Friday, 2015-07-10

sdakewhich describes excatly how the CLI tools work00:00
redrobotsdake link?00:00
adrian_ottoso I'll record this as a #agreed if we have no objections00:00
sdakeand what we have wored out with openssl and manually setting up certs00:00
adrian_ottoI'm open to hearing alternative points of view before that00:00
sdakeredrobot madhuri has link handy00:00
adrian_ottoand I will bring it up during announcements in our Tuesday team meeting coming up00:00
*** eghobo_ has quit IRC00:01
madhuriThis is in bad shape sorry for that00:02
madhuriBut it does work00:02
madhuriAlso this is the tool we might use to generate certs00:02
madhuriBut that also needs revision00:02
adrian_otto#agreed Magnum will use Barbican for an initial implementation for certificate generation and secure storage/retrieval.  We will cmmit to a second phase of development to eliminating the hard requirement on Barbican with an alternate implementation that implements the functional equivalent implemented in Magnum, which may depend on libraries, but not Barbican.00:03
adrian_ottois that fair? If not, I can undo it.00:03
adrian_ottoany opposed?00:03
adrian_ottook, do you want us to wrap up now, Madhuri?00:04
madhuriYes sure00:04
sdakemight be good to run that by the team on wed00:04
sdakemake sure there are no -2's waiting :)00:04
madhuriThank you all for your time00:04
adrian_ottothanks everyone for your time and attention. This was an important decision for us.00:04
sdaketuesday rather00:04
adrian_ottosdake, I will reference this log in our next Tuesday team meeting.00:05
adrian_ottoI am adding it to the agenda right now.00:05
adrian_ottothanks again!00:05
* adrian_otto wonders if the MeetBot crashed00:05
sdakedi dyou nick change?00:06
sdakeif not then yes i saw it exit the channel00:06
sdakeso it probably crashed00:06
madhuriredrobot, What is the good time to discuss about Barbican CA plugin?00:06
sdakethere is a global irc log for this channnel as well00:06
sdakemight as well reference that00:06
adrian_ottook, will do.00:07
adrian_otto#startmeeting containers_tls00:07
openstackMeeting started Fri Jul 10 00:07:23 2015 UTC and is due to finish in 60 minutes.  The chair is adrian_otto. Information about MeetBot at
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.00:07
openstackThe meeting name has been set to 'containers_tls'00:07
openstackMeeting ended Fri Jul 10 00:07:30 2015 UTC.  Information about MeetBot at . (v 0.1.4)00:07
openstackMinutes (text):
adrian_ottothat's a confirmed crash ;-)00:07
*** sdake_ has joined #openstack-containers00:07
*** sicarie has left #openstack-containers00:10
Kennan2adrian_otto: ping00:10
*** Kennan2 is now known as Kennan00:11
adrian_ottoKennan2: here00:11
madhuriredrobot, ?00:11
*** sdake has quit IRC00:11
Kennanhi adrian_otto: Could you check my comments on ?00:11
Kennanwe used that platform in magnum-template-manage for long time00:11
Kennando you mean we have to create bug to fix that cmd ?00:12
KennanI just followed old design :_00:12
adrian_ottoyes, I really hate taht00:12
adrian_ottosorry I let that slip through00:12
adrian_ottolet's open a bug to change the term00:13
KennanOK adrian_otto: what do you think the proper term?00:13
Kennanbetter not long :)00:13
adrian_ottoI'm open to suggestions… I have thought about it a bit00:13
adrian_ottonova_instance_type is probably the most explicit term we could use00:14
*** hongbin has joined #openstack-containers00:14
*** eghobo has quit IRC00:14
*** jruano has joined #openstack-containers00:14
adrian_ottowith suitable values < vm | ironic >00:14
adrian_ottoand we might later add docker or libvirt_lxc00:15
adrian_ottoKennan: does that make sense to you?00:15
KennanI am thinking if nova_instance_type is too binding to openstack, as you said we support native docker, is it ?00:16
adrian_ottothe other suggestion I offered was compute_form_factor00:17
adrian_ottowhich makes sense to me because internally at Rackspace that's how we refer to the different virt driver types00:17
Kennanadrian_otto: nova_instance_type seems OK, but not general. compute_form_factor: not figure it out for me :)00:17
adrian_ottobut I'm not sure if that's meaningful to others00:17
*** eghobo has joined #openstack-containers00:17
adrian_ottocompute_instance_type ?00:17
*** sdake has joined #openstack-containers00:18
adrian_ottoanother option is to *not* define this in the bay at all00:18
adrian_ottoand to put it into the glance image metadata00:18
adrian_ottowe already have a function for getting that as a dict, I think00:18
KennanI did not think can put in glance image metadata00:19
Kennanas same image can be used in ironic and vm00:19
Kennanit not have differenc00:19
adrian_ottogood point00:19
adrian_ottoit could be on the flavor00:19
adrian_ottobut I don't know if flavors have metadata without extensions added to nova00:20
KennanI still think better in baymodel now, as our coe related iterm bind to that00:20
adrian_ottook, we can do taht00:20
adrian_ottobut once I saw platform in there I gagged and gasped00:20
adrian_ottoI thought "Does that mean hardware platform? Why do we care about that?"00:21
Kennanadrian_otto: maybe nova_instance_type can be the option now, if we not have better item. (we can change that later if had)00:21
KennanI could think it today00:21
Kennanif any other options00:21
adrian_ottook, let's file a bug asking for it to be adjusted to nova_instance_type00:21
*** sdake_ has quit IRC00:21
adrian_ottoand we can revise that if we think of something better00:22
adrian_ottodo you want me to open the bug, or can you?00:22
Kennanadrian_otto: I could open bug for that00:22
*** mfalatic has quit IRC00:26
*** suro-patz1 has joined #openstack-containers00:33
*** suro-patz has quit IRC00:33
*** suro-patz1 has quit IRC00:34
*** jay-lau-513 has quit IRC00:36
*** adrian_otto has quit IRC00:37
*** diga has quit IRC00:37
*** eghobo has quit IRC00:38
*** dims has quit IRC00:40
*** ameybhide has quit IRC00:44
*** yuanying-alt has joined #openstack-containers00:45
*** h00327910_ has joined #openstack-containers00:49
*** yuanying-alt has quit IRC00:49
*** lan__ has quit IRC00:54
*** Kennan has quit IRC00:54
*** lan__ has joined #openstack-containers00:55
*** Drago has quit IRC00:57
*** Kennan has joined #openstack-containers00:58
*** yuanying_ has joined #openstack-containers01:02
*** yuanying has quit IRC01:05
*** yuanying_ has quit IRC01:07
*** mmasaki_ has joined #openstack-containers01:12
eliqiaoadrian_otto:  ping01:21
*** erkules_ has joined #openstack-containers01:22
*** rbrooker has quit IRC01:22
*** suro-patz has joined #openstack-containers01:24
*** yuanying has joined #openstack-containers01:24
eliqiaomadhuri: around ?01:24
*** dims has joined #openstack-containers01:24
*** erkules has quit IRC01:25
*** dims has quit IRC01:25
*** dims has joined #openstack-containers01:26
*** yuanying has quit IRC01:28
*** sdake_ has joined #openstack-containers01:29
*** yuanying has joined #openstack-containers01:29
*** yuanying has quit IRC01:30
*** sdake has quit IRC01:32
*** yuanying has joined #openstack-containers01:33
*** SourabhP has quit IRC01:35
*** saksham has quit IRC01:36
*** dims has quit IRC01:36
*** dims has joined #openstack-containers01:37
*** sdake has joined #openstack-containers01:38
*** tobe has joined #openstack-containers01:39
*** sdake_ has quit IRC01:41
*** dims has quit IRC01:42
madhurieliqiao, hi, what's up?01:51
*** wanghua has joined #openstack-containers01:57
*** kebray has quit IRC01:58
eliqiaomadhuri: it's re
eliqiaomadhuri: can you explain why we need to leave the bay's status as it is even the stack in heat has beed deleted ?01:59
madhuriYes about deleting the non-existing bay?01:59
eliqiaomadhuri, yes02:00
madhuriOk periodic tasks is internal task i.e. user unaware of02:00
madhuriSuch task should just update the status02:00
madhuriBut not add or delete any change/resource to system that can affect user02:00
madhuriSo in our case, if we delete the bay in periodic task that is not deleted by user.02:01
madhuriUser might not know when is their bay deleted and that is bad I guess02:02
eliqiaomadhuri let't think about this scenario, user triggering a delete operation, bay's status changed to deleting, but before it finished, m-cond restart, so it will always in deleteing.......02:02
madhuriBay is user resource and they should only decide what can be done to it02:02
madhuriAgree, then user can delete the bay again in such case02:03
eliqiaomadhuri then I am thinking if we can check if it's current status is deleting then if we found heat stack was deleted in periodic task, we delete it.02:04
eliqiaowould that make sence?02:04
*** kan__ has joined #openstack-containers02:05
madhuriYes in case only when state is DELETE_IN_PROGRESS02:05
madhuriNot when state is CREATE_COMPLETE02:05
eliqiaomadhuri, okay , I will update patch thx02:06
madhuriOk thanks02:06
eliqiaobeside, do you understand  adrian_otto 's comments?02:06
madhuriYes I guess02:06
*** mmasaki_ has quit IRC02:07
madhuriPeiodic task need some authentication and you are using admin_context02:07
madhuriwhich is not preferrable.02:07
madhuriSo for tasks like this which are triggered by internal services, we use TRUST of keystone02:08
madhuriHeat also have the same case02:08
madhuriThey are also using trust02:08
madhuriJust give me a sec, I will get you a link that might help you understand trust02:08
eliqiaomadhuri, that would be greate02:09
eliqiaoI see from the code, it will also use admin creds which configured in magnum.conf..02:10
madhuriIt is using trust id and token02:10
madhuriwhich is safe02:10
eliqiaobut I don't understand where is  trust id  and token from02:10
eliqiaothe context is a fake one02:11
eliqiaothere is no user creds02:11
madhuriOk that link would help you understand what trust is02:11
madhuriPlease go through it and let me know if I can help02:12
eliqiaomadhuri: okay, thanks I will read that link first, thanks for your help :)02:12
madhuriYou're welcome eliqiao :)02:12
eliqiaomadhuri: that link is from JP?02:13
eliqiaoI face network issue, can not open it.02:13
madhuriJust a sec02:13
*** harshs has quit IRC02:13
eliqiaois it the same ?02:14
*** mmasaki has joined #openstack-containers02:14
madhuriYes the same02:14
eliqiaomadhuri:  okay , cool thanks02:14
*** mmasaki has quit IRC02:19
*** mmasaki has joined #openstack-containers02:19
*** achanda has joined #openstack-containers02:23
eliqiaomadhuri: looks like we need to store trust_id in DB ?02:31
madhuriYes you will02:31
madhuriBut for that heat is using Barbican02:31
madhurieliqiao, Not yet supported. But they plan to use Barbican02:32
eliqiaomadhuri: you mean magnum doesn't support Barbican yet, and plan to use Barbican?02:33
*** yuanying-alt has joined #openstack-containers02:33
madhuriI said Heat is working on supporting Barbican to store secrets02:34
eliqiaomadhuri: get it.02:35
madhuriAlso Magnum will also be supporting Barbican to store certificates in TLS support02:35
madhuriBut not a hard dependency02:35
*** unicell has quit IRC02:35
*** yuanying-alt has quit IRC02:38
*** dims has joined #openstack-containers02:38
*** jruano has quit IRC02:39
*** suro-patz has quit IRC02:41
*** achanda has quit IRC02:42
*** dims has quit IRC02:43
*** Kennan has quit IRC03:07
*** Kennan has joined #openstack-containers03:07
*** jruano has joined #openstack-containers03:30
*** kebray has joined #openstack-containers03:36
*** hongbin has quit IRC03:39
*** jruano has quit IRC04:07
*** suro-patz has joined #openstack-containers04:13
*** unicell has joined #openstack-containers04:17
*** achanda has joined #openstack-containers04:20
*** yuanying-alt has joined #openstack-containers04:22
*** yuanying-alt has quit IRC04:27
*** suro-patz has quit IRC04:37
*** tobe has quit IRC04:44
*** suro-patz has joined #openstack-containers04:55
*** sdake_ has joined #openstack-containers04:58
*** sdake has quit IRC05:00
*** sdake has joined #openstack-containers05:00
*** dims has joined #openstack-containers05:03
*** fawadkhaliq has joined #openstack-containers05:03
*** sdake_ has quit IRC05:03
*** dims_ has joined #openstack-containers05:04
*** dims has quit IRC05:08
*** dims_ has quit IRC05:08
*** kebray_ has joined #openstack-containers05:21
*** Marga_ has joined #openstack-containers05:23
*** Marga_ has quit IRC05:24
*** Marga_ has joined #openstack-containers05:25
*** kebray has quit IRC05:25
*** adrian_otto has joined #openstack-containers05:30
*** david-ly_ has joined #openstack-containers05:37
*** david-lyle has quit IRC05:40
eliqiaohi adrian_otto05:42
*** SourabhP has joined #openstack-containers05:42
adrian_ottohello eliqiao05:44
eliqiaoadrian_otto: for , thanks for reviewing and I still have some questions about how to landing it.05:44
*** SourabhP_ has joined #openstack-containers05:45
eliqiaomadhuri shows me a link about trust id usage. but looks like we need a new table to store user_creds ?05:46
eliqiaoadrian_otto:  is that what you want ?05:47
*** SourabhP has quit IRC05:47
*** SourabhP_ is now known as SourabhP05:47
adrian_ottoyes, we would need that, or if we do pull in the Barbican dependency, we could rely on that to hold the trusts05:47
adrian_ottothat approach would conform more closely with the principle of least privelidge05:48
adrian_ottodoes that phrase mean something to you?05:48
adrian_ottoI don't want to make assumptions05:49
*** ig0r__ has joined #openstack-containers05:52
eliqiaoadrian_otto: sorry , don't get you.05:53
eliqiaowhich one do you prefer much more ? use table to store user_creds or Barbican?05:54
adrian_ottook, so the principle of least privilege is about only granting the minimum access level necessary to complete a particular task05:54
adrian_ottoso ideally, if we have a background job that sweeps the floor, we only give it permission to sweep, and no other verbs, or nouns05:55
adrian_ottodoes that much make sense?05:55
eliqiaoadrian_otto:  yeah, get it05:55
*** ig0r_ has quit IRC05:55
eliqiaoadrian_otto: one thing I am not very clear is, I need to query all bays in db? if we are not using an admin context, I can not query anything.05:56
adrian_ottook, so I'm suggesting we think about this a bit differently05:57
adrian_ottoif you imagine a background job that runs in a security context of a single individual tenant…05:58
adrian_ottoit can access all the bays belonging to that tenant, but not bays belonging to other tenants05:58
*** dboik has joined #openstack-containers05:58
adrian_ottoso to have something working for all tenants, we would either need one of those for each tenant… or...05:59
adrian_ottohave one that can create subtasks, each that set context to a single tenant, and remain that way until they are ended.05:59
adrian_ottodoes that part make sense?05:59
eliqiaoadrian_otto:  sound good, I will think more about it. thanks06:00
adrian_ottook, now not all of my ideas are good ones06:00
adrian_ottosometimes achieving these sorts of things is just too much of a stretch06:00
adrian_ottoso if you can't think of a great way to do this, I'm happy to reconsider the admin task you proposed in your first patch06:01
*** dims has joined #openstack-containers06:01
adrian_ottoso please don't see this as a rejection of the proposal, but a question about how we might make it better06:01
eliqiaoadrian_otto: get it. thanks, good to konw, yes, we should make things better.06:02
adrian_ottothe essential question is: "how can we do this without a "god" background job?"06:02
adrian_ottocool, thanks. I look forward to what ideas you come up with06:02
*** Kennan has quit IRC06:02
eliqiaoadrian_otto: yeah ...  thanks06:02
*** Kennan has joined #openstack-containers06:03
*** dboik has quit IRC06:03
*** Kennan2 has joined #openstack-containers06:06
*** Kennan has quit IRC06:07
*** dims has quit IRC06:09
*** suro-patz has quit IRC06:10
*** achanda has quit IRC06:11
*** yuanying-alt has joined #openstack-containers06:11
*** dims has joined #openstack-containers06:13
*** suro-patz has joined #openstack-containers06:14
*** dims_ has joined #openstack-containers06:14
*** erkules_ is now known as erkules06:15
*** erkules has joined #openstack-containers06:15
*** yuanying-alt has quit IRC06:16
*** dims has quit IRC06:17
*** dims_ has quit IRC06:21
*** dims has joined #openstack-containers06:22
*** dims has quit IRC06:27
*** adrian_otto has quit IRC06:30
*** suro-patz has quit IRC06:30
*** SourabhP has quit IRC06:34
*** sthillma has joined #openstack-containers06:36
*** kebray_ has quit IRC06:40
*** tobe has joined #openstack-containers06:45
*** sthillma has quit IRC07:02
openstackgerritKai Qiang Wu(Kennan) proposed openstack/magnum: Fix the wrong platform usage
*** j___ has quit IRC07:12
*** BertrandN has joined #openstack-containers07:15
*** BertrandN has quit IRC07:19
*** humble_ has joined #openstack-containers07:19
*** fawadkhaliq has quit IRC07:21
*** fawadk has joined #openstack-containers07:21
openstackgerritKai Qiang Wu(Kennan) proposed openstack/magnum: Add ironic support in magnum server side
openstackgerritKai Qiang Wu(Kennan) proposed openstack/magnum: Add ironic support in magnum server side
openstackgerritKai Qiang Wu(Kennan) proposed openstack/python-magnumclient: Add nova_instance_type support in client side
*** humble_ has quit IRC07:35
*** BertrandN has joined #openstack-containers07:36
openstackgerritKai Qiang Wu(Kennan) proposed openstack/python-magnumclient: Add nova_instance_type support in client side
*** alex_klimov has joined #openstack-containers07:39
*** j___ has joined #openstack-containers07:40
*** kan__ has quit IRC07:52
*** yuanying-alt has joined #openstack-containers08:00
*** yuanying-alt has quit IRC08:04
*** yuanying has quit IRC08:10
*** coolsvap|afk is now known as coolsvap|away08:26
*** fawadk has quit IRC08:27
*** coolsvap|away is now known as coolsvap08:28
*** fawadkhaliq has joined #openstack-containers08:35
*** dims has joined #openstack-containers08:40
*** Marga__ has joined #openstack-containers08:44
*** Marga_ has quit IRC08:45
*** dims has quit IRC08:45
*** Tango has quit IRC08:50
*** lan__ has quit IRC08:59
*** lan__ has joined #openstack-containers09:00
*** fawadkhaliq has quit IRC09:04
*** fawadkhaliq has joined #openstack-containers09:05
openstackgerritYusaku Sawai proposed openstack/magnum: Add `sudo` before `docker` command on dev docs
*** fawadkhaliq has quit IRC09:23
openstackgerritKai Qiang Wu(Kennan) proposed openstack/magnum: Fix unit test failure
*** Marga__ has quit IRC09:32
*** Marga_ has joined #openstack-containers09:32
Kennan2hi all , all units failed because of new relase mock, seems old mock not find something failed silent, we need to fix that09:33
Kennan2this is the fix09:33
*** fawadkhaliq has joined #openstack-containers09:36
*** yuanying-alt has joined #openstack-containers09:48
*** yuanying-alt has quit IRC09:53
*** unicell has quit IRC10:11
*** unicell has joined #openstack-containers10:11
*** j___ has quit IRC10:13
*** j___ has joined #openstack-containers10:13
*** liangbo has joined #openstack-containers10:23
*** alex_klimov has quit IRC10:27
*** dims has joined #openstack-containers10:28
*** dims_ has joined #openstack-containers10:29
*** dims has quit IRC10:33
*** dims_ has quit IRC10:34
*** alex_klimov has joined #openstack-containers10:37
*** zhenguo has quit IRC10:41
*** coolsvap is now known as coolsvap|away10:43
*** fawadkhaliq has quit IRC11:10
*** tobe has quit IRC11:15
*** tobe has joined #openstack-containers11:16
*** fawadkhaliq has joined #openstack-containers11:22
*** yuanying-alt has joined #openstack-containers11:37
*** yuanying-alt has quit IRC11:42
*** fawadkhaliq has quit IRC11:44
*** tobe has quit IRC11:45
*** slagle has joined #openstack-containers11:45
*** fawadkhaliq has joined #openstack-containers11:47
openstackgerritYusaku Sawai proposed openstack/magnum: Add `sudo` before `docker` command on dev docs
*** absubram has joined #openstack-containers12:08
*** absubram has quit IRC12:10
*** absubram has joined #openstack-containers12:10
*** fawadkhaliq has quit IRC12:15
*** dims has joined #openstack-containers12:15
*** ig0r__ has quit IRC12:17
*** dims has quit IRC12:20
*** ig0r_ has joined #openstack-containers12:23
*** david-lyle has joined #openstack-containers12:28
*** jruano has joined #openstack-containers12:29
*** julim has joined #openstack-containers12:29
*** david-ly_ has quit IRC12:30
*** liangbo has quit IRC12:34
*** dims has joined #openstack-containers12:41
*** absubram has quit IRC12:43
*** coolsvap|away has quit IRC12:47
*** kbyrne has quit IRC12:47
*** sdake_ has joined #openstack-containers12:50
*** coolsvap|away has joined #openstack-containers12:51
*** coolsvap|away is now known as coolsvap12:53
*** sdake has quit IRC12:54
*** jjfreric has joined #openstack-containers13:04
*** jjfreric has quit IRC13:11
*** jjfreric has joined #openstack-containers13:12
openstackgerritZachary Sais proposed openstack/magnum: Add .DS_Store to .gitignore
*** fawadkhaliq has joined #openstack-containers13:18
openstackgerritMerged openstack/magnum: Fix unit test failure
*** jjfreric has quit IRC13:18
*** yuanying-alt has joined #openstack-containers13:18
*** dane_leblanc has joined #openstack-containers13:22
*** dboik has joined #openstack-containers13:27
*** dboik_ has joined #openstack-containers13:28
*** dboik has quit IRC13:31
*** dave-mccowan has joined #openstack-containers13:32
openstackgerritHongbin Lu proposed openstack/magnum: Eliminate mutable default arguments
*** jjfreric has joined #openstack-containers13:35
*** liangbo has joined #openstack-containers13:36
*** jjlehr has joined #openstack-containers13:41
*** sdake_ has quit IRC13:51
*** fawadkhaliq has quit IRC13:54
*** yuanying-alt has quit IRC13:55
*** liangbo has quit IRC13:59
*** Tango has joined #openstack-containers14:01
*** kbyrne has joined #openstack-containers14:12
*** j___ has quit IRC14:16
*** liangbo has joined #openstack-containers14:18
*** wanghua has quit IRC14:23
*** hongbin has joined #openstack-containers14:26
*** kbyrne has quit IRC14:36
*** dims has quit IRC14:40
*** zsais has joined #openstack-containers14:40
*** eliqiao has quit IRC14:42
*** eliqiao has joined #openstack-containers14:42
*** coolsvap is now known as coolsvap|away14:50
openstackgerritMerged openstack/magnum: Eliminate mutable default arguments
*** jjfreric1 has joined #openstack-containers14:59
*** jjfreric has quit IRC15:01
*** dims has joined #openstack-containers15:07
*** dims_ has joined #openstack-containers15:07
*** dims has quit IRC15:11
*** liangbo has quit IRC15:13
*** kebray has joined #openstack-containers15:13
openstackgerritTon Ngo proposed openstack/magnum: Instruction for building Fedora Atomic image
*** Drago has joined #openstack-containers15:18
*** ig0r__ has joined #openstack-containers15:19
*** sdake has joined #openstack-containers15:20
*** kbyrne has joined #openstack-containers15:22
*** ig0r_ has quit IRC15:22
*** absubram has joined #openstack-containers15:28
*** harshs has joined #openstack-containers15:30
*** Drago has quit IRC15:32
*** Drago has joined #openstack-containers15:32
*** BertrandN has quit IRC15:40
*** harshs has quit IRC15:49
openstackgerritZachary Sais proposed openstack/magnum: Add .DS_Store to .gitignore
*** achanda has joined #openstack-containers15:54
*** mfalatic has joined #openstack-containers15:55
*** mfalatic_ has joined #openstack-containers15:56
*** mfalatic has quit IRC16:00
*** Drago1 has joined #openstack-containers16:05
*** Drago1 has quit IRC16:05
*** Drago1 has joined #openstack-containers16:05
*** Drago has quit IRC16:05
*** unicell has quit IRC16:08
*** alex_klimov has quit IRC16:14
*** adrian_otto has joined #openstack-containers16:17
sdakehey adrian_otto16:17
*** sdake_ has joined #openstack-containers16:19
*** Marga_ has quit IRC16:21
adrian_ottohi sdake16:22
*** sdake has quit IRC16:22
*** Marga_ has joined #openstack-containers16:26
sdake_yay for new toys : ;)16:28
*** kebray has quit IRC16:32
adrian_ottosdake, you got a retro record player?16:32
*** unicell has joined #openstack-containers16:34
*** adrian_otto has quit IRC16:38
*** adrian_otto has joined #openstack-containers16:39
*** sdake_ has quit IRC16:43
*** harshs has joined #openstack-containers16:43
*** sdake has joined #openstack-containers16:45
*** dims_ has quit IRC16:46
*** dims has joined #openstack-containers16:46
*** dims has quit IRC16:47
*** dims has joined #openstack-containers16:47
mfalatic_adrian_otto - on my doc review, any input on the Mesos verbiage you requested?16:50
*** mfalatic_ is now known as mfalatic16:51
adrian_ottocan you link me to that comment?16:57
adrian_otto^^ mfalatic16:57
adrian_ottoI'm happy to elaborate16:57
mfalaticThis is the review:
mfalaticlet me see if I can get you the exact comment.16:58
*** kebray has joined #openstack-containers16:59
adrian_ottohas the mesos bay type code been merged yet?17:00
mfalaticThat's where it came up. Not aware of whether mesos code is merged yet. Suggestion: we punt the verbiage to a future bugfix.17:01
*** bitblt has joined #openstack-containers17:01
mfalatic(I'm happy to add it if it's handy though)17:01
mfalaticbut if it requires more investigation then it can readily be added later.17:02
adrian_ottoI added a comment suggesting to skip it and file a tech-debt bug to come back later an add that17:03
adrian_ottoideally it should come from the contributors who best understand that feature so it is correctly characterized.17:03
*** suro-patz has joined #openstack-containers17:12
mfalaticI thoroughly agree.17:17
mfalaticI will be making the remaining updates today.17:17
mfalaticI'm curious... the way I abstracted the variables... do people generally like or dislike that? I always do that in my own notes to reduce typos, but if it doesn't read well I can revert that.17:18
hongbinadrian_otto: For mesos bay type, there is one patch pending on review. After that, all the code is merged.17:19
*** sdake_ has joined #openstack-containers17:19
mfalatic  Starting around line 196...17:19
mfalaticI'm tempted to revert that, just because I can see people disliking it.17:20
adrian_ottothanks hongbin17:21
adrian_ottomfalatic: the drawback of using variables is that it takes a lot more vertical space to express an example17:21
adrian_ottothe advantage is that it's a bit easier for users to substitute their own values.17:21
adrian_ottobut they could still do that by editing their CLI commands as they enter them17:22
adrian_ottowhichever way we do it, we should use a consistent style to express all our examples17:22
*** sdake has quit IRC17:23
adrian_ottoif I were you, mfalatic, I would revert that, use the prevailing style, and if you feel strongly submit another patch that changes that style everywhere, and we can debate the merits of that separately from the other improvements in the current proposal.17:23
adrian_ottofor future reference, this is one of the key drawbacks of a big patch17:24
adrian_ottowe will -1 vote for disagreeing with one small part of it, which holds up merging the good parts we like17:24
adrian_ottoso it causes more revisions to work through those various issues, and takes longer to get the improvements merged17:24
adrian_ottothis particular patch is a bit over 2X the size of my suggestion for the ideal patch size, which is about 200 lines of changes17:26
adrian_ottothat's why I'm referring to it as "Big"17:26
*** zsais has quit IRC17:27
mfalaticIt wasn't really meant to be big. :-D17:27
adrian_ottoI know ;-)17:27
adrian_ottothat's why I tried not to make a big deal about it17:28
mfalaticAs for reverting, it's trivial to do for that.17:28
mfalaticIt'd be easier if the dev manual and quickstart were all in one doc.17:29
mfalaticthen one could reference vars at the top alone. But... meh. I think this is unlike the way the docs are usually presented anyway so reverting.17:29
mfalatic(I'm guessing about 1/4-1/3 of this patch is simply fixing pre-existing problems (long lines, typos, little things)17:30
mfalaticmaybe more. anyway, looking forward to being done with it. :-)17:31
mfalaticit'll definitely be more user friendly in the end17:31
*** saksham has joined #openstack-containers17:37
openstackgerritMerged openstack/magnum: Add .DS_Store to .gitignore
*** eghobo has joined #openstack-containers17:42
*** sthillma has joined #openstack-containers17:46
*** ameybhide has joined #openstack-containers17:53
*** sdake has joined #openstack-containers17:55
*** Tango has quit IRC17:58
*** sdake_ has quit IRC17:59
*** rpothier has joined #openstack-containers18:01
*** sdake_ has joined #openstack-containers18:12
*** coolsvap|away is now known as coolsvap18:13
*** sdake has quit IRC18:16
*** sdake_ is now known as sdake18:25
*** dims has quit IRC18:29
*** Marga_ has quit IRC18:32
*** Marga_ has joined #openstack-containers18:32
*** Tango has joined #openstack-containers18:34
*** jjfreric1 has quit IRC18:34
*** PaulCzar has quit IRC18:39
*** jjfreric has joined #openstack-containers18:46
*** harlowja_ has quit IRC18:46
*** harlowja has joined #openstack-containers18:50
*** logan2 has quit IRC18:51
*** logan2 has joined #openstack-containers18:52
*** suro-patz has quit IRC18:53
*** dane_leblanc has quit IRC18:59
*** suro-patz has joined #openstack-containers18:59
openstackgerritOpenStack Proposal Bot proposed openstack/magnum: Updated from global requirements
*** dims has joined #openstack-containers19:12
mfalaticadrian_otto et. al, one last question: my doc fix seems to fix at least one other reported issue. Should I add that to the changelog or will that potentially complicate this unnecessarily?19:13
adrian_ottoyou can just close that other issue with a link to the merged patch19:14
adrian_ottoor I can close it if you point it out to me (in the event you don't have the access to do it)19:14
openstackgerritMartin Falatic proposed openstack/magnum: Updated Magnum documentation
*** achanda has quit IRC19:16
*** dims has quit IRC19:17
*** dims has joined #openstack-containers19:17
openstackgerritMartin Falatic proposed openstack/magnum: Updated Magnum documentation
mfalatic(Caught something not obvious in the git diff view)19:18
*** ig0r__ has quit IRC19:19
*** ig0r_ has joined #openstack-containers19:19
mfalaticThanks to the reviewers for the help in getting this doc updated!19:23
*** saksham has quit IRC19:35
*** kebray has quit IRC19:36
*** dane_leblanc has joined #openstack-containers19:38
mfalaticadrian_otto - I'm thinking my bug resolves (it should be clear that PUBLIC_INTERFACE is a minimum requirement)19:40
openstackLaunchpad bug 1446777 in Magnum "docs should mention that PUBLIC_INTERFACE is required in local.conf" [Medium,Triaged] - Assigned to Vilobh Meshram (vilobhmm)19:40
mfalaticadrian_otto - also, is furthered by my changes (there was a stray UUID reference in the docs)19:41
openstackLaunchpad bug 1454414 in Magnum "Use environment variable for UUID" [Undecided,Fix committed] - Assigned to hongbin (hongbin034)19:41
adrian_ottomfalatic: are you able to update those as we discussed?19:42
mfalaticI'm not quite sure how to make that update. I'm happy to try though... just need to know what the proper way is.19:42
mfalatic(and I wanted to throw them out here first in case people think these aren't helped by what I changed)19:43
mfalaticthe latter one 1454414 has a fix already merged, but I seem to have caught a stray UUID ref that escaped it.19:44
mfalaticnot sure when or how these actually get closed... not sure why that one is still open for example.19:44
mfalatic(fix_committed vs fix_released)19:44
*** wznoinsk has quit IRC19:44
*** dane_leblanc has quit IRC19:45
mfalaticactually, the other one targets k3... hmm.19:45
mfalaticadrian_otto I guess the best way to describe this is "it's complicated"... I just want to do it right.19:46
adrian_ottojust put a comment on each of those bugs, and repeat the links here when done. I will do the administration needed to process them.19:47
adrian_ottobe sure to include a link to the review that solves the issue(s)19:47
adrian_ottomfalatic: the difference between fix_committed and fix_released is only that I have not yet tagged a release that includes them19:49
adrian_ottofix_committed means the fix is merged in master19:49
adrian_ottothat's "closed" from a Magnum developer perspective19:50
*** dane_leblanc has joined #openstack-containers19:52
*** sdake_ has joined #openstack-containers19:52
mfalaticOK, I took a closer look - the UUID thing seems to be somewhat different from what I thought.19:55
mfalaticso that is unaffected.19:55
mfalaticthe other one, I added a comment but it's triaged to k3...19:55
openstackLaunchpad bug 1446777 in Magnum "docs should mention that PUBLIC_INTERFACE is required in local.conf" [Medium,Triaged] - Assigned to Vilobh Meshram (vilobhmm)19:55
*** sdake has quit IRC19:56
*** adrian_otto has quit IRC19:58
dane_leblancIs anyone familiar with the kubernetes / redis example in the developers quick-start guide? #link
mfalaticWhat's the question?20:04
*** kebray has joined #openstack-containers20:05
dane_leblancEverything seems to work for me except the last couple of steps...20:05
mfalaticThe part where you start it?20:05
dane_leblancWhere it says to do a redis-cli on the second minion20:05
dane_leblancThis line: docker exec -i -t $REDIS_ID redis-cli20:06
dane_leblancI can do the redis-cli on the master minion,20:06
dane_leblancbut redis-cli on the second (replicated) minion yeilds "Could not connect to Redis at Connection refused"20:07
mfalaticHmm. Haven't tried it but I can see if I can reproduce it.20:07
dane_leblancWondering if I need a -h <host-ip>20:07
mfalaticthis is all on one machine?20:07
mfalatic(one physical machine)20:08
dane_leblancYes, single devstack node20:08
dane_leblancI'm using 2 minions in the bay20:08
*** harlowja has quit IRC20:09
mfalaticwill give it a try.20:09
*** Marga__ has joined #openstack-containers20:09
*** Marga__ has quit IRC20:09
*** harlowja has joined #openstack-containers20:09
*** Marga__ has joined #openstack-containers20:10
dane_leblancHere is what I'm seeing: #link
*** Marga_ has quit IRC20:10
*** jruano has quit IRC20:13
*** bitblt has quit IRC20:14
mfalaticworking on it, almost to that step20:16
*** prad has joined #openstack-containers20:17
mfalaticHmm. Master fails for me: REDIS_ID=$(docker ps | grep redis:v1 | grep k8s_master | awk '{print $1}')20:19
mfalatictime="2015-07-10T20:17:54Z" level="fatal" msg="Get http:///var/run/docker.sock/v1.17/containers/json: dial unix /var/run/docker.sock: permission denied. Are you trying to connect to a TLS-enabled daemon without TLS?"20:19
*** prad has quit IRC20:19
dane_leblancmfalatic: Yeah, you have to use 'sudo docker'20:19
mfalaticthat's a bug. need to fix the docs.20:19
*** suro-patz has quit IRC20:19
*** Marga_ has joined #openstack-containers20:19
*** Marga_ has joined #openstack-containers20:20
mfalaticHmm docker ps is not returning anything on the master.20:21
openstackgerritOpenStack Proposal Bot proposed openstack/magnum: Updated from global requirements
dane_leblancI had to wait about 12 minutes before the pods were started20:21
mfalaticmy bad20:22
*** Marga__ has quit IRC20:22
dane_leblancYou've done the pod-create and service-create magnum commands already, right?20:22
dane_leblancTry 'sudo docker ps' on one of the minions20:24
mfalaticHow did you monitor the pods to see if they were ready?20:24
*** Marga__ has joined #openstack-containers20:24
dane_leblancAfter doing the service-create, I did 'sudo docker ps' on the minions (not the kube master node)20:25
*** Marga_ has quit IRC20:26
dane_leblancAt first, there was just a k8s_POD.**** pod, but after 12 minutes, 2 more showed up with names k8s_master.*** and k8s_sentinel.***20:26
mfalaticnothing yet on the minions20:26
mfalaticI'm pretty sure I did it right. hmm.20:27
dane_leblancOn the master, do you see anything with 'kubectl get minions'?20:27
mfalaticnote: redis-master is showing as "pending" in pod-show20:28
mfalaticyes, kubectl get minions shows two minions.20:28
mfalaticclaims they are ready20:28
mfalaticbut the IPs are not what I expect to see from the host.20:29
dane_leblancIPs should be the fixed (private) IPs assigned to the openstack VMs for each minion20:29
mfalaticthey appear to be so.. 10.0.0.x20:30
dane_leblanc(not the associated floating IP that you use to SSH into the minions)20:30
mfalaticI can't route to 10.0.0.x from the host though, if that matters.20:30
dane_leblancI see the same, and in 'kubectl get minions'20:31
dane_leblancI also see the pending state in 'magnum pod-show ...'20:31
dane_leblancAre you still not seeing any pods in 'sudo docker ps' on the minions?20:33
mfalaticSo basically I can get to the same things you can, but my docker isn't showing anything yet.20:33
*** adrian_otto has joined #openstack-containers20:34
mfalaticnot seeing anything on the master or the slaves.20:35
mfalaticnot quite sure how to monitor the pod from the host.20:35
dane_leblanc'magnum service-list' shows redis-sentinel?20:35
mfalaticthe progress of the pod20:35
mfalaticfrom the host, yes, I see redis-sentinel20:36
mfalatic(again, first time trying this part of the quickstart)20:38
mfalaticStatus of redis-master is still "pending"20:39
*** coolsvap is now known as coolsvap|away20:39
mfalaticThis is rather frustrating.20:41
dane_leblancDoes 'sudo journalctl -u kubelet' on the minions show anything interesting?20:42
dane_leblancI see a bunch of activity that looks good, but then a bunch of 'Cannot get host IP: host IP unknown' errors20:43
*** achanda has joined #openstack-containers20:43
mfalaticalso seeing Error syncing pod20:44
mfalaticnot sure why I don't see anything in docker ps though20:44
*** sdake_ is now known as sdake20:46
mfalaticsorry I'm not able to be of more help yet20:48
dane_leblancmfalatic: Thanks for trying!20:49
dane_leblancI wonder if anyone can get this to work?20:49
mfalaticWell, it'd be nice to know why I can't get as far as you could.20:50
mfalaticMaybe I should re-stack20:50
mfalaticreboot and restack even20:50
dane_leblancI don't know what logs to collect for debugging this20:50
mfalaticclearly it must have worked for someone at some point.20:51
jjfrericI was wrestling with that a few days ago. I haven't had a chance to retry since, but I'm glad it's not just me :)21:01
-openstackstatus- NOTICE: Gerrit will be unavailable from 22:00 to 22:30 UTC for project renames21:02
dane_leblancMaybe we can ping the group for help on this for the next magnum IRC meeting, since I don't know how many folks are tuned in on Friday afternoon21:02
*** jjfreric has left #openstack-containers21:04
*** suro-patz has joined #openstack-containers21:05
mfalaticI wonder if people are seeing what I'm seeing, which is less than what you saw even.21:08
mfalaticaw dang... gerrit choked on one of the functional gates.21:09
*** ig0r_ has quit IRC21:10
mfalaticI suspect something changed elsewhere. :-/
*** saksham has joined #openstack-containers21:11
*** achanda has quit IRC21:14
*** adrian_otto has quit IRC21:16
*** adrian_otto has joined #openstack-containers21:17
*** sdake_ has joined #openstack-containers21:24
*** achanda has joined #openstack-containers21:26
*** sdake has quit IRC21:27
*** Marga__ has quit IRC21:31
*** Marga_ has joined #openstack-containers21:31
mfalatichongbin - I see you ran recheck - was about to do that. It's a very weird error and I've been on #openstack-qa trying to see if it's me or what.21:40
mfalaticThe first error (which may be causing all the rest to fail) is: "nova keypair-add default" --> ERROR (DiscoveryFailure): Could not determine a suitable URL for the plugin21:40
mfalatichas anyone seen that before?21:40
*** achanda has quit IRC21:50
*** Marga_ has quit IRC21:51
*** achanda has joined #openstack-containers21:54
-openstackstatus- NOTICE: Gerrit is unavailable from approximately 22:00 to 22:30 UTC for project renames21:57
*** ChanServ changes topic to "Gerrit is unavailable from approximately 22:00 to 22:30 UTC for project renames"21:57
*** Drago1 has quit IRC22:00
*** Drago has joined #openstack-containers22:02
*** kebray has quit IRC22:04
*** dboik_ has quit IRC22:04
*** Drago has quit IRC22:04
*** Drago has joined #openstack-containers22:06
*** harlowja_ has joined #openstack-containers22:09
*** harlowja has quit IRC22:10
*** sdake has joined #openstack-containers22:11
*** sdake_ has quit IRC22:14
mfalaticStill getting the same error. Have others run into this nova keypair-add error?22:19
dane_leblancmfalatic: Trying to figure this out. I was looking at the functional test errors in the log you pointed to:
dane_leblancThese errors point to a null image_id22:24
dane_leblancIn, image_id, as well as magnum_url should be set either in environment, or in a functional_creds.conf file.22:25
dane_leblancBut I don't see where the functional tests set either environment or create a functional_creds.conf file.22:26
hongbinmfalatic: I haven't seen that error before.22:26
mfalaticThe first error may be with glance...22:27
*** jhova has quit IRC22:27
mfalatic"An auth plugin is required to fetch a token"22:28
mfalaticthen the nova error, then the rest. Whatever is going on, it changed in the last 24 hours and I think it's something apart from Magnum22:28
hongbinCould be. I couldn't imagine a docs change will break the test22:29
*** ChanServ changes topic to "OpenStack Containers Team"22:30
*** jjlehr has quit IRC22:41
mfalaticI would think it'd affect everyone though...22:45
mfalaticIf anyone else has an open review and wants to try a recheck on it....22:45
*** ameybhide has quit IRC22:53
*** hongbin has quit IRC22:58
*** jruano has joined #openstack-containers23:04
dane_leblancMust be an upstream gating test breakage. The glance error seems as if the openstack credentials are set before the 'glance image-show ...', but the openstack credentials should be set in the (magnum/tests/contrib/post_test_hook.sh23:12
dane_leblanc), line "source openrc admin admin"23:12
*** dims has quit IRC23:13
dane_leblancMeant to say "credentials AREN'T set"23:13
*** dims has joined #openstack-containers23:17
*** eghobo_ has joined #openstack-containers23:31
*** eghobo has quit IRC23:34
*** openstackgerrit has quit IRC23:39
*** openstackgerrit has joined #openstack-containers23:39
*** jruano has quit IRC23:42
mfalaticdane_leblanc Hmm... I concur, but what's the proper course of action? Filing a bug on... qa?23:50
*** Daviey has quit IRC23:52

Generated by 2.14.0 by Marius Gedminas - find it at!