SamYaple | ayoung: http://paste.openstack.org/show/486128/ | 19:12 |
ayoung | SamYaple, anyway I can put them somewhere other than /etc | 19:13 |
ayoung | I want to make this a developer friendly approach, very little Sudo/Root magic other than launching things via docker | 19:13 |
SamYaple | ayoung: no thats hardcoded in kolla-ansible script. youll ahve to change ./tools/kolla-ansible | 19:13 |
SamYaple | but if you pop open the bash script it should be straight forward | 19:14 |
SamYaple | however, you wil have to adjsut other values to not use /etc | 19:14 |
SamYaple | let me look those up for you | 19:14 |
SamYaple | fyi ayoung you do need to build the following images (you can use this command) `./tools/build.py keystone mariadb rsyslog kolla-toolbox` | 19:15 |
SamYaple | those are all you will need (saves alot of build time) | 19:15 |
*** esikachev has quit IRC | 19:16 | |
*** iotapi322 has joined #kolla | 19:16 | |
iotapi322 | Hi after connecting last night trying to get a build, sdrake suggested that I try with the stable/liberty release, however I can't seem to find the command "kolla-build" in this particular stable branch. | 19:17 |
iotapi322 | what command should i run to build the containers? | 19:17 |
inc0 | iotapi322, tools/build.py | 19:18 |
inc0 | standalone script with same command options as kolla-build | 19:18 |
ayoung | SamYaple, is this tox/venv based ? | 19:18 |
inc0 | actually kolla-build calls it underneath;) | 19:19 |
iotapi322 | so for an all in one in the /etc/kolla/globals file do you suggest the openstack_release to be latest or to be liberty? | 19:19 |
SamYaple | ayoung: you can run it in a venv, there is no tox command to build _just_ what you have asked for though | 19:20 |
ayoung | SamYaple, Nah, I meant for building the venv in the first place | 19:20 |
ayoung | SamYaple, tox -r -e py34 ? | 19:20 |
ayoung | or py27 ? | 19:21 |
SamYaple | shoudl work for either 27 or 34 | 19:21 |
inc0 | iotapi322, so that's probably something we have to change as it's misleading. openstack_release in global.yml is tag which you build your images with | 19:21 |
SamYaple | though 27 is the most tested | 19:21 |
ayoung | thats what I was asking thanks | 19:21 |
inc0 | or rather what image tag it will look for | 19:21 |
inc0 | by default both are 2.0 | 19:21 |
inc0 | in master | 19:21 |
inc0 | in liberty/stable I think its still latest | 19:22 |
inc0 | defaults should work in any case | 19:22 |
SamYaple | shoot ayoung. testing this outside of /etc/ has reileved a problem. ill have a patch up momentarily (blame inc0! .... and those that reviewed his patch) | 19:22 |
ayoung | INFO:__main__:Found the docker image folder at /opt/stack/kolla/docker | 19:22 |
ayoung | Traceback (most recent call last): | 19:22 |
SamYaple | revieled* | 19:22 |
ayoung | revealed? | 19:23 |
SamYaple | thats the one | 19:23 |
ayoung | k | 19:23 |
iotapi322 | Ah i understand. | 19:23 |
inc0 | SamYaple, what? now hold on, I have my rights | 19:23 |
ayoung | so long as you were not reviling inc0 cuz he's on my good list | 19:23 |
SamYaple | inc0: :P its old code | 19:24 |
SamYaple | inc0: just a missed thing from syslog patch | 19:24 |
*** salv-orlando has quit IRC | 19:24 | |
inc0 | duh | 19:24 |
inc0 | it's not my fault, I did most of stuff;) | 19:25 |
inc0 | well anyway, add patch I want to see what I'm blamed for | 19:25 |
inc0 | my name needs cleaning now | 19:25 |
*** mimizone has quit IRC | 19:26 | |
iotapi322 | building these docker containers and dealing with the timeouts is terrible. Does everyone just run their own docker repositories? Or do you guys use proxy's on your local networks or what? | 19:29 |
*** mimizone has joined #kolla | 19:29 | |
*** mimizone has quit IRC | 19:30 | |
*** mimizone has joined #kolla | 19:31 | |
inc0 | for proxy we have -i and -I options in build | 19:31 |
inc0 | you can inject proxy setting code | 19:31 |
inc0 | I build all images locally anyway | 19:32 |
iotapi322 | Yeah i'm trying to setup a squid proxy here at my local network for this. i keep blowing up on : | 19:34 |
iotapi322 | Cannot open: http://repo.percona.com/release/7/RPMS/x86_64/percona-release-0.1-3.noarch.rpm. Skipping. INFO:__main__:base:Error: Nothing to do INFO:__main__:base: | 19:34 |
ayoung | SamYaple, is kolla making assumptions about access control to DOcker? | 19:35 |
ayoung | docker.errors.DockerException: Error while fetching server API version: ('Connection aborted.', error(13, 'Permission denied')) | 19:35 |
ayoung | (py27)[ayoung@ayoung54 | 19:35 |
*** aginwala has joined #kolla | 19:36 | |
*** limamauricio has quit IRC | 19:36 | |
*** aginwala has quit IRC | 19:36 | |
iotapi322 | ayoung - i had to have the python-docker tools installed | 19:37 |
*** mbound has joined #kolla | 19:38 | |
ayoung | iotapi322, on F23? | 19:39 |
SamYaple | ayoung: youll have to forgive the slowness, im working at beer capacity | 19:40 |
ayoung | SamYaple, heh | 19:40 |
ayoung | Brno? | 19:40 |
SamYaple | ayoung: kolla is just assuming it can connect to the docker socket | 19:40 |
ayoung | SamYaple, I figured it out: | 19:41 |
ayoung | http://askubuntu.com/questions/477551/how-can-i-use-docker-without-sudo | 19:41 |
ayoung | probably a security hole, but I can live with it for now | 19:41 |
SamYaple | yep thats the one | 19:41 |
SamYaple | its a docker security hole to be honest | 19:41 |
SamYaple | but it can be rememdedy with tls conenction | 19:41 |
ayoung | docker is a security hole | 19:41 |
ayoung | no iot can't | 19:41 |
ayoung | it | 19:41 |
SamYaple | yea but docker != security | 19:41 |
ayoung | yep | 19:42 |
SamYaple | i mean tcp+tls | 19:42 |
SamYaple | why can it not? | 19:42 |
ayoung | if I really, really, reaaly cared what I would do is: | 19:42 |
ayoung | createa develop user | 19:42 |
ayoung | su - developer | 19:42 |
ayoung | and do everything as that user | 19:42 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Fix non-root deploys https://review.openstack.org/276887 | 19:42 |
ayoung | not as ayoung | 19:42 |
SamYaple | ayoung: ^^ that _should_ do it, but im testing now | 19:42 |
SamYaple | added an option to kolla-ansible script too | 19:42 |
ayoung | SamYaple, cool...I'm still a couple steps behind, but I;'ll catch up | 19:43 |
*** mbound has quit IRC | 19:43 | |
SamYaple | ayoung: youve come at a bad time too, docker 1.10 was released yesterday and our gate is AFU | 19:43 |
SamYaple | probably will hit you too, still working through the issue | 19:43 |
SamYaple | this should all work with docker 1.9 and docker-py 1.6.0 | 19:43 |
SamYaple | if you run into any issues that is | 19:44 |
ayoung | SamYaple, There is no such thing as a good time. Good time means everything is stable, which means everyone has moved on to new projects and your code is legacy | 19:44 |
SamYaple | tis true | 19:45 |
ayoung | SamYaple, so, here is the plan | 19:45 |
ayoung | get this to work with the code in the built docker images | 19:45 |
ayoung | then createa new keystone image where I mount /opt/stack via -v | 19:45 |
ayoung | then, inside the keystone container, to a python setup.py of the keystone code so that it points at the /opt/stack/keystone repo | 19:46 |
ayoung | and run Keystone that way | 19:46 |
ayoung | use that to check commits in my laptop git repo | 19:46 |
ayoung | think it will work? | 19:46 |
inc0 | bbl | 19:47 |
*** inc0 has quit IRC | 19:47 | |
SamYaple | ayoung: i test patched the way i _think_ you are describing, yes | 19:47 |
SamYaple | but one of these ways it should work | 19:47 |
ayoung | SamYaple, good to know. | 19:47 |
ayoung | I don't want to have to do a complete rebuild each time I want to test a change from Keystone if I can help it | 19:47 |
SamYaple | yea that model doesnt work for devel | 19:48 |
ayoung | SamYaple, and devstack is way too invasive | 19:48 |
SamYaple | you can rerun setup.py from in the container, and you can bindmount from outside the container... so this _should_ all work | 19:48 |
SamYaple | but if it doesnt work the way you want you I gaurantee we are a patchset away from making it do so | 19:48 |
ayoung | SamYaple, what version of the keystone code is going to be running when I do the ./tools/.... build you linked above? | 19:49 |
SamYaple | ayoung: it pulls master tarball from tarballs.openstack.org | 19:49 |
SamYaple | you can configure that to point to a git repo | 19:49 |
SamYaple | and i _think_ we have local as well | 19:49 |
ayoung | So right now it will be Keystone Mitaka RC 1? | 19:49 |
ayoung | Er Milestone 1 | 19:49 |
SamYaple | whatever master tarball is | 19:49 |
SamYaple | ayoung: http://tarballs.openstack.org/keystone/keystone-master.tar.gz | 19:50 |
ayoung | keystone-master.tar.gz05-Feb-2016 09:09 1.0M | 19:50 |
SamYaple | but again, this is configurable | 19:50 |
ayoung | Its OK, I just want to know the delta | 19:50 |
ayoung | that should be close enough to a git checkout that I won't be jumping major versions | 19:50 |
*** aginwala has joined #kolla | 19:51 | |
ayoung | SamYaple, OK...I have a completed build... | 19:52 |
SamYaple | ayoung: im still testing my patch for non-root deploy | 19:52 |
ayoung | SamYaple, I'll test with what you pasted abovce, and then try your patch after | 19:52 |
ayoung | minimize the number of moving pieces. | 19:53 |
SamYaple | ayoung: awesome. get it working root first, then non-root will make more sense | 19:53 |
ayoung | SamYaple, Ok you have kolla_base_distro: "ubuntu" | 19:53 |
ayoung | can I make that Fedora? | 19:53 |
SamYaple | you can make it centos | 19:53 |
SamYaple | im not sure the fedora images build | 19:53 |
SamYaple | centos we gate against | 19:53 |
SamYaple | as long as you dont run nova, youll be fine | 19:53 |
*** sdake has quit IRC | 19:54 | |
ayoung | OK. Let me try with Centos First... | 19:55 |
ayoung | I'll try Fedora later | 19:55 |
SamYaple | btw ayoung i did a full keystone v3 kolla build in liberty based on one of you blog posts. worked well with kolla. hoping to have a v3 only option for kolla before mitaka drops | 19:55 |
SamYaple | i say worked well... the service that worked with v3 worked well ;) | 19:55 |
ayoung | SamYaple, that makes me so happy | 19:55 |
*** Marga_ has quit IRC | 20:04 | |
ayoung | SamYaple, so it looks like it is insisting on at least an empty dict in /etc/kolla/passwords | 20:13 |
ayoung | and /etc/kolla/rsyslog needs to be writible, I think | 20:14 |
*** dave-mccowan has quit IRC | 20:14 | |
openstackgerrit | Merged openstack/kolla: Fix two missing toolbox references https://review.openstack.org/276841 | 20:14 |
ayoung | ah, needs to exist and be a directory and writable | 20:15 |
SamYaple | ayoung: ive just done a full deploy from an unprivileged user | 20:17 |
SamYaple | so should be good with my patch. let me post some steps | 20:18 |
SamYaple | ayoung: http://paste.openstack.org/show/486143/ | 20:20 |
SamYaple | /home/sam/kolla_config contains globals.yml and passwords.yml only | 20:20 |
SamYaple | (the deploy populates it with more files of course) | 20:20 |
*** neilus has joined #kolla | 20:21 | |
ayoung | SamYaple, next up was "ERROR! ERROR! 'docker_registry_password' is undefined" | 20:21 |
SamYaple | ayoung: just copy the passwords.yml from the repo | 20:22 |
SamYaple | those variables need to be set | 20:22 |
SamYaple | its a jinja2/ansible thing | 20:22 |
*** sdake has joined #kolla | 20:23 | |
SamYaple | youll only need to tweak the mariabd and keystone password variable | 20:23 |
ayoung | SamYaple, makes sense. What do you mean by tweak? | 20:23 |
SamYaple | default password is "password" | 20:24 |
SamYaple | if you want it to be different, change the value | 20:24 |
SamYaple | you dont have to change the value | 20:24 |
*** salv-orlando has joined #kolla | 20:24 | |
SamYaple | if you dont adjust the variables in passwords.yml youll end up with a database with root:password with keystone user keystone:password | 20:25 |
ayoung | SamYaple, so, I ran this without a password.yml file, and then again with, and it errored on rsyslog | 20:25 |
ayoung | An exception occurred during task execution. To see the full traceback, use -vvv. The error was: KeyError: 0 | 20:25 |
ayoung | fatal: [localhost]: FAILED! => {"changed": false, "failed": true, "parsed": false} | 20:25 |
ayoung | task was | 20:25 |
ayoung | TASK [common : Starting rsyslog container] | 20:25 |
SamYaple | what version of docker and docker-py are you running? | 20:25 |
*** neilus has quit IRC | 20:26 | |
ayoung | docker-1.9.1-6.git6ec29ef.fc23.x86_64 | 20:36 |
ayoung | docker-py==1.7.0 | 20:36 |
SamYaple | try 1.6.0 for docker-py | 20:36 |
ayoung | SamYaple, that is in venv | 20:36 |
ayoung | hmmm | 20:36 |
*** dave-mccowan has joined #kolla | 20:37 | |
ayoung | ok setting it to =1.6... | 20:37 |
SamYaple | so we have an ansible module, kolla_docker. when i wrote it I did not implement the "command" part of the api and I *think* thats causing our troubles | 20:37 |
SamYaple | im still investigating | 20:37 |
ayoung | hmmm | 20:38 |
ayoung | <1.7? | 20:38 |
SamYaple | yea | 20:38 |
ayoung | Nope | 20:38 |
ayoung | doesn't like that | 20:38 |
SamYaple | how so? | 20:38 |
SamYaple | like you cant install it? | 20:38 |
ayoung | disregard. install ran | 20:39 |
ayoung | ok testing again | 20:39 |
openstackgerrit | Sam Yaple proposed openstack/kolla: Fix non-root deploys https://review.openstack.org/276887 | 20:40 |
SamYaple | docker-engine 1.9.1 + docker-py 1.6.0 works for me (and its worked for master for a month or so) | 20:41 |
*** inc0 has joined #kolla | 20:43 | |
ayoung | SamYaple, nope, no difference | 20:44 |
ayoung | TASK [common : Starting rsyslog container] ************************************* | 20:44 |
ayoung | An exception occurred during task execution. To see the full traceback, use -vvv. The error was: KeyError: 0 | 20:44 |
ayoung | fatal: [localhost]: FAILED! => {"changed": false, "failed": true, "parsed": false} | 20:44 |
ayoung | lemme try with -vvvv | 20:45 |
SamYaple | ... KeyError.... | 20:45 |
SamYaple | is this master branch? | 20:45 |
*** achanda has quit IRC | 20:45 | |
dmsimard | SamYaple: last time I saw something really weird you happened to have seen it too | 20:51 |
dmsimard | Have you come across this on master? :P | 20:51 |
dmsimard | http://logs.openstack.org/92/276492/6/check/gate-puppet-openstack-integration-scenario001-tempest-dsvm-centos7/78b9c32/logs/nova/nova-compute.txt.gz#_2016-02-05_20_21_53_983 | 20:51 |
dmsimard | tempest failing on [most|all] VM related tests | 20:52 |
SamYaple | dmsimard: policy.json related methings | 20:52 |
*** aginwala has quit IRC | 20:53 | |
ayoung | SamYaple, so, yes, this is master, and I'm looking at the code now | 20:58 |
SamYaple | ayoung: im in docker-dev, they are confused right now as well | 20:58 |
SamYaple | working on it | 20:58 |
ayoung | SamYaple, http://git.openstack.org/cgit/openstack/kolla/tree/ansible/library/kolla_docker.py#n564 | 20:59 |
ayoung | Its the return from | 20:59 |
ayoung | http://git.openstack.org/cgit/openstack/kolla/tree/ansible/library/kolla_docker.py#n553 | 20:59 |
*** achanda has joined #kolla | 20:59 | |
SamYaple | ayoung: are you ansible 2.0? | 21:00 |
ayoung | SamYaple, nope 1.9 | 21:01 |
SamYaple | 1.9.4? | 21:01 |
ayoung | I lied | 21:01 |
ayoung | ansible --version | 21:01 |
ayoung | ansible 2.0.0.2 | 21:01 |
SamYaple | yup | 21:01 |
SamYaple | thatll do it | 21:01 |
SamYaple | we havent converted to 2.0 yet | 21:01 |
ayoung | rpmquery -f `which ansible` | 21:01 |
ayoung | ansible-2.0.0.2-1.fc23.noarch | 21:01 |
ayoung | Gah... | 21:01 |
SamYaple | always another thing | 21:02 |
ayoung | let me see if I can downgrade | 21:02 |
SamYaple | but good to know that error | 21:02 |
ayoung | I don't need 2 yuet | 21:02 |
ayoung | yet | 21:02 |
SamYaple | ill patch it with a better error message | 21:02 |
SamYaple | so reqs shoudl be docker-engine==1.9.1 docker-py==1.6.0 ansible==1.9.4 | 21:02 |
ayoung | dnf downgrade ftw! | 21:03 |
SamYaple | it is _unlikely_ that we will switch to ansible 2.0 for mitaka | 21:03 |
SamYaple | possibly, well be talking about it at midcycle | 21:03 |
SamYaple | but if not mitaka, first thing in newton cycle | 21:03 |
ayoung | Nah, too soon | 21:03 |
SamYaple | unfortuantely we have to | 21:03 |
SamYaple | ansible bugs they wont fix :/ | 21:03 |
ayoung | Yeah, Newton is right. I thought I was still on 1.9 | 21:03 |
ayoung | coulda swore I checked that | 21:03 |
ayoung | Ah..that was docker, not Ansible ...silly me | 21:04 |
SamYaple | so many numbers | 21:04 |
*** esikachev has joined #kolla | 21:05 | |
*** aginwala has joined #kolla | 21:06 | |
*** achanda has quit IRC | 21:06 | |
ayoung | SamYaple, TASK: [common | Starting rsyslog container] *********************************** | 21:09 |
ayoung | failed: [localhost] => {"failed": true} | 21:09 |
ayoung | msg: The requested image does not exist: kollaglue/centos-source-rsyslog:2.0.0 | 21:09 |
ayoung | hmmm | 21:09 |
ayoung | kollaglue/centos-binary-rsyslog 2.0.0 6b727d9e6e42 5 minutes ago 339 MB | 21:09 |
ayoung | SamYaple, what's the difference? whether it was built as part of the | 21:10 |
SamYaple | ayoung: binary vs source :P | 21:10 |
ayoung | ./tools/build.py keystone mariadb rsyslog kolla-toolbox | 21:10 |
SamYaple | sorry thats my bad | 21:10 |
SamYaple | ./tools/build.py --type source keystone mariadb rsyslog kolla-toolbox | 21:10 |
ayoung | kolla_install_type | 21:10 |
SamYaple | yea | 21:10 |
SamYaple | you built rdo packages | 21:11 |
ayoung | in globals, change that to binary | 21:11 |
SamYaple | globals is for ansible only | 21:11 |
SamYaple | but if you change globals to binary it will deploy yes | 21:11 |
ayoung | k | 21:11 |
SamYaple | you did not build containers with the toolchain in them so you won't be able to do the setup.py dev tests | 21:11 |
SamYaple | just fyi | 21:11 |
ayoung | SamYaple, ok so the source usese the tarball, the binary builds using Deloarean packages? | 21:12 |
ayoung | Delorean | 21:12 |
SamYaple | correct | 21:12 |
*** achanda has joined #kolla | 21:12 | |
ayoung | SamYaple, I kindof love this....it feels right after Puppet | 21:12 |
SamYaple | i also kind of love this, and once you get it setup it is _much_ _much_ easier to work with | 21:13 |
SamYaple | but the learning curve is a bit high if you dont know docker or ansible | 21:13 |
ayoung | SamYaple, I did a whole Proof of concept using Ansible and Packstack last summit. I've been battling tripleo and Puppet stuff. The tooling is essential, and I think its taken a while to get the right vision here | 21:14 |
ayoung | I've been working with Ansible for a while now, but Docker is new to me | 21:14 |
ayoung | and I like that we have a whole docker ansible savvy team here | 21:14 |
SamYaple | i have used docker for a few years | 21:15 |
SamYaple | then i worked at RAX during the OSAD devel (not as a dev) | 21:15 |
ayoung | I'd love to replace devstack with something along these lines. Narrow the distance between development and deployments | 21:15 |
SamYaple | so that introduced me to ansible | 21:15 |
SamYaple | i started a docker+ansible project, then merged it with kolla | 21:15 |
SamYaple | so i know the pain of deployment | 21:16 |
SamYaple | would love to see devstack replaceemnt with kolla workflow | 21:16 |
*** sdake has quit IRC | 21:16 | |
ayoung | SamYaple, so, maybe we need: ./tools/build.py --type git keystone mariadb rsyslog kolla-toolbox | 21:16 |
SamYaple | well you can deploy from git now | 21:16 |
SamYaple | its a config option | 21:17 |
ayoung | ok...good enough | 21:17 |
ayoung | I'll get there | 21:17 |
SamYaple | let me find it... | 21:17 |
SamYaple | yea ayoung if you tox gen the config youll get a kolla-build.conf config | 21:17 |
SamYaple | you can change that to git instead of url and point it to a repo | 21:18 |
ayoung | OK...might bug you again if I have trouble there...still going for positive thread here on the source build | 21:18 |
SamYaple | yea get _something_ going and then we can move forward | 21:18 |
SamYaple | unfortuantely i think soe of this stuff is just tribal knowlege on master | 21:19 |
SamYaple | some* | 21:19 |
ayoung | SamYaple, I might push to use something like this for Keystone functional tests | 21:19 |
ayoung | Need tests that run against a live MySQL and LDAP backend, Horizon with some Web SSO Identity provider and the like | 21:20 |
ayoung | this is much lower barrier to entry than devstack and full VMs | 21:20 |
SamYaple | ayoung: I will 100% help you with that | 21:21 |
SamYaple | I did all the gate work here for kolla | 21:21 |
SamYaple | (dont look its awful!) | 21:21 |
SamYaple | we are at an empasse with teh gate | 21:21 |
SamYaple | we are out of ram | 21:21 |
ayoung | SamYaple, um...postgresql? | 21:22 |
ayoung | ah, just an RPM | 21:22 |
SamYaple | i dont understand the question | 21:22 |
*** aginwala has quit IRC | 21:22 | |
ayoung | INFO:__main__:openstack-base: Verifying : postgresql-devel-9.2.14-1.el7_1.x86_64 6/39 | 21:22 |
ayoung | I was looking at the log as build was running and saw a postgresql reference. | 21:23 |
*** aginwala has joined #kolla | 21:23 | |
ayoung | SamYaple, let me get this working, and I can then turn around and talk with the rest of Keystone Core and say "hey, this works" | 21:24 |
ayoung | "lets use it to test" | 21:24 |
SamYaple | ayoung: yea thats for... something i think thats for gnocchie | 21:24 |
SamYaple | gnocchi* | 21:24 |
ayoung | and I think they will be happy with it | 21:24 |
SamYaple | something needs the devel lib for postgresql | 21:24 |
SamYaple | cool! | 21:24 |
*** esikachev has quit IRC | 21:24 | |
ayoung | Yeah, I saw sqlite as well...few dependencies | 21:25 |
SamYaple | well we have an openstack-base container that has lots of common deps | 21:25 |
SamYaple | but not all contaienrs require all those deps | 21:25 |
inc0 | ayoung, I meant to write howto-use-kolla-instead-of-devstack but never got to do it | 21:25 |
ayoung | inc0, well, I might co-author it with you... | 21:25 |
SamYaple | it greatly increases the speed of the build, and reduces the _total_ size ayoung, but its not perfect | 21:25 |
ayoung | SamYaple, yeah, I grok the tradeoff | 21:26 |
inc0 | way I wanted to propose it is to run everything in kolla *besides* service you're working on | 21:26 |
ayoung | OK, source build completed. deploy.... | 21:26 |
inc0 | you can run that normally in terminal and don't have to deal with containers during development | 21:26 |
ayoung | OSError: [Errno 13] Permission denied: '/etc/kolla/mariadb' | 21:26 |
SamYaple | inc0: ayoung is trying to use it to test, but thats how i use it | 21:26 |
ayoung | let me make that... | 21:26 |
SamYaple | ayoung: ok you *may* want to switch to my patch if you are unprivileged user | 21:27 |
ayoung | I should probably chown /etc/kolla ,no? | 21:27 |
SamYaple | otherwise youll want to run it as root hoenstly | 21:27 |
inc0 | or run plays as root | 21:27 |
ayoung | We'll see.... | 21:27 |
SamYaple | ayoung: just chown /etc/kolla as your user | 21:27 |
SamYaple | youll be fine | 21:27 |
ayoung | TASK: [mariadb | Copying over galera.cnf] ************************************* | 21:28 |
ayoung | fatal: [localhost] => {'msg': "One or more undefined variables: 'dict object' has no attribute u'ansible_dummy0'", 'failed': True} | 21:28 |
ayoung | fatal: [localhost] => {'msg': 'One or more items failed.', 'failed': True, 'changed': False, 'results': [{'msg': "One or more undefined variables: 'dict object' has no attribute u'ansible_dummy0'", 'failed': True}]} | 21:28 |
inc0 | for CI testing, yeah that's cool...how many times we've seen broken deps that someone included... | 21:28 |
SamYaple | ayoung: oh right. ok so i left a comment in there | 21:28 |
SamYaple | dummy0 needs to be like eth0 for you | 21:28 |
inc0 | and didnt show up in CI because devstack and something else already installed it | 21:28 |
SamYaple | and kolla_internal_interface should be your ip address on eth0 | 21:28 |
ayoung | Ah in the config | 21:28 |
*** Marga_ has joined #kolla | 21:28 | |
SamYaple | normally kolla_internal_interface is a VIP that keepalive assigns, bt since you are running without keepalived you need to set it to your interface ip | 21:29 |
ayoung | SamYaple, so my host has eth0 as 10.0.0.13 is that what I want for kolla_internal_address | 21:30 |
SamYaple | yes | 21:30 |
SamYaple | 10.0.0.13 is kolla_internal_address | 21:30 |
SamYaple | network_interface: eth0 | 21:30 |
SamYaple | we dont bind to 0.0.0.0 so we need some specfics | 21:31 |
ayoung | actually not eth0. it is enp0s25. Where did that come from? | 21:31 |
SamYaple | i said eth0 before | 21:31 |
SamYaple | but youre a smart fella youll get it done! | 21:31 |
ayoung | Fedora is wwweierded | 21:31 |
*** achanda has quit IRC | 21:31 | |
SamYaple | dnf install biosdevname | 21:32 |
ayoung | SamYaple, OK, so , as I wait for this to deploy. THe git option you posted before builds and image from those files, but they are not linked from the base image, right? They are copied, and if I want to test out a new change I need to regen the keystone image? | 21:32 |
SamYaple | udev and there stupid consitent (inconsistent) naming rules | 21:32 |
*** aginwala has quit IRC | 21:32 | |
SamYaple | ayoung: that is correct | 21:33 |
ayoung | Hmmm...no mariadb running | 21:33 |
SamYaple | has it crashed? | 21:33 |
ayoung | failed: [localhost] => {"attempts": 6, "changed": false, "cmd": ["docker", "exec", "mariadb", "ls", "/var/lib/mysql/mariadb.pid"], "delta": "0:00:00.019984", "end": "2016-02-05 16:32:42.132941", "failed": true, "rc": 1, "start": "2016-02-05 16:32:42.112957", "stdout_lines": [], "warnings": []} | 21:33 |
SamYaple | ah | 21:33 |
inc0 | didnt it got merged? | 21:33 |
inc0 | duh | 21:33 |
SamYaple | so you should see container running | 21:33 |
SamYaple | `docker ps -a` | 21:33 |
*** mbound has joined #kolla | 21:33 | |
ayoung | not running | 21:34 |
SamYaple | do `docker logs mariadb` | 21:34 |
ayoung | rsyslog and kolla_toolbox only | 21:34 |
SamYaple | but it exists? | 21:34 |
ayoung | chown: changing ownership of '/var/lib/mysql': Permission denied | 21:34 |
ayoung | um...you need to stop running as root | 21:34 |
SamYaple | thats not the host | 21:34 |
SamYaple | thats in the container | 21:34 |
ayoung | hmmm | 21:34 |
SamYaple | but like I said, use my patch at this point | 21:35 |
SamYaple | youll be happier | 21:35 |
SamYaple | or run as root | 21:35 |
SamYaple | this in between will only give you headache | 21:35 |
ayoung | OK, I'll grab your patch | 21:35 |
inc0 | SamYaple, can you plz remind me link? | 21:35 |
SamYaple | inc0: https://review.openstack.org/#/c/276887/ | 21:35 |
inc0 | because it's ridiculus, it's there for quite some time | 21:35 |
inc0 | ah not this one | 21:35 |
inc0 | but ok, will review | 21:36 |
ayoung | SamYaple, so now copy /etc/kolla to... | 21:36 |
ayoung | or just rerun as is? | 21:36 |
*** aginwala has joined #kolla | 21:37 | |
SamYaple | ayoung: http://paste.openstack.org/show/486155/ | 21:37 |
SamYaple | youll need to clean your environement | 21:37 |
SamYaple | we have a clean-containers script in the tools/ folder | 21:37 |
ayoung | SamYaple, images are OK? | 21:38 |
*** esikachev has joined #kolla | 21:38 | |
openstackgerrit | Sam Yaple proposed openstack/kolla: Fix non-root deploys https://review.openstack.org/276887 | 21:38 |
SamYaple | ayoung: yes | 21:38 |
SamYaple | ignre that new patch | 21:38 |
SamYaple | it wont affect you | 21:38 |
ayoung | ./tools/kolla-ansible --configdir /home/ayoung/kolla deploy | 21:39 |
ayoung | running now... | 21:40 |
SamYaple | did you clean all containers first? | 21:40 |
SamYaple | and volumes? | 21:40 |
ayoung | not volumes | 21:40 |
ayoung | just containers | 21:40 |
SamYaple | if yo u ran script it should have done both | 21:40 |
SamYaple | if you cleaned up manually volumes probably stuck around | 21:40 |
ayoung | ./tools/cleanup-containers | 21:40 |
SamYaple | yup should have gotten volulmes | 21:40 |
ayoung | still failed at mysql | 21:40 |
ayoung | stderr: ls: cannot access /var/lib/mysql/mariadb.pid: No such file or directory | 21:40 |
SamYaple | yea | 21:41 |
inc0 | docker volume rm $(docker volume ls) | 21:41 |
SamYaple | ayoung: yea volume stuck around :/ | 21:41 |
*** achanda has joined #kolla | 21:41 | |
SamYaple | clean containres and do what inc0 has | 21:41 |
ayoung | ok let me | 21:41 |
SamYaple | you can confirm with `docker volume ls` that volumes are removed | 21:41 |
ayoung | fc7cdb24b442 kollaglue/centos-source-mariadb:2.0.0 "kolla_start" 4 minutes ago Up 4 minutes mariadb | 21:41 |
ayoung | still tehre | 21:41 |
ayoung | let me kill that manually | 21:41 |
inc0 | docker rm $(docker ps -a) && docker volume rm $(docker volume ls) | 21:41 |
SamYaple | lol yea that | 21:41 |
SamYaple | if you want to nuke everything | 21:42 |
inc0 | docker rm -f $(docker ps -a) && docker volume rm $(docker volume ls) | 21:42 |
ayoung | Error response from daemon: Cannot kill container fc7cdb24b442: [2] Container does not exist: container destroyed | 21:42 |
ayoung | Error: failed to kill containers: [fc7cdb24b442] | 21:42 |
ayoung | hmmm | 21:42 |
inc0 | thats disturbing | 21:42 |
ayoung | SamYaple, that won't kill images though, right? | 21:42 |
inc0 | nah, images are safe | 21:42 |
ayoung | OK, I restarted the docker service, container was still there, but now I was able to kill it and ps shows it is gone | 21:43 |
SamYaple | ayoung: correct images are safe, controled by `docker images` | 21:43 |
*** Marga_ has quit IRC | 21:43 | |
ayoung | how do I list volumes? | 21:44 |
*** mbound has quit IRC | 21:44 | |
inc0 | docker volume ls | 21:44 |
SamYaple | `docker volume ls` | 21:44 |
ayoung | man docker does not have volume specified | 21:45 |
inc0 | what do you mean? | 21:45 |
*** esikachev has quit IRC | 21:46 | |
SamYaple | ayoung: yea docker just recently got man pages | 21:46 |
ayoung | tried to open a new terminal and getpt failed: Permission denied | 21:46 |
SamYaple | check your perms on /dev/ptmx | 21:47 |
inc0 | aaa you run kolla and gnome;) | 21:47 |
inc0 | I have no idea why that happends | 21:47 |
SamYaple | inc0: nah i get the same probelm too | 21:47 |
SamYaple | fix perms on /dev/ptmx | 21:47 |
ayoung | $ sudo ls -la /dev/ptmx | 21:47 |
ayoung | lrwxrwxrwx. 1 root root 8 Feb 5 16:45 /dev/ptmx -> pts/ptmx | 21:47 |
SamYaple | ok so check /dev/pts/ptmx :P | 21:47 |
inc0 | yeah I did too | 21:47 |
ayoung | c---------. 1 root root 5, 2 Jan 30 11:42 /dev/pts/ptmx | 21:48 |
ayoung | I assume 700? | 21:48 |
SamYaple | fix that | 21:48 |
SamYaple | yea | 21:48 |
SamYaple | or 777 | 21:48 |
*** aginwala has quit IRC | 21:48 | |
SamYaple | i dont know default | 21:48 |
SamYaple | supposedly thats fixed in docker 1.10 | 21:48 |
ayoung | 777 | 21:48 |
ayoung | 700 still perm denied | 21:48 |
*** Marga_ has joined #kolla | 21:49 | |
*** Marga_ has joined #kolla | 21:49 | |
ayoung | I wonder if that is supposed to be read only...let's see | 21:49 |
*** Marga_ has joined #kolla | 21:50 | |
ayoung | 666 worked | 21:50 |
ayoung | Is that going to be happening a lot? | 21:50 |
SamYaple | ayoung: its a docker bug | 21:50 |
SamYaple | supposedly fixed in 1.10 | 21:51 |
SamYaple | i think it happens at daemon start | 21:51 |
inc0 | and wont happen on servers | 21:51 |
*** aginwala has joined #kolla | 21:51 | |
SamYaple | it happens... randomly | 21:51 |
inc0 | at least never experienced it | 21:51 |
SamYaple | yea and inc0 is right | 21:51 |
ayoung | mariadb failed again | 21:51 |
SamYaple | inc0: binding /dev /for syslog broke docker! | 21:51 |
ayoung | stderr: ls: cannot access /var/lib/mysql/mariadb.pid: No such file or directory | 21:52 |
SamYaple | ayoung: can you check logs again? | 21:52 |
SamYaple | same chown issue? | 21:52 |
ayoung | the funny thing is, I was successful getting the mariadb server to launch with the Fedora docker container | 21:52 |
ayoung | SamYaple, Its running again, one sec | 21:52 |
ayoung | chown: changing ownership of '/var/lib/mysql': Permission denied | 21:53 |
SamYaple | ugh | 21:53 |
SamYaple | so we do not run as root inside the containers ayoung | 21:53 |
SamYaple | but we need _some_ root commands. and thats whats tripping you up | 21:53 |
SamYaple | ayoung: https://github.com/openstack/kolla/blob/master/docker/mariadb/extend_start.sh#L23 | 21:54 |
SamYaple | the container itself has a sudoers entry for that, but i suspect its selinux at play here | 21:54 |
ayoung | SamYaple, let me see what the Fedora one does... | 21:54 |
SamYaple | do you have it set to enforce mode? | 21:54 |
ayoung | SamYaple, selinux? | 21:55 |
SamYaple | ayoung: yep | 21:55 |
ayoung | $ getenforce | 21:55 |
ayoung | Enforcing | 21:55 |
SamYaple | youll have to not.... kolla + selinux wont work until someone writes the profiles | 21:55 |
SamYaple | no one has written the profiles | 21:55 |
ayoung | https://github.com/fedora-cloud/Fedora-Dockerfiles/tree/master/mariadb/scripts | 21:56 |
ayoung | maridb should work ok | 21:56 |
SamYaple | ayoung: this is not related to mariadb | 21:57 |
SamYaple | but rather kolla | 21:57 |
SamYaple | because we are calling sudo in the container we are violating some selinux things | 21:57 |
SamYaple | we need profiles to allow those things | 21:57 |
SamYaple | no one has written those profiles | 21:57 |
SamYaple | thus we need selinux turned off :( | 21:58 |
ayoung | Docker runs as root, why do we need to sudo? | 21:58 |
SamYaple | because the containers run as unprivileged users | 21:58 |
inc0 | we don't run root inside containers | 21:58 |
SamYaple | so if anyone breaks out of a service with an exploit, they are running as unprivileged user | 21:58 |
SamYaple | inc0: ive always hated that drop-root stuff :/ | 21:59 |
inc0 | yeah I don't really get it | 21:59 |
ayoung | SamYaple, so you think if I setenforce permissive it will work? | 21:59 |
SamYaple | ayoung: i do | 21:59 |
ayoung | let me try | 21:59 |
inc0 | openstack services are normally run as root | 21:59 |
inc0 | but put this service inside container and it's security risk | 22:00 |
inc0 | I mean...it's good mindset to remove permissions that aren't needed | 22:00 |
inc0 | but it causes problems | 22:00 |
SamYaple | i agree, but i wanted to get stablier first | 22:01 |
ayoung | Well, I think it is possible to run mysql without needing root permissions | 22:01 |
SamYaple | ayoung: it absolutely is | 22:01 |
SamYaple | thats not the issue | 22:01 |
ayoung | if it listens on a port above 1000 | 22:01 |
SamYaple | so we use named volumes with docker | 22:02 |
SamYaple | those mount in as root | 22:02 |
SamYaple | we can do nothing to change that as of yet | 22:02 |
ayoung | looking good | 22:02 |
SamYaple | but those perms need to change for us to use the named volume | 22:02 |
ayoung | I just saw the Keystone stuff fly past | 22:02 |
SamYaple | hence the sudo | 22:02 |
SamYaple | ayoung: you are free to stop it then | 22:02 |
SamYaple | its not going to do anything else | 22:02 |
ayoung | Horizon failed. No surprise. I adde that to the ignor in globlas, right? | 22:03 |
*** aginwala has quit IRC | 22:03 | |
SamYaple | probably | 22:03 |
SamYaple | next time run with just `kolla-ansible deploy --tags keystone,mariadb` | 22:03 |
SamYaple | it will filter all the rest | 22:03 |
inc0 | so if you have mariadb and stuff deployed | 22:03 |
inc0 | yeah, I wanted to say about tags | 22:03 |
ayoung | I might actually enable rabbit next run. NOtifications from Keystone are sent over the mq | 22:04 |
ayoung | would like to make that part of testing | 22:04 |
*** dims has joined #kolla | 22:04 | |
ayoung | I think I'm good here, except for SELinux. | 22:04 |
*** dimsum__ has quit IRC | 22:04 | |
ayoung | OK, that was fast | 22:04 |
ayoung | how to test? | 22:04 |
ayoung | I want to do openstack token issue | 22:04 |
SamYaple | ayoung: normal openstack ocmmands | 22:05 |
SamYaple | endpoint is at 10.0.0.13 | 22:05 |
ayoung | is there an rc file generated? | 22:05 |
inc0 | netstat -nalp | grep 5000 | 22:05 |
SamYaple | ayoung: yea | 22:05 |
inc0 | is should be there | 22:05 |
SamYaple | kolla-ansible post-deploy | 22:05 |
SamYaple | that will spit out admin-openrc.sh to your config directory | 22:05 |
SamYaple | next to globals.yml | 22:05 |
SamYaple | you can source that then `openstack user list` | 22:05 |
ayoung | um...I think it does not honor the conf dir, though | 22:06 |
ayoung | looking in /etc/kolla, there is a lot there still | 22:06 |
ayoung | including the dev dir... | 22:06 |
ayoung | $ openstack token issue succeeded | 22:07 |
ayoung | OK, thanks guys, I'm going to take this success for today. | 22:07 |
ayoung | SamYaple, inc0 you two going to Austin? | 22:08 |
inc0 | yeah, I actually live 70miles from Austin | 22:08 |
SamYaple | ayoung: indeed | 22:08 |
ayoung | SamYaple, inc0, first beer is on me | 22:08 |
SamYaple | ayoung: i just moved ~1000 miles away! | 22:08 |
*** neilus has joined #kolla | 22:08 | |
*** jtriley has quit IRC | 22:08 | |
inc0 | ayoung, will remember that for sure | 22:09 |
ayoung | inc0, I hope they still let you get a hotel. | 22:09 |
inc0 | I'll get airbnb or sth | 22:09 |
SamYaple | ayoung: yea lots to be desired for non-root stuff. we do no testing on that as of yet. I was hoping we would this cycle but it was not made a priority... I will try to make it one next cycle | 22:09 |
inc0 | I'm staying in Austin for alcohol, that one is certain | 22:09 |
ayoung | OK...so, now that I have this, let me test out the features that just landed in Keystone. | 22:10 |
SamYaple | ayoung: how was the speed of deploy for you? | 22:11 |
ayoung | SamYaple, pretty snappy | 22:11 |
ayoung | I didn;'t time it, but not long enough for me to get bored and go read the news | 22:12 |
SamYaple | cool | 22:12 |
SamYaple | full openstack AIO deploy for me takes under 2 minutes | 22:12 |
SamYaple | thats with prebuilt images of course | 22:12 |
*** neilus has quit IRC | 22:13 | |
dmsimard | ayoung: curious, are you deploying source or binary ? centos or ubuntu ? | 22:14 |
SamYaple | dmsimard: centos source | 22:16 |
dmsimard | ok | 22:17 |
*** bmace has quit IRC | 22:18 | |
*** bmace has joined #kolla | 22:19 | |
ayoung | SamYaple, OK, so what are these "profiles" that need to be written for the SELinux stuff to work? | 22:19 |
SamYaple | might be called contexts ayoung ? | 22:20 |
SamYaple | profiels are for apparmor | 22:20 |
ayoung | SamYaple, you mean SELinux policy? | 22:21 |
SamYaple | that sounds right | 22:21 |
SamYaple | that is right yup | 22:21 |
ayoung | hmm...OK. | 22:21 |
SamYaple | ayoung: https://blueprints.launchpad.net/kolla/+spec/enable-selinux | 22:22 |
ayoung | I'm ot 100% certain what the right thing to do there is. Would you expect that, once the cluster is deployed, I can re-enable selinux and everythign would work? | 22:22 |
SamYaple | nihilifer: you make any progress on this? | 22:22 |
SamYaple | ayoung: yes, once deployed | 22:22 |
ayoung | OK...it hasn;t blown up yet | 22:22 |
SamYaple | not for a full stack | 22:22 |
SamYaple | but just what you deploy it will work | 22:22 |
nihilifer | SamYaple: you mean selinux? | 22:22 |
SamYaple | nihilifer: yea | 22:22 |
nihilifer | then nope | 22:22 |
nihilifer | well, currently the only thing we can do regarding selinux, is adding ":z" suffix to volumes | 22:24 |
SamYaple | what does that do? | 22:24 |
nihilifer | applies svirt_sandbox_file_t flag to the volume | 22:24 |
SamYaple | ok guys I to go move my stuff into my new house | 22:25 |
SamYaple | ill be around later | 22:25 |
nihilifer | in selinux, you need this flag to be applied on the directory to mount it as a volume | 22:25 |
nihilifer | ok | 22:25 |
SamYaple | ping me if you hit anymore issue ayoung, my irc is always ticking | 22:25 |
SamYaple | thanks for the info nihilifer | 22:25 |
nihilifer | np | 22:25 |
ayoung | SamYaple, thanks, will do. Right now I'm going to make sure some keystone features are working. | 22:25 |
ayoung | Probably all I'll get to tonight | 22:25 |
*** aginwala has joined #kolla | 22:26 | |
*** jtriley has joined #kolla | 22:27 | |
inc0 | ayoung, if that helps, I sit right next to Dolph, so if you guys run into problems I can provide quick help | 22:34 |
*** aginwala has quit IRC | 22:39 | |
*** salv-orl_ has joined #kolla | 22:40 | |
*** salv-orlando has quit IRC | 22:43 | |
*** mbound has joined #kolla | 22:45 | |
*** mbound has quit IRC | 22:50 | |
*** aginwala has joined #kolla | 22:51 | |
*** jtriley has quit IRC | 23:02 | |
*** dims_ has joined #kolla | 23:06 | |
*** neilus has joined #kolla | 23:08 | |
*** dims has quit IRC | 23:10 | |
*** neilus has quit IRC | 23:12 | |
*** farid has quit IRC | 23:20 | |
*** aginwala has quit IRC | 23:28 | |
*** aginwala has joined #kolla | 23:31 | |
*** aginwala has quit IRC | 23:34 | |
*** inc0 has quit IRC | 23:37 | |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Dependencies clean up / simplification https://review.openstack.org/272912 | 23:38 |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Use a new per-service format https://review.openstack.org/273843 | 23:38 |
*** stvnoyes has quit IRC | 23:43 | |
*** stvnoyes has joined #kolla | 23:44 | |
*** aginwala has joined #kolla | 23:45 | |
openstackgerrit | Angus Salkeld proposed openstack/kolla-mesos: Use a new per-service format https://review.openstack.org/273843 | 23:45 |
*** aginwala has quit IRC | 23:46 | |
*** aginwala has joined #kolla | 23:46 | |
*** aginwala has quit IRC | 23:53 | |
*** aginwala has joined #kolla | 23:56 | |
*** iotapi322 has quit IRC | 23:59 | |
*** jtriley has joined #kolla | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!