cloudnull | any cores around mind firing https://review.openstack.org/#/c/568872/ though ? | 00:10 |
---|---|---|
*** jra has joined #openstack-ansible | 00:14 | |
jra | What's the state of multiple cell support in OSA for pike and/or queens? | 00:15 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Convert rsyslog to an include_task with group_vars https://review.openstack.org/556002 | 00:18 |
cloudnull | jra its all cells v2 now | 00:19 |
cloudnull | and that's implemented | 00:19 |
cloudnull | though i dont know anyone using more than cell 0 at this point | 00:19 |
jra | cloudnull: so if I want to create multiple cells and attach compute nodes to them, how do I use OSA to accomplish that? | 00:19 |
jra | We do have a use case for more than cell0, although I'm also interested in alternatives to that... | 00:20 |
cloudnull | you'd probably want to extend the environment for nova to add a specific cells grouping that way you can have group vars per cell. then set https://github.com/openstack/openstack-ansible-os_nova/blob/master/defaults/main.yml#L97-L99 per cell. | 00:21 |
cloudnull | that said, i've not done this so i really have no idea. | 00:21 |
cloudnull | I'm sure there will be some learnings that will need to be applied | 00:22 |
cloudnull | how big is the environment, or is this just a logical seperation of nodes ? | 00:22 |
openstackgerrit | Andy Smith proposed openstack/openstack-ansible-os_aodh master: Update to use oslo.messaging services for RPC and Notify https://review.openstack.org/568989 | 00:23 |
jra | ~10k cores, with two distinct service domains: one requires high availability but can tolerate overcommitting resources, the other requires high performance but can tolerate some volatility | 00:23 |
jra | to me, that sounded like a good fit for cells | 00:23 |
jra | we're currently getting by with aggregates and affinities, but it means only one scheduling strategy shared for both | 00:23 |
jra | which is not ideal | 00:23 |
cloudnull | Could you use different availability zones ? | 00:26 |
cloudnull | Set the over commit ratio and scheduler hints using group/host vars? | 00:27 |
*** cjloader has joined #openstack-ansible | 00:27 | |
jra | Probably; the main reason we wanted cells was the prospect of being able to specify different scheduler filters/weighers per cell | 00:27 |
cloudnull | Doesn't resolve all of the scheduler limitations but probably gets you super close without introdcing cells. | 00:28 |
cloudnull | Ah ok. | 00:28 |
jra | ie, in one cell I want to schedule for least-loaded, but in the other I want to schedule to pack memory | 00:28 |
cloudnull | While the code exists I'm 100% positive it could be improved. | 00:28 |
cloudnull | I'd be game to help work out the issues. | 00:29 |
cloudnull | But there is a bit of unknown there. | 00:29 |
cloudnull | Will you be at summit next week? | 00:29 |
jra | I will be, yes | 00:30 |
cloudnull | Cool! Maybe we can get a hack session going and figure this out ? | 00:30 |
cloudnull | evrardjp will be there too . | 00:30 |
*** vnogin has joined #openstack-ansible | 00:30 | |
cloudnull | I'm not sure whom else. I also know spotz is going to be there. | 00:31 |
jra | That'd be great, I'm definitely interested to know if this is something we can do | 00:31 |
*** vnogin has quit IRC | 00:31 | |
cloudnull | I've got a smallish lab (25 hosts) I might be able to give it a go there. | 00:31 |
jra | Supposedly we're trying to go prod with this in about 6 weeks, so we'll see whether it's going to be realistic... | 00:31 |
*** cjloader has quit IRC | 00:31 | |
*** dxiri has joined #openstack-ansible | 00:32 | |
cloudnull | You should be able to go prod and break things up with cells later as needed. | 00:32 |
jra | Yeah, we had dreams of "getting it right" at deploy time | 00:33 |
jra | :) | 00:33 |
cloudnull | With a six week time table, I'd probably stick with multiple AZs and host aggregates. | 00:33 |
cloudnull | Then introduce multiple cells later. | 00:33 |
jra | You're probably right, I was kinda hoping I'd discover this was all well-trodden ground | 00:33 |
cloudnull | Hahaha | 00:34 |
cloudnull | I wish. Cells v2 is still kinda fresh. | 00:34 |
*** vnogin has joined #openstack-ansible | 00:34 | |
jra | Yeah. I come from an enterprise IT background, and the pace of change in cloud land is taking some getting used to | 00:34 |
cloudnull | At Rackspace , on our public cloud we use cells v1 and that has some good docs/known limitations. Idk if the same is true with cells v2 | 00:34 |
jra | yeah, cells v1 looked a bit painful to implement | 00:35 |
cloudnull | Yes. It was. | 00:35 |
jra | we were all excited about v2, which looked like the "fixes everything" release | 00:35 |
*** vnogin has quit IRC | 00:39 | |
*** chigang__ has joined #openstack-ansible | 00:39 | |
*** dxiri has quit IRC | 00:42 | |
cloudnull | https://twitter.com/cloudnull/status/996913778183811072?s=19 | 00:42 |
cloudnull | Sent up the bat signal ;) | 00:42 |
*** dxiri has joined #openstack-ansible | 00:42 | |
cloudnull | Likely get some good attention going into summit, and when we're there we'll have a lot of good resources on hand. | 00:43 |
cloudnull | :D | 00:43 |
jra | Nice! Thanks | 00:44 |
jra | Awesome. Will onboarding be the best place to catch up with you all? | 00:46 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Use fan out when deploying with nspawn https://review.openstack.org/568331 | 00:46 |
*** dxiri has quit IRC | 00:47 | |
cloudnull | That'll be one place. we have a google hang out if you want to join i can invite you . | 00:47 |
cloudnull | DM me an email address and I'll add you | 00:48 |
jra | done | 00:52 |
cloudnull | awesome | 00:55 |
* cloudnull gotta run be back later. | 00:55 | |
*** osnaya_ has quit IRC | 00:55 | |
*** jra has quit IRC | 00:58 | |
*** vnogin has joined #openstack-ansible | 01:12 | |
*** cmart has quit IRC | 01:14 | |
*** vnogin has quit IRC | 01:17 | |
*** klamath has quit IRC | 01:26 | |
*** klamath has joined #openstack-ansible | 01:26 | |
*** cmart has joined #openstack-ansible | 01:30 | |
idlemind | cloudnull thx for checking on stable/pike for me ... i ended up having to run over to minnesota last minute :) | 01:33 |
*** markvoelker_ has quit IRC | 01:34 | |
cloudnull | Np. All good? Did you ever get the repo build going? | 01:35 |
idlemind | getting back into later on or tomorrow morning | 01:35 |
idlemind | been a whirlwind 48 hours | 01:35 |
idlemind | no rest for the wicked i suppose | 01:36 |
cloudnull | well i hope everything is ok. | 01:37 |
*** markvoelker has joined #openstack-ansible | 01:38 | |
*** vnogin has joined #openstack-ansible | 01:40 | |
idlemind | oh ya; just normal stuff | 01:42 |
idlemind | ya know "network is on fire" | 01:42 |
idlemind | several commands later ... all good | 01:42 |
*** vnogin has quit IRC | 01:45 | |
openstackgerrit | Merged openstack/openstack-ansible-lxc_hosts master: Trivial: Fix the pep8 warning https://review.openstack.org/566836 | 01:51 |
*** vnogin has joined #openstack-ansible | 01:58 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Use fan out when deploying with nspawn https://review.openstack.org/568331 | 02:00 |
*** vnogin has quit IRC | 02:02 | |
*** vnogin has joined #openstack-ansible | 02:05 | |
cloudnull | hahahah. C'est la vie | 02:10 |
*** vnogin has quit IRC | 02:10 | |
*** vnogin has joined #openstack-ansible | 02:13 | |
*** vnogin has quit IRC | 02:17 | |
*** cmart has quit IRC | 02:17 | |
*** klamath_ has joined #openstack-ansible | 02:18 | |
*** cjloader has joined #openstack-ansible | 02:19 | |
*** klamath has quit IRC | 02:22 | |
*** klamath_ has quit IRC | 02:23 | |
*** cjloader has quit IRC | 02:24 | |
*** vnogin has joined #openstack-ansible | 02:29 | |
*** cmart has joined #openstack-ansible | 02:31 | |
*** vnogin has quit IRC | 02:33 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Use fan out when deploying with nspawn https://review.openstack.org/568331 | 02:38 |
*** lbragstad has quit IRC | 02:39 | |
*** masber has quit IRC | 02:49 | |
*** dave-mcc_ has quit IRC | 02:52 | |
*** masber has joined #openstack-ansible | 02:52 | |
*** vnogin has joined #openstack-ansible | 02:55 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Use fan out when deploying with nspawn https://review.openstack.org/568331 | 02:58 |
*** vnogin has quit IRC | 03:00 | |
*** cmart has quit IRC | 03:06 | |
*** qwebirc48444 has quit IRC | 03:11 | |
*** vnogin has joined #openstack-ansible | 03:11 | |
*** KeithMnemonic has joined #openstack-ansible | 03:15 | |
*** vnogin has quit IRC | 03:16 | |
*** cjloader has joined #openstack-ansible | 03:19 | |
*** klamath has joined #openstack-ansible | 03:20 | |
*** nicolasbock has quit IRC | 03:23 | |
*** cjloader has quit IRC | 03:24 | |
*** tux_ has joined #openstack-ansible | 03:26 | |
tux_ | Hello folks | 03:26 |
tux_ | If i don't want storage node then should i ## comment that in "openstack_user_config.yml" file right? | 03:27 |
*** vnogin has joined #openstack-ansible | 03:28 | |
cloudnull | yup | 03:30 |
cloudnull | then nothing cinder will be deployed | 03:30 |
tux_ | Thanks! @cloudnull | 03:31 |
tux_ | Glad to see you :) | 03:32 |
*** klamath has quit IRC | 03:33 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-lxc_hosts master: Add task to regather facts after run https://review.openstack.org/568632 | 03:33 |
cloudnull | 2u2 :) | 03:33 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Remove the package cache update https://review.openstack.org/568615 | 03:34 |
*** jbadiapa has quit IRC | 03:35 | |
*** jbadiapa has joined #openstack-ansible | 03:35 | |
*** vnogin has quit IRC | 03:36 | |
*** KeithMnemonic has quit IRC | 03:41 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Begin testing opensuse and centos with nspawn https://review.openstack.org/568197 | 03:41 |
*** cjloader has joined #openstack-ansible | 03:46 | |
tux_ | is Python 2.7.5 sufficient to install pike 16.0.10 tags? or should i use python 3.4.x? | 03:46 |
*** udesale has joined #openstack-ansible | 03:49 | |
*** vnogin has joined #openstack-ansible | 03:54 | |
*** masuberu has joined #openstack-ansible | 03:56 | |
*** vnogin has quit IRC | 03:59 | |
*** masber has quit IRC | 04:00 | |
*** masber has joined #openstack-ansible | 04:02 | |
*** masuberu has quit IRC | 04:05 | |
*** vnogin has joined #openstack-ansible | 04:08 | |
*** chigang__ has quit IRC | 04:08 | |
*** udesale has quit IRC | 04:10 | |
*** vnogin has quit IRC | 04:12 | |
*** udesale has joined #openstack-ansible | 04:13 | |
*** vnogin has joined #openstack-ansible | 04:14 | |
*** tux_ has quit IRC | 04:14 | |
cloudnull | 2.7 is totally fine | 04:15 |
*** vnogin has quit IRC | 04:19 | |
*** gyee has quit IRC | 04:26 | |
*** udesale has quit IRC | 04:36 | |
*** udesale has joined #openstack-ansible | 04:37 | |
*** vnogin has joined #openstack-ansible | 04:46 | |
*** cjloader has quit IRC | 04:47 | |
*** cjloader has joined #openstack-ansible | 04:48 | |
*** vnogin has quit IRC | 04:50 | |
*** cjloader has quit IRC | 04:53 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Use fan out when deploying with nspawn https://review.openstack.org/568331 | 04:54 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Begin testing opensuse and centos with nspawn https://review.openstack.org/568197 | 04:55 |
*** tux_ has joined #openstack-ansible | 05:04 | |
*** vnogin has joined #openstack-ansible | 05:07 | |
*** radeks has joined #openstack-ansible | 05:07 | |
*** vnogin has quit IRC | 05:11 | |
*** vnogin has joined #openstack-ansible | 05:16 | |
*** vnogin_ has joined #openstack-ansible | 05:18 | |
*** vnogin has quit IRC | 05:18 | |
*** udesale has quit IRC | 05:20 | |
*** udesale has joined #openstack-ansible | 05:21 | |
*** vnogin_ has quit IRC | 05:22 | |
*** chhagarw has joined #openstack-ansible | 05:24 | |
*** udesale has quit IRC | 05:26 | |
*** klamath has joined #openstack-ansible | 05:30 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/queens: Update lxc_hosts role SHA for cache resolver fix https://review.openstack.org/569013 | 05:32 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/pike: Update lxc_hosts role SHA for cache resolver fix https://review.openstack.org/569014 | 05:33 |
*** klamath has quit IRC | 05:34 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Update lxc_hosts role SHA for cache resolver fix https://review.openstack.org/569015 | 05:36 |
*** tux_ has quit IRC | 05:37 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Use fan out when deploying with nspawn https://review.openstack.org/568331 | 05:39 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Begin testing opensuse and centos with nspawn https://review.openstack.org/568197 | 05:40 |
*** udesale has joined #openstack-ansible | 05:43 | |
*** vnogin has joined #openstack-ansible | 05:46 | |
*** mma has joined #openstack-ansible | 05:48 | |
*** vnogin has quit IRC | 05:51 | |
*** admin0 has quit IRC | 05:54 | |
*** admin0 has joined #openstack-ansible | 05:55 | |
*** bhujay has joined #openstack-ansible | 05:55 | |
*** Guest90622 is now known as gokhan | 05:59 | |
*** armaan has joined #openstack-ansible | 06:12 | |
*** dxiri has joined #openstack-ansible | 06:17 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_hosts stable/queens: Add mount options for better machinectl performance https://review.openstack.org/569023 | 06:18 |
*** vnogin has joined #openstack-ansible | 06:19 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_hosts stable/queens: Add mount options for better machinectl performance https://review.openstack.org/569023 | 06:21 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-lxc_hosts stable/queens: Enable quota system and set qgroups https://review.openstack.org/568867 | 06:22 |
*** dxiri has quit IRC | 06:22 | |
*** vnogin has quit IRC | 06:24 | |
*** klamath has joined #openstack-ansible | 06:28 | |
*** vnogin has joined #openstack-ansible | 06:30 | |
*** klamath has quit IRC | 06:33 | |
*** vnogin has quit IRC | 06:35 | |
*** niraj_singh has joined #openstack-ansible | 06:39 | |
*** DanyC_ has quit IRC | 06:40 | |
*** DanyC has joined #openstack-ansible | 06:41 | |
*** vnogin has joined #openstack-ansible | 06:43 | |
*** DanyC has quit IRC | 06:45 | |
*** vnogin has quit IRC | 06:47 | |
*** dxiri has joined #openstack-ansible | 06:48 | |
*** mbuil has joined #openstack-ansible | 06:52 | |
*** dxiri has quit IRC | 06:53 | |
*** lvdombrkr has joined #openstack-ansible | 06:57 | |
*** vnogin has joined #openstack-ansible | 07:00 | |
*** shardy has joined #openstack-ansible | 07:02 | |
*** vnogin has quit IRC | 07:05 | |
*** vnogin has joined #openstack-ansible | 07:11 | |
*** vnogin has quit IRC | 07:16 | |
*** epalper has joined #openstack-ansible | 07:17 | |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: infra-journal-remote: Skip playbook if no log_hosts group is defined https://review.openstack.org/569039 | 07:17 |
*** cjloader has joined #openstack-ansible | 07:18 | |
*** cjloader has quit IRC | 07:23 | |
*** dxiri has joined #openstack-ansible | 07:32 | |
odyssey4me | hwoarang evrardjp related to that patch, perhaps we can finalise the review for https://review.openstack.org/#/c/556002/17 so that we don't run both systems in parallel | 07:36 |
*** dxiri has quit IRC | 07:37 | |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: infra-journal-remote: Skip playbook if no log_hosts group is defined https://review.openstack.org/569039 | 07:39 |
hwoarang | oh that's a big patch | 07:39 |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: infra-journal-remote: Skip playbook on undefined or emtpy log_hosts group https://review.openstack.org/569039 | 07:40 |
odyssey4me | hwoarang not so big - it's mostly moving some things around | 07:41 |
*** vnogin has joined #openstack-ansible | 07:43 | |
evrardjp | it's not a big patch but... We don't have any functional testing of it | 07:43 |
evrardjp | we don't know if the files are really present (i.e. the rotation happens) | 07:44 |
odyssey4me | evrardjp that's no different to before | 07:45 |
odyssey4me | the test for that would belong in the role anyway, not here | 07:46 |
evrardjp | odyssey4me: not really, because we know the role works | 07:46 |
evrardjp | because of the test is done there | 07:46 |
evrardjp | here my concern is integration | 07:46 |
evrardjp | what if private: yes causes problems with variable passing | 07:46 |
evrardjp | what if a regression in ansible happens and breaks this behavior? We'll not see it immediately | 07:47 |
evrardjp | I will file a bug | 07:47 |
odyssey4me | well, in my own testing I can confirm that it doesn't - I've been using it in other roles - private: yes just prevents facts set in the role coming back to the play | 07:47 |
evrardjp | it is of low importance | 07:47 |
*** vnogin has quit IRC | 07:47 | |
*** fghaas has joined #openstack-ansible | 07:48 | |
*** pvradu has joined #openstack-ansible | 07:48 | |
odyssey4me | so - in other news, we have a gate wedge | 07:48 |
odyssey4me | two problems: one is that centos is broken for all os_* roles (at least) due to the problem we already know about (somepackage conflict) | 07:49 |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible master: infra-journal-remote: Skip playbook on empty emtpy log_hosts group https://review.openstack.org/569039 | 07:49 |
odyssey4me | the other is that with the merge of https://review.openstack.org/559127 all os_* roles need inventory changes for those tasks to work | 07:49 |
openstackgerrit | Jean-Philippe Evrard proposed openstack/openstack-ansible master: Convert rsyslog to an include_task https://review.openstack.org/556002 | 07:50 |
*** pvradu has quit IRC | 07:50 | |
evrardjp | odyssey4me: could you revote ^ | 07:50 |
evrardjp | I removed with group vars, as this group vars handling was removed from the initial patch during patch dev | 07:50 |
*** Taseer has quit IRC | 07:50 | |
evrardjp | (from the commit message) | 07:51 |
*** DanyC has joined #openstack-ansible | 07:51 | |
odyssey4me | evrardjp ah, good call | 07:51 |
odyssey4me | great - now we can move on :) | 07:51 |
odyssey4me | right, so I'm thinking we temporarily make centos tests non-voting - clearing the way for all roles to merge anything working | 07:52 |
odyssey4me | also, I can modify the tests repo to do the tasks using the right group depending on what's defined | 07:52 |
odyssey4me | everyone happy with that approach? | 07:53 |
evrardjp | not sure what you mean | 07:53 |
evrardjp | just a sec | 07:53 |
*** Taseer has joined #openstack-ansible | 07:54 | |
odyssey4me | let me take an example - https://review.openstack.org/#/c/558620/ is a perfectly good patch, but failing on centos | 07:54 |
odyssey4me | oh, interesting - this is a different failure: http://logs.openstack.org/20/558620/15/check/openstack-ansible-functional-centos-7/2776eb9/logs/ara-report/result/d21c9764-b859-4ab2-b02a-8f712acaee7d/ | 07:55 |
*** tosky has joined #openstack-ansible | 07:55 | |
odyssey4me | FYI some of these are good to go, I think: https://review.openstack.org/#/q/topic:bp/hybrid-messaging | 07:55 |
evrardjp | that's the usual copr failure | 07:57 |
*** lvdombrkr89 has joined #openstack-ansible | 07:58 | |
odyssey4me | usual? is this consistent? | 07:58 |
evrardjp | not consistent, else it would be fixed. | 07:58 |
evrardjp | just regularily showing up | 07:58 |
*** radeks_ has joined #openstack-ansible | 07:58 | |
evrardjp | odyssey4me: mmm | 07:59 |
evrardjp | I have a problem with those patches | 07:59 |
evrardjp | it's smart, but you have to use the same server for all your infra ? | 08:00 |
evrardjp | (as oslomsg_rpc_servers isn't scoped?) | 08:00 |
*** lvdombrkr has quit IRC | 08:00 | |
*** Taseer has quit IRC | 08:01 | |
odyssey4me | well, I suppose previously it was *possible* to split every service onto their own MQ system... in practise, though, I wonder if that's ever been used | 08:02 |
odyssey4me | RPC vs Notifications, sure | 08:02 |
evrardjp | rackspace public cloud was doing it. They said it was a mandatory thing if they wanted to use OSA. | 08:03 |
evrardjp | it's still achievable though | 08:03 |
evrardjp | it's just harder. | 08:03 |
odyssey4me | oh really? | 08:03 |
evrardjp | yeah, but in the meantime people working on this PoC left. :p | 08:03 |
evrardjp | I don't know where they are about that now. | 08:04 |
odyssey4me | well, so I plan to fast follow these patches with https://review.openstack.org/#/q/topic:move-mq-setup-into-roles and can happily put the namespacing back if you like | 08:04 |
odyssey4me | or I suppose we could so it now in the role | 08:04 |
evrardjp | odyssey4me: so you know, the idea was to have different component managed easily: so switching all nova things, or switching all glance things. | 08:05 |
evrardjp | without touching the rest | 08:05 |
evrardjp | I am fine with namespacing back, but it doesn't hurt to do it right at the first time? | 08:05 |
odyssey4me | well, we could do it simply like this in the role without requiring more group_vars wiring: glance_oslomsg_rpc_servers: "{{ oslomsg_rpc_servers | default('127.0.0.1') }}" | 08:06 |
evrardjp | yes | 08:06 |
evrardjp | convention over configuration. | 08:06 |
odyssey4me | I prefer that approach - less group vars | 08:06 |
evrardjp | I am fine with that, we have set that in the past. | 08:06 |
evrardjp | we should just list all the global variables we are using by default somewhere. | 08:07 |
odyssey4me | given the memory usage thanks to our inventory & group_vars, and also the duplication of vars all over the place - I'm preferring that approach now and going forward | 08:07 |
evrardjp | I agree with you | 08:07 |
evrardjp | we are setting convention, and if the convention is documented, we are good. | 08:07 |
odyssey4me | rather than list, ideally we should figure out how we can auto_generate things like that | 08:07 |
evrardjp | one step at a time :) | 08:08 |
odyssey4me | for example, the secrets used on a role could be auto-documented by trawling the role for key words in the docs publishing process | 08:08 |
evrardjp | Yeah, or a well defined comment | 08:09 |
odyssey4me | we could also use some sort of var name convention to make figuring out other required vars and publishing them | 08:09 |
evrardjp | yes, or what I proposed with the variable loading into a namespace | 08:09 |
odyssey4me | or yes, we could also use some other convention - like a comment or whatever | 08:09 |
evrardjp | many solutions | 08:09 |
*** Taseer has joined #openstack-ansible | 08:09 | |
evrardjp | but I think this is not so important | 08:09 |
odyssey4me | whatever we do, we should make it easier to lint to reduce the pressure on reviewers | 08:10 |
evrardjp | important is to keep feature parity, and clean the vars. | 08:10 |
*** pvradu has joined #openstack-ansible | 08:16 | |
*** blinkiz has joined #openstack-ansible | 08:18 | |
*** cjloader has joined #openstack-ansible | 08:19 | |
evrardjp | odyssey4me: We should have done the add bionic testing as -nv, because now we'll have things -nv things voting, and we won't immediately now what is what | 08:22 |
blinkiz | In all openstack_user_config.yml examples internal IPv4 addresses like 172.29.2xx.xx is used. Am trying to wrap my head around which services needs a public IPv4 address. Horizon needs one for example. Is that external_lb_vip_address? But here it also says external_lb_vip_address is for all Openstack API calls. Does openstack-ansible set up a router somewhere that am missing? Is it any routing between all the internal networks? | 08:22 |
evrardjp | because http://logs.openstack.org/32/567032/4/check/openstack-ansible-dir-ubuntu-bionic/65fb7d1/job-output.txt.gz#_2018-05-15_04_14_01_755935 will require more work | 08:22 |
evrardjp | blinkiz: we have haproxy as reverse proxy | 08:23 |
*** cjloader has quit IRC | 08:23 | |
evrardjp | blinkiz: it listens on external_lb_vip_address and internal_lb_vip_addresses. It has access to internal mgmt network | 08:23 |
evrardjp | set a fqdn for external_lb_vip_address, and everything will look great | 08:24 |
blinkiz | Thanks for the information | 08:25 |
niraj_singh | evrardjp: please look into this http://paste.openstack.org/show/721147/ | 08:25 |
blinkiz | How does management network communicate with the outside world? Does openstack-ansible install a router in this network? Am I supposed to create a router in management network so containers can communicate with the internet? | 08:26 |
*** vnogin has joined #openstack-ansible | 08:27 | |
jrosser | blinkiz: there are a set of instructions for setting up your hosts here https://docs.openstack.org/project-deploy-guide/openstack-ansible/queens/targethosts.html. You need one interface that somehow allows internet access for install packages and the like, you can choose to do that via the management network if you wish | 08:28 |
odyssey4me | evrardjp I don't understand - where was a non-voting test added without -nv ? | 08:28 |
odyssey4me | as far as I'm concerned, if the test was passing, make it voting - if not, then it should be -nv and not voting... not without -nv and non-voting | 08:29 |
evrardjp[m] | I completely agree, but ppl disagreed with me in the irc. | 08:30 |
evrardjp[m] | I completely agree, but ppl disagreed with me in the irc. | 08:30 |
blinkiz | Am a Internet Service Provider so I have almost unlimited IPv4 addresses to use. Is it a bad idea to give public IPv4 addresses to all services on management network? Does the containers take care of there own security or do they need NAT before accessing internet? | 08:31 |
evrardjp[m] | link about what you are talking about odyssey4me ? | 08:31 |
*** vnogin has quit IRC | 08:31 | |
odyssey4me | evrardjp[m] I'm hoping you'll point me at an example of a non-voting test that doesn't have -nv ? | 08:31 |
odyssey4me | given that this was your complaint | 08:32 |
jrosser | blinkiz: imho it is a very bad idea to just hang the whole infra on the internet, nothing stops you though | 08:32 |
evrardjp | I am very confused | 08:32 |
evrardjp | what are you talking about? | 08:32 |
evrardjp | I remember things like that in the past | 08:32 |
odyssey4me | 'evrardjp:odyssey4me: We should have done the add bionic testing as -nv, because now we'll have things -nv things voting, and we won't immediately now what is what' | 08:32 |
evrardjp | we had to "temporarily" disable centos from voting, so instead of changing all the roles, we simply marked in tests the centos test as voting: false without changing the jobs in the roles | 08:33 |
evrardjp | oh | 08:33 |
evrardjp | different thing :) | 08:33 |
evrardjp | sorry english :p | 08:33 |
evrardjp | yeah | 08:33 |
evrardjp | it's not that big of a deal | 08:33 |
evrardjp | don't worry | 08:33 |
*** radeks has quit IRC | 08:34 | |
evrardjp | I just think some of those bionic patches should get included as -nv, and meta added later. | 08:34 |
odyssey4me | evrardjp yes, that sort of thing is a brief workaround - not a permanent thing | 08:34 |
odyssey4me | evrardjp sure, if the patch isn't ready, but you want the testing put in - then go for it just like that | 08:34 |
odyssey4me | we don't have to make it work right away | 08:34 |
*** gkadam has joined #openstack-ansible | 08:35 | |
odyssey4me | there are some who may prefer to leave the patch up until it's made to work... either way works for me | 08:36 |
odyssey4me | I guess if there's a list of unmerged patches up, it makes it easy to identify which roles/repositories still need work | 08:36 |
*** bhujay has quit IRC | 08:36 | |
blinkiz | If I have understood it right, internet access for the container/services is only needed for updating packages, time and a few simple things. Nothing about the functionality of openstack or communication within openstack is done over this internet/routed connection. Is this correct? | 08:38 |
blinkiz | So I need myself set up a simple router so the container has a default gateway in the management network? | 08:39 |
blinkiz | Or have I understood it wrong? Maybe containers communicate with the infrastructure hosts that is providing default gateway for containers? | 08:40 |
blinkiz | Well well, guess I just need to try and see how it turns out :-) | 08:41 |
*** flaviosr has quit IRC | 08:41 | |
jrosser | its a little more subtle than that | 08:41 |
*** flaviosr has joined #openstack-ansible | 08:42 | |
*** vnovakov has joined #openstack-ansible | 08:42 | |
jrosser | blinkiz: https://docs.openstack.org/project-deploy-guide/openstack-ansible/draft/app-networking.html | 08:42 |
jrosser | blinkiz: just make sure the hosts themselves have a route to the internet, don't worry about the containers themselves | 08:44 |
blinkiz | Oh, I see. yeah, that makes it simple. I can assign external IPv4 addresses on br-mgmt for all real hosts. Keep cidr_networks: containers: 172.29.236.0/22.. Okay.. | 08:46 |
odyssey4me | blinkiz exposing your mgmt address to the world is a terrible idea, but you can do it if you want | 08:47 |
odyssey4me | just make sure you firewall it | 08:47 |
odyssey4me | better to create a public bridge which is firewalled, and add a route to the internet out on that bridge | 08:48 |
blinkiz | Okay, got it. I clearly have misunderstood some things about the br-mgmt network. | 08:49 |
jrosser | blinkiz: ^ take the hint :) don't put public addresses where you don't need them | 08:49 |
jrosser | the simplest thing you can do is to arrange a firewall/NAT router on the mgmt network as the examples show and leave the whole addressing scheme just as-is | 08:50 |
jrosser | then everything will be familiar to folks here if you want some extra help | 08:50 |
openstackgerrit | Periyasamy Palanisamy proposed openstack/openstack-ansible-os_neutron master: Public n/w access for vxlan tenant n/w in ml2.ovs https://review.openstack.org/567809 | 08:52 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: Optionally use the old rabbitmq setup tasks https://review.openstack.org/569061 | 08:56 |
*** vnogin has joined #openstack-ansible | 08:59 | |
*** pvradu has quit IRC | 09:01 | |
*** bhujay has joined #openstack-ansible | 09:01 | |
*** dxiri has joined #openstack-ansible | 09:03 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: Optionally use the old rabbitmq setup tasks https://review.openstack.org/569061 | 09:03 |
odyssey4me | evrardjp ^ that should unblock role tests for the os_roles again | 09:04 |
*** vnogin has quit IRC | 09:04 | |
odyssey4me | I wonder if I should switch those up to use the 'vars:' key word at the same time | 09:04 |
*** mma has quit IRC | 09:04 | |
*** dxiri has quit IRC | 09:07 | |
*** pvradu_ has joined #openstack-ansible | 09:08 | |
*** pvradu has joined #openstack-ansible | 09:09 | |
*** pvradu has quit IRC | 09:09 | |
*** pvradu_ has quit IRC | 09:09 | |
*** pvradu has joined #openstack-ansible | 09:09 | |
*** pvradu has quit IRC | 09:14 | |
evrardjp | odyssey4me: vars? | 09:15 |
evrardjp | oh yes | 09:15 |
evrardjp | it's cleaner I think | 09:17 |
evrardjp | but it works, so meh | 09:17 |
evrardjp | you'll change that anyway | 09:17 |
evrardjp | oh it doesn't pass tests | 09:17 |
evrardjp | oh yeah forgot to wrap | 09:18 |
evrardjp | commented | 09:18 |
*** storshoo has quit IRC | 09:21 | |
odyssey4me | when to quote, and when not to quote - it's so annoyingly inconsistent | 09:21 |
*** vnogin has joined #openstack-ansible | 09:23 | |
*** pvradu has joined #openstack-ansible | 09:24 | |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: Optionally use the old rabbitmq setup tasks https://review.openstack.org/569061 | 09:25 |
*** vnogin has quit IRC | 09:28 | |
*** storshoo has joined #openstack-ansible | 09:29 | |
*** pvradu has quit IRC | 09:32 | |
*** mma has joined #openstack-ansible | 09:38 | |
*** pvradu has joined #openstack-ansible | 09:40 | |
*** mma has quit IRC | 09:42 | |
*** vnogin has joined #openstack-ansible | 09:42 | |
*** dxiri has joined #openstack-ansible | 09:48 | |
*** dxiri has quit IRC | 09:53 | |
fghaas | anyone I could bother for a Verified on https://review.openstack.org/#/c/568360/ ? (thanks odyssey4me for fixing up lxc_hosts!) | 09:57 |
niraj_singh | evrardjp: hi may you please look into this http://paste.openstack.org/show/721150/ | 09:57 |
fghaas | s/Verified/Workflow/ of course (I'm having a "Florian is confused" day today, sorry) | 10:03 |
*** mma has joined #openstack-ansible | 10:15 | |
odyssey4me | cores - we need https://review.openstack.org/#/c/569061/ merged to unblock the os_* roles in master - please review asap | 10:22 |
*** klamath has joined #openstack-ansible | 10:30 | |
*** nicolasbock has joined #openstack-ansible | 10:30 | |
*** dxiri has joined #openstack-ansible | 10:33 | |
*** klamath has quit IRC | 10:35 | |
*** dxiri has quit IRC | 10:37 | |
*** armaan has quit IRC | 10:48 | |
*** armaan has joined #openstack-ansible | 10:49 | |
*** mbuil has quit IRC | 10:54 | |
*** klamath has joined #openstack-ansible | 10:56 | |
*** guilhermesp has quit IRC | 11:02 | |
*** guilhermesp_ is now known as guilhermesp | 11:02 | |
*** ansmith has quit IRC | 11:04 | |
*** dave-mccowan has joined #openstack-ansible | 11:08 | |
*** cjloader has joined #openstack-ansible | 11:19 | |
*** armaan has quit IRC | 11:23 | |
*** armaan has joined #openstack-ansible | 11:23 | |
*** cjloader has quit IRC | 11:23 | |
spotz | odyssey4me: done | 11:26 |
odyssey4me | thanks spotz | 11:29 |
spotz | fghaas: Florian you're reviewed too | 11:29 |
*** electrofelix has joined #openstack-ansible | 11:30 | |
fghaas | spotz: thanks, please subscribe to my HaaS (Hugs as a Service) API | 11:31 |
spotz | jra if you'd like to be added to the hangout to meet up with the team just PM me your email if you haven't given it to cloudnull yet | 11:32 |
spotz | fghaas: You be in Vancouver? | 11:33 |
*** lbragstad has joined #openstack-ansible | 11:33 | |
fghaas | spotz: I will be in a hammock under my cherry tree in Austria with two little kids in my lap, and my next Summit will be Berlin :) | 11:34 |
spotz | fghaas: I'd almost trade you, I really liked Salzburg and Vienna:) | 11:35 |
*** klamath has quit IRC | 11:45 | |
*** dxiri has joined #openstack-ansible | 11:48 | |
*** udesale has quit IRC | 11:53 | |
*** dxiri has quit IRC | 11:53 | |
*** vnovakov has quit IRC | 12:04 | |
*** nsingh has joined #openstack-ansible | 12:15 | |
mgariepy | morning everyone | 12:15 |
openstackgerrit | Merged openstack/openstack-ansible-tests master: Optionally use the old rabbitmq setup tasks https://review.openstack.org/569061 | 12:18 |
*** niraj_singh has quit IRC | 12:18 | |
*** sawblade6 has quit IRC | 12:19 | |
odyssey4me | o/ mgariepy if you have a little time to work through the review queue, that'd be swell :) | 12:24 |
hwoarang | what is it to review? | 12:25 |
*** niraj_singh has joined #openstack-ansible | 12:26 | |
*** nsingh has quit IRC | 12:26 | |
odyssey4me | hwoarang want a list? | 12:26 |
odyssey4me | https://review.openstack.org/#/q/Iba329a23d3388a588cc8d030c97c20664edbbaee | 12:26 |
*** vnogin has quit IRC | 12:27 | |
odyssey4me | https://review.openstack.org/569023 / https://review.openstack.org/568485 / https://review.openstack.org/568626 / https://review.openstack.org/568586 | 12:28 |
odyssey4me | https://review.openstack.org/#/q/I4986ff4d21e9fa8b77115ab80c66de99afe45f6c | 12:28 |
odyssey4me | https://review.openstack.org/568589 / https://review.openstack.org/#/q/status:open+I5bcc8f78ec8a1a175f0bfae3f38101d26e7e3803 | 12:29 |
*** vnogin has joined #openstack-ansible | 12:30 | |
openstackgerrit | Christian Zunker proposed openstack/openstack-ansible-os_cloudkitty master: Configure cloudkitty-api to use uwsgi https://review.openstack.org/567168 | 12:30 |
hwoarang | odyssey4me: ok on it | 12:31 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-tests master: Move database creation into role (glance) https://review.openstack.org/568492 | 12:32 |
openstackgerrit | Merged openstack/openstack-ansible-os_neutron stable/pike: Add default for neutron_ovs_service_name https://review.openstack.org/568360 | 12:32 |
*** tosky has quit IRC | 12:32 | |
odyssey4me | hwoarang https://review.openstack.org/568492 too :) | 12:32 |
odyssey4me | and https://review.openstack.org/568867 | 12:33 |
hwoarang | we prob need to use the star system to indicate reviews that need attention | 12:33 |
evrardjp | odyssey4me: the gift that keeps on giving :) | 12:33 |
evrardjp | yup building a dashboard | 12:34 |
evrardjp | but whose stars? | 12:34 |
evrardjp | I am using it for things I should remember doing | 12:34 |
evrardjp | sometimes it's for review sometimes it's for patching, sometimes something else... | 12:34 |
hwoarang | the dashboard i think puts cores+stas on the top | 12:34 |
odyssey4me | you may find this URL handy to bookmark: https://review.openstack.org/#/q/project:%22%255Eopenstack/(monitorstack%257Copenstack-ansible).*%22+status:open+label:Verified%253E%253D1%252Czuul+NOT+label:Workflow%253C%253D-1+NOT+label:Code-Review%253C%253D-1+NOT+owner:self | 12:34 |
*** vnogin has quit IRC | 12:34 | |
hwoarang | odyssey4me: can you put that on the wiki along with the other dashboard? | 12:35 |
*** guilhermesp has quit IRC | 12:35 | |
odyssey4me | that's everything zuul has verified, and everything that has no -1 review | 12:35 |
*** guilhermesp has joined #openstack-ansible | 12:36 | |
odyssey4me | ... and that is not your own patch | 12:36 |
*** tosky has joined #openstack-ansible | 12:36 | |
*** sawblade6 has joined #openstack-ansible | 12:37 | |
*** armaan has quit IRC | 12:40 | |
evrardjp | I am using the same in gertty | 12:40 |
*** chhavi__ has joined #openstack-ansible | 12:40 | |
*** armaan has joined #openstack-ansible | 12:41 | |
*** nsingh has joined #openstack-ansible | 12:42 | |
*** chhagarw has quit IRC | 12:44 | |
*** niraj_singh has quit IRC | 12:45 | |
*** vnovakov has joined #openstack-ansible | 12:46 | |
odyssey4me | evrardjp good questions, we'll have to wait for cloudnull to comment - but any changes will need to be made to master, then put on top of that chain | 12:46 |
odyssey4me | I believe that cloudnull is likely on holiday today, unfortunately. | 12:47 |
*** mbuil has joined #openstack-ansible | 12:49 | |
*** ansmith has joined #openstack-ansible | 12:50 | |
openstackgerrit | Merged openstack/openstack-ansible-os_blazar master: Updated from OpenStack Ansible Tests https://review.openstack.org/567036 | 12:57 |
evrardjp | odyssey4me: yup :( | 13:03 |
evrardjp | odyssey4me: and then the summit | 13:03 |
evrardjp | odyssey4me: for https://review.openstack.org/#/c/568485/3 yes, but I'd like to see a review of hwoarang before giving my +2 :p | 13:04 |
evrardjp | I am already positive to the idea :) | 13:05 |
evrardjp | or just a comment "I am fine with that" :D | 13:05 |
evrardjp | (that's me an my carefulness) | 13:05 |
odyssey4me | evrardjp and cloudnull has answered in https://review.openstack.org/568867 :) | 13:06 |
ansmith | odyssey4me, evrardjp: thanks for comments on messaging changes, I should have asked re. the need to have messaging namespace per service | 13:07 |
ansmith | I could see that it was supported but wasnt sure it was needed | 13:08 |
odyssey4me | ansmith yeah, it's probably an edge case for very large clouds, given that the request actually came from rax pubcloud | 13:08 |
odyssey4me | but there's no harm in doing in - unfortunately it does mean another round of revisions | 13:09 |
*** cmart has joined #openstack-ansible | 13:09 | |
ansmith | no worries, happy to oblige :) | 13:09 |
odyssey4me | there was only one other thing bugging me, and that was the transport url being fairly unreadable - if you could figure out a way to simplify that line it'd be great - both evrardjp and I made some suggestions | 13:09 |
ansmith | yes, I saw your suggestions, will take a pass at making it easier to understand. I started with what was already in place... | 13:10 |
* mnaser hasn't been as active | 13:10 | |
mnaser | summit things, sorry everyone | 13:11 |
odyssey4me | ansmith yeah, what was there was just as unreadable which is why I initially ignored it - but if we can make it better in this round, then it saves some more patches on top | 13:11 |
evrardjp | odyssey4me: ansmith I think it's worth doing later, as an optimization. The first rounds of commits are good. | 13:14 |
evrardjp | I mean if we change the scoping :D | 13:14 |
ansmith | evrardjp: ack | 13:14 |
odyssey4me | yeah, I'm totally not requiring a change there - very happy to leave the current unreadable thing in place and to make something nicer later... | 13:15 |
*** armaan has quit IRC | 13:15 | |
odyssey4me | scope creep can be a terrible thing | 13:15 |
*** armaan has joined #openstack-ansible | 13:15 | |
evrardjp | yeah | 13:15 |
hwoarang | ok my turn to ask for reviews | 13:19 |
hwoarang | https://review.openstack.org/#/c/567808/ https://review.openstack.org/#/c/568311/ please :) | 13:20 |
*** fghaas has left #openstack-ansible | 13:23 | |
evrardjp | ok | 13:23 |
evrardjp | may I ask reviews on ppt files? | 13:24 |
evrardjp | :p | 13:24 |
hwoarang | -ENOFOUND | 13:24 |
odyssey4me | hwoarang oh nice, thanks for simplifying the fact setting along with this | 13:24 |
hwoarang | thank you for the prototype ;p | 13:25 |
dmsimard|off | by the way, this vscode extension for Ansible is REALLY good: https://github.com/VSChina/vscode-ansible | 13:25 |
odyssey4me | hwoarang both lgtm, good work :) | 13:27 |
hwoarang | thank you for the reviews | 13:28 |
evrardjp | dmsimard|off: yes :) | 13:29 |
evrardjp | vscode is pretty awesome | 13:29 |
*** esberglu has joined #openstack-ansible | 13:32 | |
mgariepy | does anyone knows if can we specify another library path for a particular playbook ? | 13:36 |
openstackgerrit | Taseer Ahmed proposed openstack/openstack-ansible-os_blazar master: Introduce os_blazar role in gerrit https://review.openstack.org/567112 | 13:37 |
evrardjp | mgariepy: playbook or role? | 13:38 |
jrosser | mgariepy: LD_LIBRARY_PATH? | 13:38 |
jrosser | which kind :) ? | 13:38 |
mgariepy | when deploying ceph with lvm I need to add the ceph-ansible library path to get the ceph_volume module. | 13:38 |
evrardjp | mgariepy: use the env var, at worse. | 13:38 |
evrardjp | mgariepy: isn't this done by default in our integrated repo? | 13:38 |
mgariepy | openstack-ansible ceph-install.yml -M /etc/ansible/roles/ceph-ansible/library/ --tag ceph-osd | 13:39 |
mgariepy | i need that in queens when deploying ceph with OSA | 13:39 |
*** sdake_ is now known as sdake | 13:39 | |
*** jwitko_ has joined #openstack-ansible | 13:39 | |
mgariepy | evrardjp, where ? | 13:40 |
*** throwsb1 has joined #openstack-ansible | 13:41 | |
openstackgerrit | wangqi proposed openstack/openstack-ansible master: Replace deprecated library function os.popen() with subprocess https://review.openstack.org/569116 | 13:44 |
mgariepy | jrosser, ansible lib :) | 13:45 |
openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible master: add Ceph-ansible library https://review.openstack.org/569119 | 13:51 |
evrardjp | mgariepy: I don't think we have this | 13:51 |
mgariepy | ^^ | 13:52 |
evrardjp | andymccr: logan- didn't we plan to add the ceph libraries into the path by default? | 13:52 |
andymccr | https://review.openstack.org/#/c/550749/ | 13:53 |
andymccr | should probably add the action plugin | 13:53 |
*** armaan has quit IRC | 13:56 | |
*** armaan has joined #openstack-ansible | 13:57 | |
andymccr | it was getting lint failures | 13:58 |
openstackgerrit | Andy McCrae proposed openstack/openstack-ansible master: Unpin ceph-ansible and add ceph-ansible library path https://review.openstack.org/550749 | 13:58 |
andymccr | but i couldnt figure out why | 13:58 |
andymccr | if anybody who is a lint expert could take a look that'd be awesome :) | 13:58 |
*** ianychoi_ has quit IRC | 13:59 | |
*** armaan has quit IRC | 14:00 | |
*** armaan has joined #openstack-ansible | 14:00 | |
openstackgerrit | Merged openstack/openstack-ansible-repo_server stable/queens: Support devpi as the upstream pip server https://review.openstack.org/568586 | 14:00 |
mnaser | andymccr: odd, that linter seems to be messing things up, the task is prefectly ok | 14:03 |
mnaser | perfectly* | 14:03 |
andymccr | mnaser: yeah. i was squinting hard at it | 14:03 |
andymccr | i was kinda hoping i had pebkac'd it somewhere and that somebody could point it out to me :D | 14:04 |
mnaser | andymccr: https://github.com/ansible/ansible/issues/38220 | 14:04 |
mnaser | please don't tell me ceph introduced a 'vars' module | 14:04 |
mnaser | they're very good at not namespacing stuff :X | 14:05 |
andymccr | hmm | 14:05 |
andymccr | i dont think a vars one | 14:05 |
mnaser | nope, don't see one either | 14:05 |
andymccr | they are introducing a validate.py soon (which is why i included the actions plugins) | 14:05 |
SamYaple | andymccr: i dont think youve messed things up. it looks like the repo server isnt running. i dont see how you patch could have caused that | 14:05 |
andymccr | SamYaple: right?! | 14:06 |
SamYaple | unless those directories dont exist when they are first accessed? | 14:06 |
SamYaple | that might be causing problems? | 14:06 |
andymccr | ahhh | 14:06 |
andymccr | that would make sense | 14:06 |
andymccr | lemme test that | 14:06 |
*** armaan has quit IRC | 14:15 | |
*** armaan has joined #openstack-ansible | 14:15 | |
openstackgerrit | Merged openstack/openstack-ansible-os_barbican master: Add support for using distribution packages for OpenStack services https://review.openstack.org/567808 | 14:16 |
*** johanssone has quit IRC | 14:16 | |
*** ianychoi has joined #openstack-ansible | 14:18 | |
openstackgerrit | Merged openstack/openstack-ansible stable/ocata: Update lxc_hosts role SHA for cache resolver fix https://review.openstack.org/569015 | 14:23 |
openstackgerrit | Merged openstack/openstack-ansible-os_aodh master: Add support for using distribution packages for OpenStack services https://review.openstack.org/568311 | 14:23 |
mgariepy | andymccr, /etc/ansible/roles/plugins/action/config_template.py /etc/ansible/roles/ceph-ansible/plugins/actions/config_template.py | 14:23 |
*** kstev has joined #openstack-ansible | 14:23 | |
mgariepy | can this be the issue ? | 14:23 |
andymccr | mgariepy: i dont thinkso, i hadnt included actions before (failed the same way) | 14:24 |
*** KeithMnemonic has joined #openstack-ansible | 14:28 | |
*** openstacking_123 has joined #openstack-ansible | 14:28 | |
*** hamzy_ has joined #openstack-ansible | 14:28 | |
*** hamzy has quit IRC | 14:29 | |
*** dxiri has joined #openstack-ansible | 14:29 | |
*** pvradu has quit IRC | 14:30 | |
*** openstacking_123 has quit IRC | 14:33 | |
openstackgerrit | kourosh vivan proposed openstack/openstack-ansible-os_ceilometer stable/pike: Adds ceilometer-upgrade task https://review.openstack.org/569128 | 14:33 |
hwoarang | andymccr: odyssey4me: evrardjp: i am trying to understand the ceph failures on opensuse http://logs.openstack.org/periodic/git.openstack.org/openstack/openstack-ansible/master/openstack-ansible-deploy-ceph-opensuse-423/1f240b1/job-output.txt.gz#_2018-05-17_07_37_49_416603 .Do you happen to have any suggestions what that error suggests to me? a plain virsh secret-define on vanilla opensuse works so i think nova may not be configured | 14:40 |
hwoarang | properly | 14:40 |
mgariepy | hwoarang, once osa failed you cannot run it ? | 14:43 |
*** epalper has quit IRC | 14:43 | |
hwoarang | mgariepy: i'm not sure i understand | 14:43 |
hwoarang | ah you mean that vanilla works, but after osa it fails | 14:43 |
*** epalper has joined #openstack-ansible | 14:43 | |
*** hamzy_ is now known as hamzy | 14:43 | |
hwoarang | yeah not sure i am running on vagrant now but it's slooooow | 14:43 |
hwoarang | and because it's vagrant (ie no kvm there) it may configure nova in a different way | 14:44 |
mgariepy | seems like a virsh misconfiguration to me more than a nova one. | 14:44 |
*** dxiri has quit IRC | 14:45 | |
hwoarang | yeah possibly but i am not very familiar with the nova role so i was looking for some quick pointers before i dig deeper :) | 14:45 |
openstackgerrit | Merged openstack/openstack-ansible stable/pike: Update lxc_hosts role SHA for cache resolver fix https://review.openstack.org/569014 | 14:46 |
openstackgerrit | Merged openstack/openstack-ansible stable/queens: Update lxc_hosts role SHA for cache resolver fix https://review.openstack.org/569013 | 14:46 |
openstackgerrit | Merged openstack/openstack-ansible master: Convert rsyslog to an include_task https://review.openstack.org/556002 | 14:46 |
openstackgerrit | Merged openstack/openstack-ansible stable/queens: Ensure facts are updated for new containers https://review.openstack.org/568626 | 14:46 |
mgariepy | hwoarang, what pkg are installed on your vanilla install libvirt* ? | 14:47 |
mgariepy | maybe you are missing something | 14:48 |
hwoarang | hmm | 14:48 |
hwoarang | vanilla is vanilla + libvirt (and deps). the nova role should have the same packages since libvirt should pull the same packages | 14:49 |
mgariepy | libvirt dependency maybe ? | 14:50 |
hwoarang | i think it could be a libvirt configuration issue. i may have to compare the conf files but sadly the integration repo does not collect any /etc/ information yet | 14:50 |
mgariepy | hwoarang, https://github.com/openstack/openstack-ansible-os_nova/blob/master/vars/suse-42.yml#L41-L56 | 14:50 |
*** weezS has joined #openstack-ansible | 14:51 | |
hwoarang | hmm | 14:52 |
hwoarang | let me check | 14:52 |
openstackgerrit | Merged openstack/openstack-ansible stable/ocata: Run inventory-manage.py from the ansible-runtime venv https://review.openstack.org/568466 | 14:53 |
hwoarang | could be that libvirt-daemon-driver-secret is missing | 14:54 |
hwoarang | lets try that thanks! | 14:54 |
*** cjloader has joined #openstack-ansible | 14:54 | |
mgariepy | yep i would guess it's the issue :) | 14:56 |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_nova master: SUSE: Add libvirt-daemon-driver-secret to list of packages https://review.openstack.org/569137 | 14:56 |
*** dxiri has joined #openstack-ansible | 14:58 | |
*** Guest40373 has joined #openstack-ansible | 14:59 | |
*** blinkiz has quit IRC | 15:03 | |
openstackgerrit | Marc Gariépy (mgariepy) proposed openstack/openstack-ansible master: Only bind to internal IP. https://review.openstack.org/569141 | 15:08 |
*** lvdombrkr89 has quit IRC | 15:09 | |
*** epalper has quit IRC | 15:20 | |
*** KeithMnemonic has quit IRC | 15:22 | |
openstackgerrit | Merged openstack/openstack-ansible stable/pike: Run inventory-manage.py from the ansible-runtime venv https://review.openstack.org/568464 | 15:24 |
*** gyee has joined #openstack-ansible | 15:41 | |
*** lvdombrkr has joined #openstack-ansible | 15:43 | |
*** foutatoro has joined #openstack-ansible | 15:47 | |
*** armaan has quit IRC | 15:50 | |
*** armaan has joined #openstack-ansible | 15:51 | |
*** chyka has joined #openstack-ansible | 16:10 | |
* jrosser gets the ops repo elk+grafana running..... nice :) | 16:17 | |
openstackgerrit | Markos Chandras (hwoarang) proposed openstack/openstack-ansible-os_ceilometer master: Add support for using distribution packages for OpenStack services https://review.openstack.org/569161 | 16:20 |
*** DanyC has quit IRC | 16:22 | |
*** DanyC has joined #openstack-ansible | 16:22 | |
*** DanyC has quit IRC | 16:27 | |
*** pvradu has joined #openstack-ansible | 16:31 | |
*** bhujay has quit IRC | 16:39 | |
*** guhcampos has joined #openstack-ansible | 16:48 | |
*** mbuil has quit IRC | 16:51 | |
*** dxiri has quit IRC | 16:55 | |
*** Guest40373 has quit IRC | 16:58 | |
*** mma has quit IRC | 16:59 | |
*** mma has joined #openstack-ansible | 16:59 | |
*** chhavi__ has quit IRC | 17:18 | |
*** DanyC has joined #openstack-ansible | 17:18 | |
openstackgerrit | Merged openstack/openstack-ansible master: infra-journal-remote: Skip playbook on empty emtpy log_hosts group https://review.openstack.org/569039 | 17:20 |
*** lvdombrkr has quit IRC | 17:23 | |
evrardjp | hwoarang: sorry I didn't got the chance tocheck at this. | 17:24 |
evrardjp | OMG it's already that time? | 17:25 |
evrardjp | that flew. | 17:25 |
*** udesale has joined #openstack-ansible | 17:32 | |
*** udesale has quit IRC | 17:33 | |
*** radeks_ has quit IRC | 17:36 | |
*** ivve has joined #openstack-ansible | 17:41 | |
*** sawblade6 has quit IRC | 17:46 | |
*** dxiri has joined #openstack-ansible | 17:49 | |
*** tomtom001 has quit IRC | 17:52 | |
*** tomtom001 has joined #openstack-ansible | 17:54 | |
*** sawblade6 has joined #openstack-ansible | 18:00 | |
*** guhcampos_ has joined #openstack-ansible | 18:01 | |
*** guhcampos has quit IRC | 18:04 | |
*** pvradu has quit IRC | 18:05 | |
*** cjloader_ has joined #openstack-ansible | 18:07 | |
*** cjloader has quit IRC | 18:07 | |
*** cjloader_ has quit IRC | 18:07 | |
*** cjloader has joined #openstack-ansible | 18:07 | |
*** tomtom001 has quit IRC | 18:09 | |
*** tomtom001 has joined #openstack-ansible | 18:12 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Begin testing opensuse and centos with nspawn https://review.openstack.org/568197 | 18:17 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Use fan out when deploying with nspawn https://review.openstack.org/568331 | 18:19 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Begin testing opensuse and centos with nspawn https://review.openstack.org/568197 | 18:21 |
*** cmart has quit IRC | 18:28 | |
cloudnull | jrosser nice! | 18:33 |
cloudnull | is that on a single logging node? | 18:33 |
*** cjloader has quit IRC | 18:34 | |
*** cjloader has joined #openstack-ansible | 18:35 | |
*** cjloader has quit IRC | 18:35 | |
*** cjloader has joined #openstack-ansible | 18:36 | |
jrosser | cloudnull: i did 3 | 18:38 |
jrosser | becasue i could :) | 18:39 |
jrosser | most of the difficulty i had was entirely self induced needing to mirror the repos and get proxy support working for the install | 18:39 |
openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Remove default pip_links value https://review.openstack.org/549012 | 18:40 |
*** dxiri has quit IRC | 18:47 | |
*** jwitko__ has joined #openstack-ansible | 18:49 | |
*** jwitko_ has quit IRC | 18:53 | |
*** chyka has quit IRC | 19:00 | |
*** cmart has joined #openstack-ansible | 19:05 | |
mmercer | mgariepy: i think you misunderstood what i was saying -- the reason that its bad is because the option is embed 2x in the same config file -- see lines 19,20 and 63 as an example. its not the only option that is dual defined, but its one of the more critical examples | 19:12 |
*** chyka has joined #openstack-ansible | 19:13 | |
mgariepy | mmercer, i did understood, i was just saying that some option can be there more than once | 19:13 |
mmercer | * nods * its generally recommended not to do it that way though, as you can run into mixed behavior/unexpected behavior | 19:14 |
mgariepy | ie: it's perfectly fine to have multiple ListenAddress in sshd_config :) | 19:14 |
mmercer | true, but what if you are expecting it to be the one you define as opposed to the one that already exists, now you have both | 19:15 |
mmercer | with many of the parameters, the assumption should err on the 'explicitly defined' | 19:15 |
mmercer | since we are referring to these as hardening suggestions | 19:15 |
mgariepy | is there a way to validate sshd config automatically ? | 19:15 |
mmercer | -t | 19:16 |
mmercer | validate: /usr/bin/sshd -t | 19:16 |
mmercer | but as several directive can be treated additively, validation wont necessarily prevent multi definition | 19:17 |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add mount options whenever formatting disks https://review.openstack.org/569226 | 19:18 |
openstackgerrit | David Wilde proposed openstack/openstack-ansible-ops master: Fix MNAIO Flat Networking https://review.openstack.org/569227 | 19:21 |
d34dh0r53 | cloudnull: ^ when you get a chance | 19:22 |
*** guhcampos_ has quit IRC | 19:23 | |
cloudnull | ^^ odyssey4me when you get a chance | 19:28 |
cloudnull | :) | 19:28 |
d34dh0r53 | :) | 19:29 |
*** lkoranda has quit IRC | 19:30 | |
*** pvradu has joined #openstack-ansible | 19:34 | |
*** radeks has joined #openstack-ansible | 19:34 | |
*** goldenfr_ has quit IRC | 19:36 | |
*** dxiri has joined #openstack-ansible | 19:46 | |
*** pvradu has quit IRC | 19:47 | |
jrosser | cloudnull: did you get the multi-system-metrics dashboard to show the derivative of network in/out bytes? | 19:50 |
*** cmart has quit IRC | 19:53 | |
cloudnull | the dashboard should show that when >1 host is selected, however its possible I have a logic bug there. | 19:54 |
jrosser | the network packet io graphs looks like flatlines - then i zoom out to "today" and its ahha! as they have a nice linear-ish slope to them | 19:56 |
jrosser | even with multiple hosts selected i think? | 19:56 |
*** cmart has joined #openstack-ansible | 20:07 | |
*** gkadam has quit IRC | 20:08 | |
cloudnull | I wonder if thats due to something in the query ? | 20:09 |
jrosser | like this i think https://pastebin.com/5zRRLZcS | 20:10 |
*** notmyname has joined #openstack-ansible | 20:23 | |
notmyname | evrardjp: around? | 20:24 |
notmyname | evrardjp: pm sent. I got your stats for you | 20:25 |
notmyname | find me in #openstack-swift (or about a dozen other openstack channels) if you have questions | 20:26 |
evrardjp | for that I can be back :) | 20:26 |
evrardjp | notmyname: thank you! | 20:26 |
*** DanyC has quit IRC | 20:43 | |
*** DanyC has joined #openstack-ansible | 20:43 | |
*** ansmith has quit IRC | 20:46 | |
*** DanyC has quit IRC | 20:47 | |
*** throwsb1 has quit IRC | 20:53 | |
*** vnogin has joined #openstack-ansible | 21:04 | |
*** notmyname has quit IRC | 21:15 | |
*** mma has quit IRC | 21:16 | |
*** esberglu has quit IRC | 21:16 | |
*** mma has joined #openstack-ansible | 21:17 | |
*** notmyname has joined #openstack-ansible | 21:18 | |
*** jwitko_ has joined #openstack-ansible | 21:21 | |
*** jwitko_ has quit IRC | 21:21 | |
*** mma has quit IRC | 21:21 | |
*** jwitko_ has joined #openstack-ansible | 21:21 | |
*** jwitko__ has quit IRC | 21:25 | |
*** cjloader_ has joined #openstack-ansible | 21:28 | |
*** cjloader has quit IRC | 21:28 | |
*** jwitko_ has quit IRC | 21:28 | |
*** cjloader_ has quit IRC | 21:33 | |
*** nurdie has joined #openstack-ansible | 21:33 | |
*** ansmith has joined #openstack-ansible | 21:35 | |
*** KeithMnemonic has joined #openstack-ansible | 21:39 | |
*** hughsaunders has quit IRC | 21:44 | |
*** hughsaunders has joined #openstack-ansible | 21:46 | |
*** nurdie has quit IRC | 21:47 | |
*** chyka_ has joined #openstack-ansible | 21:52 | |
*** kstev has quit IRC | 21:55 | |
*** chyka has quit IRC | 21:56 | |
*** chyka_ has quit IRC | 21:57 | |
openstackgerrit | Merged openstack/openstack-ansible master: Remove default pip_links value https://review.openstack.org/549012 | 22:01 |
*** vnogin has quit IRC | 22:17 | |
*** cmart has quit IRC | 22:22 | |
*** foutatoro has quit IRC | 22:25 | |
*** cjloader has joined #openstack-ansible | 22:26 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add mount options whenever formatting disks https://review.openstack.org/569226 | 22:27 |
*** chyka has joined #openstack-ansible | 22:28 | |
*** cjloader has quit IRC | 22:28 | |
*** cjloader has joined #openstack-ansible | 22:28 | |
*** chyka has quit IRC | 22:32 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add mount options whenever formatting disks https://review.openstack.org/569226 | 22:35 |
*** threestrands has joined #openstack-ansible | 22:37 | |
*** d34dh0r53 has quit IRC | 22:47 | |
*** d34dh0r53 has joined #openstack-ansible | 22:47 | |
*** cloudnull has quit IRC | 22:48 | |
*** eglute has quit IRC | 22:48 | |
*** cloudnull has joined #openstack-ansible | 22:48 | |
*** eglute has joined #openstack-ansible | 22:49 | |
*** vnogin has joined #openstack-ansible | 22:51 | |
*** vnogin has quit IRC | 22:55 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible master: Add mount options whenever formatting disks https://review.openstack.org/569226 | 22:59 |
*** vnogin has joined #openstack-ansible | 23:07 | |
*** radeks has quit IRC | 23:11 | |
*** vnogin has quit IRC | 23:11 | |
*** vnogin has joined #openstack-ansible | 23:18 | |
*** vnogin has quit IRC | 23:23 | |
*** cjloader has quit IRC | 23:29 | |
*** klamath has joined #openstack-ansible | 23:35 | |
*** tosky has quit IRC | 23:36 | |
openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-nspawn_hosts master: Update prep scripts to pull in a few updates https://review.openstack.org/569292 | 23:37 |
*** vnogin has joined #openstack-ansible | 23:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!