| *** Marga_ has quit IRC | 00:05 | |
| *** Marga_ has joined #openstack-operators | 00:05 | |
| *** markvoelker has quit IRC | 00:19 | |
| *** furlongm has joined #openstack-operators | 00:30 | |
| *** SimonChung2 has quit IRC | 00:32 | |
| *** SimonChung has joined #openstack-operators | 00:32 | |
| *** SimonChung1 has joined #openstack-operators | 00:33 | |
| *** SimonChung1 has quit IRC | 00:33 | |
| *** SimonChung has quit IRC | 00:33 | |
| *** SimonChung2 has joined #openstack-operators | 00:33 | |
| *** mdorman has quit IRC | 00:36 | |
| *** david-lyle is now known as david-lyle_afk | 00:41 | |
| *** Marga_ has quit IRC | 00:46 | |
| *** alop has quit IRC | 00:46 | |
| *** Marga_ has joined #openstack-operators | 00:46 | |
| *** Marga_ has quit IRC | 00:47 | |
| *** Marga_ has joined #openstack-operators | 00:47 | |
| *** markvoelker has joined #openstack-operators | 00:52 | |
| *** markvoelker has quit IRC | 00:59 | |
| *** trad511 has quit IRC | 01:04 | |
| *** Marga_ has quit IRC | 01:06 | |
| *** Marga_ has joined #openstack-operators | 01:07 | |
| *** Marga_ has quit IRC | 01:12 | |
| *** markvoelker has joined #openstack-operators | 01:29 | |
| *** SimonChung has joined #openstack-operators | 01:37 | |
| *** SimonChung2 has quit IRC | 01:37 | |
| *** devlaps has quit IRC | 01:55 | |
| *** SimonChung has quit IRC | 01:57 | |
| *** Piet has quit IRC | 02:45 | |
| *** SimonChung has joined #openstack-operators | 03:01 | |
| *** SimonChung1 has joined #openstack-operators | 03:05 | |
| *** SimonChung has quit IRC | 03:07 | |
| *** harlowja_ is now known as harlowja_away | 03:15 | |
| *** snk has quit IRC | 04:13 | |
| *** snk has joined #openstack-operators | 04:16 | |
| *** rbrooker has quit IRC | 04:39 | |
| *** markvoelker has quit IRC | 04:42 | |
| *** markvoelker has joined #openstack-operators | 04:42 | |
| *** markvoelker has quit IRC | 04:46 | |
| *** markvoelker has joined #openstack-operators | 05:13 | |
| *** sanjayu has joined #openstack-operators | 05:15 | |
| *** markvoelker has quit IRC | 05:17 | |
| *** VW_ has joined #openstack-operators | 05:19 | |
| *** VW_ has quit IRC | 05:21 | |
| *** VW_ has joined #openstack-operators | 05:22 | |
| *** sandywalsh has joined #openstack-operators | 05:38 | |
| *** sandywalsh_ has quit IRC | 05:40 | |
| *** Marga_ has joined #openstack-operators | 05:40 | |
| *** Marga_ has quit IRC | 05:41 | |
| *** Marga_ has joined #openstack-operators | 05:42 | |
| *** zigo has quit IRC | 05:54 | |
| *** zigo has joined #openstack-operators | 05:55 | |
| *** markvoelker has joined #openstack-operators | 06:13 | |
| *** markvoelker has quit IRC | 06:18 | |
| *** signed8bit has joined #openstack-operators | 06:22 | |
| *** signed8bit_ZZZzz has quit IRC | 06:23 | |
| *** belmoreira has joined #openstack-operators | 06:26 | |
| *** VW_ has quit IRC | 06:28 | |
| *** slaweq has joined #openstack-operators | 06:41 | |
| *** slaweq has quit IRC | 07:08 | |
| *** slaweq has joined #openstack-operators | 07:10 | |
| *** markvoelker has joined #openstack-operators | 07:14 | |
| *** markvoelker has quit IRC | 07:19 | |
| *** VW_ has joined #openstack-operators | 07:27 | |
| *** VW_ has quit IRC | 07:27 | |
| *** VW__ has joined #openstack-operators | 07:28 | |
| *** Marga_ has quit IRC | 07:29 | |
| *** slaweq has quit IRC | 08:02 | |
| *** harlowja_away has quit IRC | 08:03 | |
| *** slaweq has joined #openstack-operators | 08:04 | |
| *** slaweq has quit IRC | 08:09 | |
| *** markvoelker has joined #openstack-operators | 08:15 | |
| *** markvoelker has quit IRC | 08:20 | |
| *** zz_avozza is now known as zz_zz_avozza | 08:44 | |
| *** bvandenh has joined #openstack-operators | 09:10 | |
| *** matrohon has joined #openstack-operators | 09:14 | |
| *** markvoelker has joined #openstack-operators | 09:16 | |
| *** derekh has joined #openstack-operators | 09:19 | |
| *** markvoelker has quit IRC | 09:21 | |
| *** racedo_ has joined #openstack-operators | 09:31 | |
| *** VW__ has quit IRC | 09:34 | |
| *** VW_ has joined #openstack-operators | 09:39 | |
| *** VW_ has quit IRC | 09:48 | |
| *** VW_ has joined #openstack-operators | 09:49 | |
| *** zz_zz_avozza is now known as avozza | 09:52 | |
| *** markvoelker has joined #openstack-operators | 10:17 | |
| *** markvoelker has quit IRC | 10:22 | |
| *** pboros has joined #openstack-operators | 10:36 | |
| *** VW_ has quit IRC | 10:39 | |
| *** pboros_ has joined #openstack-operators | 10:57 | |
| *** pboros has quit IRC | 10:59 | |
| *** blair has quit IRC | 11:02 | |
| *** avozza is now known as zz_avozza | 11:07 | |
| *** VW_ has joined #openstack-operators | 11:07 | |
| *** zz_avozza is now known as avozza | 11:16 | |
| *** markvoelker has joined #openstack-operators | 11:18 | |
| *** markvoelker has quit IRC | 11:23 | |
| *** VW_ has quit IRC | 11:25 | |
| *** VW_ has joined #openstack-operators | 11:27 | |
| *** VW_ has quit IRC | 11:27 | |
| *** VW_ has joined #openstack-operators | 11:28 | |
| *** blair has joined #openstack-operators | 11:32 | |
| *** blair has quit IRC | 11:34 | |
| *** blairo has joined #openstack-operators | 11:34 | |
| *** blairo has quit IRC | 11:39 | |
| *** markvoelker has joined #openstack-operators | 12:19 | |
| *** markvoelker has quit IRC | 12:24 | |
| *** VW_ has quit IRC | 12:29 | |
| *** VW_ has joined #openstack-operators | 12:30 | |
| *** avozza is now known as zz_avozza | 12:42 | |
| *** VW_ has quit IRC | 12:47 | |
| *** blair has joined #openstack-operators | 12:50 | |
| *** blair has quit IRC | 12:55 | |
| *** markvoelker has joined #openstack-operators | 13:02 | |
| *** zz_avozza is now known as avozza | 13:10 | |
| *** bilco105 has quit IRC | 13:23 | |
| *** bilco105 has joined #openstack-operators | 13:24 | |
| *** markvoelker_ has joined #openstack-operators | 13:42 | |
| *** markvoelker has quit IRC | 13:42 | |
| *** racedo_ is now known as racedo | 13:59 | |
| *** cpschult has joined #openstack-operators | 14:10 | |
| *** trad511 has joined #openstack-operators | 14:14 | |
| *** VW_ has joined #openstack-operators | 14:22 | |
| *** j05h has quit IRC | 14:25 | |
| *** VW_ has quit IRC | 14:26 | |
| *** VW_ has joined #openstack-operators | 14:33 | |
| *** sanjayu has quit IRC | 14:33 | |
| *** csoukup has joined #openstack-operators | 14:34 | |
| *** markvoelker_ has quit IRC | 14:38 | |
| *** markvoelker has joined #openstack-operators | 14:39 | |
| *** j05h has joined #openstack-operators | 14:51 | |
| *** delattec has joined #openstack-operators | 15:14 | |
| *** avozza is now known as zz_avozza | 15:36 | |
| *** mdorman has joined #openstack-operators | 15:47 | |
| *** jaypipes has quit IRC | 15:51 | |
| *** jaypipes has joined #openstack-operators | 16:01 | |
| *** david-lyle_afk is now known as david-lyle | 16:04 | |
| *** alaski has joined #openstack-operators | 16:06 | |
| VW_ | do the meeting commands work in these channels? | 16:08 |
|---|---|---|
| klindgren__ | I dont think so | 16:08 |
| mdorman | i doubt it, but could try | 16:08 |
| klindgren__ | I dont see an openstack bot in here | 16:08 |
| VW_ | no worries | 16:09 |
| VW_ | I'll take notes in the etherpad | 16:09 |
| VW_ | if you folks could help | 16:09 |
| mdorman | could you paste the link? | 16:09 |
| belmoreira | https://etherpad.openstack.org/p/Large-Deployment-Team-Meetings | 16:09 |
| VW_ | yep - that | 16:10 |
| VW_ | thanks belmoreira | 16:10 |
| VW_ | Ok, let's get started | 16:10 |
| VW_ | alaski: you here? | 16:10 |
| alaski | yep | 16:10 |
| VW_ | cool - thanks! | 16:10 |
| VW_ | so, folks, in the last meeting, there were questions about how cells v2 was coming along | 16:11 |
| VW_ | I asked alaski to join for a bit | 16:11 |
| mdorman | cool | 16:11 |
| VW_ | I know he has to run by mid-hour, but I thought we could catch up on the progress and find out how we can help | 16:11 |
| alaski | sure | 16:11 |
| alaski | there's been work on a few fronts | 16:12 |
| alaski | we've put some effort into getting cellsv1 tested better | 16:12 |
| alaski | there's a check job upstream that isn't voting yet, but is down to about 2 failures we need to fix to get it voting | 16:12 |
| alaski | for cellsv2 we've been focusing on the new database that's going to be added | 16:13 |
| alaski | that work is being tracked under https://blueprints.launchpad.net/nova/+spec/cells-v2-mapping right now | 16:13 |
| alaski | the goal is to have that in place with a few tables by the end of Kilo, though nothing will be using it yet | 16:14 |
| alaski | there's been some parallel work so that Nova will be able to communicate with multiple databases | 16:14 |
| alaski | but that should be mostly invisible to operators | 16:14 |
| VW_ | so, would this be in addition to the current DB we run at the region/top level, alaski? | 16:15 |
| VW_ | or replace it | 16:15 |
| mdorman | ^ my question too | 16:15 |
| alaski | this would replace it | 16:15 |
| VW_ | and by multiple, nova would talk to it and all the cells directly? | 16:15 |
| alaski | the idea is that there will be an api level database, and a cell level database, and they will now be different | 16:15 |
| alaski | VW_: yes | 16:15 |
| VW_ | got it | 16:15 |
| alaski | nova-api would talk to multiple cell databases directly | 16:15 |
| mdorman | are their plans for a migration strategy to go from v1 to v2 schema? or would that not be part of the kilo release? | 16:16 |
| alaski | that will not be done in Kilo | 16:16 |
| mdorman | alright. | 16:16 |
| alaski | but the basic idea is that the current nova db will become a cell db, and a current cell db can remain. migration would be primarily populating the new api db | 16:17 |
| alaski | so there's not a schema change, we're just adding a new one up top, and replacing the top one if you're currently using cells | 16:18 |
| mdorman | gottcha | 16:18 |
| mdorman | makes sense | 16:18 |
| alaski | and the primary focus in K is getting that db in place and getting Nova able to work with multiple dbs | 16:19 |
| alaski | L is going to get into migrating data, scheduling, networking and more | 16:19 |
| mdorman | kk | 16:19 |
| alaski | oh, and multiple message queues in K (hopefully) | 16:19 |
| alaski | discussions have begun with Neutron folks as to how the projects will work together with cells | 16:20 |
| *** j05h has quit IRC | 16:20 | |
| VW_ | awesome | 16:20 |
| mdorman | how would it look if multiple DBs are supported, but not multiple message queues? | 16:20 |
| alaski | mdorman: multiple message queues will need to be supported before cellsv2 can be used | 16:20 |
| mdorman | ok. | 16:21 |
| VW_ | wouldn't they be though? The new top level DB is basically telling the API which message queue to sue, right | 16:21 |
| VW_ | s/sue/use | 16:21 |
| alaski | VW_: right | 16:21 |
| alaski | we just need to add support for that in the code | 16:21 |
| alaski | but currently the focus in on the database | 16:21 |
| alaski | the summary is that we're mainly working on some underpinnings in K, and nothing will be usable until L | 16:22 |
| VW_ | awesome, alaski - thanks for sharing | 16:22 |
| VW_ | what can this group do to help? | 16:23 |
| *** zz_avozza is now known as avozza | 16:23 | |
| VW_ | I know that among operators, we have a larger perctange using (or needing to use for scale) cells | 16:23 |
| alaski | continuing to ask questions and providing feedback, especially when we get to scheduling and networking | 16:24 |
| *** med_` is now known as medberry | 16:24 | |
| VW_ | so folks, like myself, belmoreira and mdorman have a vested interest in helping it succeed :) | 16:24 |
| mdorman | :) | 16:24 |
| *** medberry is now known as medberry2 | 16:24 | |
| mdorman | alaski: what is the best avenue for that? are there specs that we can help review? other forums? | 16:24 |
| VW_ | any specific blueprints that you haven't gotten enough feedback on, alaski? | 16:24 |
| alaski | https://review.openstack.org/#/c/141486/ is way out of date now, but that's about scheduling | 16:25 |
| VW_ | cool - we'll try to stoke the fires on that one then | 16:25 |
| alaski | mdorman: specs are the best, but if there's a place I can raise visibility for other avenues please let me know | 16:25 |
| *** Piet has joined #openstack-operators | 16:26 | |
| VW_ | you might remind this group of your meeting times, alaski | 16:26 |
| mdorman | openstack-operators ML i think is probably the best for that | 16:26 |
| VW_ | we should probably be better about being in those too | 16:26 |
| alaski | mdorman: gotcha, I can send info there when new specs come up | 16:26 |
| alaski | https://wiki.openstack.org/wiki/Meetings/NovaCellsv2#Next_Meeting | 16:26 |
| VW_ | cool - we are trying to spread the word on that. I mentioned it last month during the alternate meeting time too | 16:27 |
| VW_ | I know you have to run, alaski. Thanks for joining | 16:27 |
| alaski | not a problem | 16:27 |
| mdorman | yeah, thanks for the insights | 16:28 |
| VW_ | there is an ops mid-cycle in two weeks as well. I won't make it, but I'll get any of the LDT folks there to help spred the word | 16:28 |
| alaski | feel free to ping me in #openstack-nova or on email with any further questions | 16:28 |
| *** j05h has joined #openstack-operators | 16:28 | |
| VW_ | that brings us to the other main thing I wanted to talk about | 16:29 |
| VW_ | the upcoming mid-cycle | 16:29 |
| VW_ | tom has an hour and a half set aside for LDT on Monday | 16:29 |
| *** delatte has joined #openstack-operators | 16:30 | |
| mdorman | looking back at previous meeting notes for follow up items we could address there... | 16:31 |
| VW_ | belmoreira, mdorman or any other Large Deployment Team folks - are you attending? | 16:31 |
| mdorman | i and klindgren__ are | 16:31 |
| belmoreira | i'm not attending | 16:31 |
| *** klindgren__ is now known as klindgren | 16:32 | |
| *** delattec has quit IRC | 16:32 | |
| mdorman | i think mfisch and dvorak are as well | 16:32 |
| VW_ | ok cool, mdorman. I'll have a couple folks from Rackspace there too. I just had family travel already scheduled | 16:32 |
| mdorman | yup | 16:32 |
| mdorman | so is this our last meeting before that meetup, or would we have one mor? | 16:33 |
| VW_ | this is our last one before | 16:34 |
| mdorman | ok. | 16:34 |
| VW_ | I'm almost wondering if it would be useful to build a list of wants around scheduling and cell/host control | 16:34 |
| VW_ | statuses, filters, etc | 16:34 |
| mdorman | seems like that could have some potential | 16:34 |
| VW_ | it would be something very tangible and clearly fits in a space where alaski and company need some feedback | 16:35 |
| VW_ | it could be less focused on specific specs and more of a wish/want list | 16:35 |
| VW_ | but could inlcude feedback on the spec above as well | 16:35 |
| mdorman | right | 16:35 |
| belmoreira | yes | 16:35 |
| VW_ | for example, I REALLY want to be able to "disable" a cell so no new builds go there | 16:35 |
| VW_ | the folks in Phill could circulate what they come up with back to the ML for those of us not there | 16:36 |
| VW_ | then we can finalize it in the March meeting, and pass it on to the devs | 16:36 |
| mdorman | i like it | 16:36 |
| VW_ | would be a huge win for the working group process in general | 16:36 |
| belmoreira | yes | 16:37 |
| mdorman | are there other general topics like that for any of the other projects where we could do something similar (neutron?) | 16:37 |
| belmoreira | there are other interesting topics that are going on that ops should discuss, in my view... | 16:37 |
| belmoreira | mdorman: :) | 16:37 |
| belmoreira | nova net -> neutron migration | 16:38 |
| VW_ | I agree, belmoreira. And many will be at the larger meeting. I'm just trying to make sure we make the most of the time slot Tom gave LDT | 16:38 |
| VW_ | but yeah, something Neutron-ish sounds and smells like a good place to focus some time | 16:39 |
| VW_ | :D | 16:39 |
| VW_ | Tom also gave me a preview of the Vancover schedule and we have an LDT timeslot there too | 16:40 |
| mdorman | nice | 16:41 |
| belmoreira | other point not really related with development is how LDT are doing ops (procedures) | 16:41 |
| belmoreira | it would be nice to have a discussion on this | 16:41 |
| VW_ | yeah, belmoreira - that is a good idea | 16:41 |
| belmoreira | for example, how to retire hundreds of servers | 16:41 |
| mdorman | you thinking sort of like a show-and-tel on that? | 16:41 |
| VW_ | That might be a great thing to do in Vancouver | 16:42 |
| VW_ | let me go see if I can still pull up the schedule | 16:42 |
| belmoreira | image management is other important topic in my view | 16:43 |
| VW_ | yes, belmoreira +eleventy billion :) | 16:43 |
| klindgren | yes | 16:43 |
| VW_ | I'm really wanting to hear more about your squid/caching stuff | 16:44 |
| VW_ | and we are looking at some active precaching on the hypervisor | 16:44 |
| VW_ | very early stages, though | 16:44 |
| klindgren | I have an ansible playbook to precache images on a server | 16:44 |
| klindgren | using bittorrent | 16:44 |
| belmoreira | that is really interesting | 16:45 |
| belmoreira | I would love to hear more about it | 16:45 |
| VW_ | sounds like if we took one of our two 40 minute session in Vancover to have an open discussion about image management, it wouldn't be a waste of time then? | 16:45 |
| VW_ | :) | 16:46 |
| mdorman | probably not | 16:46 |
| mdorman | should we try to include some time at the meetup to discuss wants for vancouver agenda? | 16:46 |
| klindgren | yea - I would love for "community public" and "provider public" to be segregated somehow | 16:46 |
| VW_ | cool then - consider it done pending finalization of the official schedule by the Foundation | 16:46 |
| belmoreira | great | 16:47 |
| *** dmsimard_away is now known as dmsimard | 16:47 | |
| klindgren | and that when people delete images you dont end up with vm's that can no longer migrate because the backing image has been destroyed | 16:47 |
| VW_ | we probably could cover that in Philly, mdorman, but we only have an hour and a half | 16:47 |
| VW_ | we still have March and April LDT meetings too | 16:47 |
| mdorman | ah yeah good point. | 16:47 |
| VW_ | yeah klindgren - there is some useful best practices there | 16:48 |
| VW_ | I didn't chime into the thread a while back, but we actually release a new image with each update for that reason | 16:48 |
| VW_ | but we are working longer term, for a fancier way, so that we don't have to keep as many old images around | 16:48 |
| *** alop has joined #openstack-operators | 16:49 | |
| mdorman | cool | 16:49 |
| VW_ | for now, I'd like to keep one of the Vancover sessions (40 minutes) open to produce something, but if we don't have a canidate like the scheduling stuff, we can have another sharing session like the images work | 16:50 |
| VW_ | or expand images to both | 16:50 |
| mdorman | sounds good. | 16:50 |
| VW_ | because I'm sure folks have plenty to talk about there ;) | 16:50 |
| klindgren | Slightly off topic - but is has anyone else recently updated to juno and now getting neutron read timeouts? | 16:51 |
| dvorak | what does the message look like? | 16:51 |
| klindgren | We are getting it randomly - but consistently on metadata calls | 16:52 |
| VW_ | check for DB deadlocks | 16:52 |
| klindgren | Failed to get metadata for ip: 10.22.253.11 | 16:52 |
| klindgren | <trace> | 16:52 |
| klindgren | ConnectionFailed: Connection to neutron failed: HTTPSConnectionPool(host='openstack.int.godaddy.com', port=9696): Read timed out. (read timeout=60) | 16:52 |
| VW_ | a recent deploy of ours (which is late 2014 code) has seen an increase in those on the Neutron side | 16:53 |
| VW_ | caused us a lot of headache | 16:53 |
| klindgren | what solved it? | 16:53 |
| VW_ | we are still solving, actually | 16:53 |
| VW_ | :\ | 16:54 |
| klindgren | ah - anyway would love to talk about what you tried after the meeting | 16:54 |
| VW_ | have you already put that on the list? I can try and get some folks to chime in that know more about the specifics | 16:54 |
| klindgren | nope - was going to ask in #openstack-neutron but to be honest- that always seems like a huge waste | 16:54 |
| VW_ | send it to #openstack-neutron and #openstack-opeators and I'll try to kick some folks to take a look | 16:55 |
| klindgren | kk | 16:55 |
| VW_ | we may be dealing with completely different problems | 16:55 |
| VW_ | but if our neutron folks have any insight that will help, I'd like them to share | 16:55 |
| klindgren | agreed but I tired things that "should fix it" | 16:55 |
| klindgren | increased workers/timeouts | 16:56 |
| klindgren | and nothing helped | 16:56 |
| klindgren | :-/ | 16:56 |
| VW_ | I'm in our UK office till the 28th, but I can harrass them from afar and, worst case, in person when I get back | 16:56 |
| VW_ | OK, well, sorry for the meeting room confusion | 16:56 |
| * VW_ is clearly way behind on wiki work | 16:56 | |
| VW_ | but thank you guys for joining today. Looks like we have some useful stuff to work on | 16:57 |
| VW_ | mdorman, klindgren I might pull you two together with a couple of my RS folks going so we can plan ahead for the session based on the above | 16:57 |
| klindgren | kk | 16:58 |
| mdorman | sounds good. | 16:58 |
| VW_ | alright - well thinks guys | 16:59 |
| VW_ | I'll let you get back to all your operations related fires :) | 16:59 |
| klindgren | :-) | 16:59 |
| belmoreira | thanks :) | 16:59 |
| mgagne | anyone using custom weighers in Nova? I find it inconvenient that Nova will load ALL weighers found in scheduler/weights by default unlike filters. And tests breaking because I added one (or more) weighers in scheduler/weights. | 17:02 |
| *** Marga_ has joined #openstack-operators | 17:08 | |
| *** Marga_ has quit IRC | 17:09 | |
| *** Marga_ has joined #openstack-operators | 17:10 | |
| VW_ | mgagne: I pinged one of our devs with the above. He wasn't quite sure what you were asking, but mentioned the following option to control the weights used: | 17:19 |
| VW_ | cfg.ListOpt('scheduler_weight_classes', | 17:19 |
| VW_ | default=['nova.scheduler.weights.all_weighers'], | 17:19 |
| VW_ | help='Which weight class names to use for weighing hosts'), | 17:19 |
| mgagne | VW_: I agree, my message wasn't that clear =) | 17:19 |
| mgagne | VW_: I'm already using that config. The point is that I find it inconvenient that nova.scheduler.weights.all_weighers is the default. It will tries to load all weighers. | 17:21 |
| mgagne | VW_: This same config is used in tests and someone cannot drop in a new weigher without updating a test that counts the number of weighers found. That's inconvenient when you have more than one custom weighers. You have to apply your patch in the right order, each one incrementing the previous value. | 17:21 |
| VW_ | hmm - OK - let me see if I can find more | 17:23 |
| mgagne | VW_: my rant is mainly about the fact ALL weighers are loaded by default (using auto-discovery) unlike filters which loads a bunch of "standard" filters. | 17:23 |
| VW_ | gotcha | 17:24 |
| VW_ | we pretty much just use the offsets to discourage builds | 17:25 |
| mgagne | say it again? | 17:25 |
| VW_ | and now we've actually writted a custom filter to block all builds if a specific offset is set | 17:25 |
| VW_ | yes, sorry - let me explain better | 17:26 |
| VW_ | when you say weights, my mind initially goes to weighting of cells in the DB | 17:26 |
| mgagne | hehe | 17:27 |
| VW_ | cause we use that a lot | 17:27 |
| VW_ | since there is no "disable" for a cell | 17:27 |
| VW_ | so apologies for going all left field on you | 17:27 |
| mgagne | oh right =) | 17:27 |
| mgagne | yes, that feature could come in handy | 17:28 |
| mgagne | I think RS mentioned it a couple of time last summit during cells presentations | 17:28 |
| *** bvandenh has quit IRC | 17:32 | |
| *** derekh has quit IRC | 17:34 | |
| VW_ | yes we did | 17:34 |
| VW_ | I yell at our cores about it all the time | 17:34 |
| VW_ | :) | 17:34 |
| mdorman | devops, man! | 17:34 |
| VW_ | no, I know | 17:34 |
| VW_ | we actually wrote a custom filter for now | 17:35 |
| mgagne | mdorman: I have a love/hate relationship with the concept of devops =) | 17:35 |
| mdorman | heh, yeah | 17:35 |
| VW_ | but I'd like proper disabling in the DB with reason fields, etc - like hosts | 17:35 |
| mdorman | one of those sounds-good-in-theory things | 17:35 |
| mgagne | mdorman: the only article I found that kind of summarize my opinion can be read on puppetlabs: http://puppetlabs.com/blog/salesforce-devops-journey | 17:35 |
| VW_ | there could be others, though - admin only for example | 17:36 |
| mgagne | mdorman: mainly because people I know only wish to do the "good" part of ops, not the bad and boring one =) | 17:37 |
| VW_ | we stopeed calling it devops and only practice it-buzz-word-du-jor now :P | 17:38 |
| mgagne | how about superhero? :D | 17:38 |
| mgagne | with a rockstar logo | 17:38 |
| VW_ | I actually started calling it opsdev jokingly a while back because most of my teams are systems engineers first and foremost. They've been trying to take all the "devops" tactics for deploying apps on infrastructure and turning a fleet of infrastructure nodes into an application itself | 17:39 |
| *** yapeng has joined #openstack-operators | 17:40 | |
| VW_ | but anyway, I totally get where mdorman is coming from on the misunderstanding of what it is and isn't | 17:41 |
| *** matrohon has quit IRC | 17:42 | |
| VW_ | alright gents, I've ignored the day job for too long. back to meeting hell.... | 17:45 |
| mgagne | good luck | 17:45 |
| mgagne | the vote for presentations website is slow as hell :-/ | 17:53 |
| mdorman | +1 | 17:59 |
| mgagne | Pages take forever to load, nobody ain't got time for that | 18:00 |
| *** avozza is now known as zz_avozza | 18:09 | |
| *** harlowja has joined #openstack-operators | 18:10 | |
| *** belmoreira has quit IRC | 18:11 | |
| *** SimonChung1 has quit IRC | 18:17 | |
| mgagne | for those who needs real same host resize: https://github.com/mgagne/nova/commit/bbc92b18f7e5e6105c7608469e080f6c4e8ea5ac | 18:30 |
| *** cpschult has quit IRC | 18:30 | |
| *** Piet has quit IRC | 18:34 | |
| *** yapeng has quit IRC | 18:37 | |
| *** yapeng_ has joined #openstack-operators | 18:37 | |
| *** Piet has joined #openstack-operators | 19:00 | |
| *** trad511 has quit IRC | 19:08 | |
| *** delatte has quit IRC | 19:15 | |
| *** VW_ has quit IRC | 19:22 | |
| *** cpschult has joined #openstack-operators | 19:39 | |
| *** yapeng_ has quit IRC | 19:52 | |
| *** Marga_ has quit IRC | 19:54 | |
| *** VW_ has joined #openstack-operators | 20:12 | |
| *** turnerg has joined #openstack-operators | 20:12 | |
| *** VW_ has quit IRC | 20:18 | |
| *** VW_ has joined #openstack-operators | 20:19 | |
| *** SimonChung has joined #openstack-operators | 20:19 | |
| *** Marga_ has joined #openstack-operators | 20:25 | |
| *** Marga_ has quit IRC | 20:30 | |
| *** VW_ has quit IRC | 20:31 | |
| *** blair has joined #openstack-operators | 20:36 | |
| *** blair has quit IRC | 20:40 | |
| *** Piet has quit IRC | 20:46 | |
| *** mdorman has quit IRC | 20:48 | |
| *** vinsh has joined #openstack-operators | 20:50 | |
| *** trad511 has joined #openstack-operators | 20:51 | |
| *** Marga_ has joined #openstack-operators | 21:19 | |
| *** Marga_ has quit IRC | 21:20 | |
| *** Marga_ has joined #openstack-operators | 21:20 | |
| *** blair has joined #openstack-operators | 21:22 | |
| *** signed8bit has quit IRC | 21:24 | |
| *** signed8bit has joined #openstack-operators | 21:24 | |
| *** Piet has joined #openstack-operators | 21:29 | |
| *** alop has quit IRC | 21:30 | |
| *** Marga_ has quit IRC | 21:30 | |
| *** Marga_ has joined #openstack-operators | 21:31 | |
| *** alop has joined #openstack-operators | 21:32 | |
| *** yapeng has joined #openstack-operators | 21:39 | |
| jlk | Any of you exposing keystone v3 (in the catalog) at all? | 21:46 |
| *** jaypipes has quit IRC | 21:53 | |
| *** trad511 has quit IRC | 21:57 | |
| mgagne | not yet ^^' | 22:01 |
| *** Marga_ has quit IRC | 22:01 | |
| mgagne | working on it | 22:01 |
| mgagne | have to review the new policy.json =) | 22:02 |
| *** trad511 has joined #openstack-operators | 22:10 | |
| *** trad511 has left #openstack-operators | 22:10 | |
| *** blair has quit IRC | 22:11 | |
| *** SimonChung has quit IRC | 22:11 | |
| *** SimonChung1 has joined #openstack-operators | 22:12 | |
| *** SimonChung has joined #openstack-operators | 22:12 | |
| *** SimonChung1 has quit IRC | 22:12 | |
| *** mdorman has joined #openstack-operators | 22:15 | |
| *** trad512 has joined #openstack-operators | 22:18 | |
| jlk | I'm having a hell of a time with it | 22:21 |
| *** turnerg has quit IRC | 22:24 | |
| *** signed8bit has quit IRC | 22:26 | |
| klindgren | we appear to have the v3 keystone endpoint | 22:27 |
| jlk | trying to figure out how to actually expose it | 22:27 |
| klindgren | I think we had to do it for some custom stuff | 22:27 |
| jlk | I mean, the /v3 URI is there, but isn't listed in the service catalog | 22:27 |
| klindgren | around projects | 22:27 |
| jlk | trying to put it in there... things go weird | 22:27 |
| jlk | most clients and such are setup to expect a auth end point of /v2.0/ | 22:27 |
| *** signed8bit has joined #openstack-operators | 22:28 | |
| jlk | some say to put a unversoined URI in to the service catalog | 22:28 |
| mdorman | klindgren: i htink we only use it for our project api stuff so that we can assign roles to groups as well as users. iirc v2 you can only assign roles to users, not groups | 22:28 |
| jlk | but then things are even weirder. | 22:28 |
| klindgren | yea I think so as well | 22:28 |
| klindgren | however we have both the v2 and v3 exposed in the serivce catalog (endpoint0list) | 22:29 |
| jlk | http://docs.openstack.org/developer/keystone/http-api.html#how-do-i-migrate-from-v2-0-to-v3 | 22:29 |
| jlk | how do you distinctly name them? | 22:29 |
| klindgren | we also have a keystone and keystonev3 service created in the service-list | 22:29 |
| jlk | for things like cinder, clients explicitly look for volumes or volumesv2 | 22:30 |
| jlk | as a endpoint type | 22:30 |
| jlk | I am unsuccessful in getting openstack client or keystoneclient to work with unversioned endpoint | 22:31 |
| jlk | but maybe they mean with the code, not with the actual command line client | 22:32 |
| klindgren | jlk here is what we have: | 22:33 |
| klindgren | keystone endpoint-list | 22:34 |
| klindgren | | <redact> | zone1 | https://openstack-dev.int.godaddy.com:5000/v3 | https://openstack-dev.int.godaddy.com:5000/v3 | https://openstack-dev.int.godaddy.com:35357/v3 | <keyv3> | | 22:34 |
| klindgren | | <redact> | zone1 | https://openstack-dev.int.godaddy.com:5000/v2.0 | https://openstack-dev.int.godaddy.com:5000/v2.0 | https://openstack-dev.int.godaddy.com:35357/v2.0 | <keyv2> | | 22:34 |
| klindgren | keystone service-list | 22:34 |
| klindgren | | keyv2 | keystone | identity | OpenStack Identity Service | | 22:34 |
| klindgren | | keyv3 | keystonev3 | identityv3 | OpenStack Identity Service v3 | | 22:34 |
| klindgren | but like mike said - I believe we are using v2 for pretty much everything | 22:35 |
| jlk | yeah, okay. | 22:35 |
| klindgren | and only using v3 for a very specific function | 22:35 |
| jlk | oooh, using code directly I get discovery to work right | 22:35 |
| jlk | >>> keystone = client.Client(auth_url='http://localhost:5000') | 22:35 |
| jlk | >>> type(keystone) | 22:35 |
| jlk | <class 'keystoneclient.v3.client.Client'> | 22:35 |
| jlk | I just don't think the cli interface is setup to utilize that | 22:36 |
| klindgren | also via the cli stuff - we specifically configure auth url to b the v2.0 | 22:37 |
| *** trad512 has quit IRC | 22:37 | |
| *** medberry2 is now known as med_ | 22:37 | |
| jlk | how do you make use of the v3 stuff then? | 22:37 |
| *** med_ has quit IRC | 22:37 | |
| *** med_ has joined #openstack-operators | 22:37 | |
| jlk | what code out there specifically looks for "identityv3" ? | 22:37 |
| klindgren | 1 very specific project | 22:37 |
| klindgren | hehehe | 22:37 |
| klindgren | I hsould say we have keystonev3 enabled - but its only used by a single internal project | 22:38 |
| *** signed8bit has quit IRC | 22:38 | |
| klindgren | so - sorry for the wild goose chase - we aren't doing what I thought you wanted :-/ | 22:38 |
| jlk | you sort of are | 22:39 |
| jlk | I need v3 end point to do some federatoin stuff, but I just don't get how to make this work | 22:39 |
| mgagne | when I tested from CLI, I used the versionized URL and OS_IDENTITY_API_VERSION=3 | 22:41 |
| mgagne | keystonev3 and identityv3 in the catalog | 22:41 |
| *** VW_ has joined #openstack-operators | 22:46 | |
| *** vinsh has quit IRC | 22:48 | |
| *** matrohon has joined #openstack-operators | 22:56 | |
| *** yapeng has quit IRC | 23:08 | |
| *** blair has joined #openstack-operators | 23:15 | |
| *** pboros_ has quit IRC | 23:24 | |
| *** Marga_ has joined #openstack-operators | 23:30 | |
| *** vinsh has joined #openstack-operators | 23:32 | |
| *** vinsh has quit IRC | 23:32 | |
| *** SimonChung has quit IRC | 23:46 | |
| *** SimonChung1 has joined #openstack-operators | 23:46 | |
| *** csoukup has quit IRC | 23:47 | |
| *** markvoelker has quit IRC | 23:49 | |
| *** matrohon has quit IRC | 23:56 | |
| *** vinsh has joined #openstack-operators | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!