| *** alexpilo_ has quit IRC | 00:01 | |
| *** alexpilotti has joined #openstack-ironic | 00:01 | |
| openstackgerrit | Merged openstack/bifrost: Updated from global requirements https://review.openstack.org/313712 | 00:03 |
|---|---|---|
| *** jvgrant_ has joined #openstack-ironic | 00:05 | |
| *** jvgrant has quit IRC | 00:07 | |
| *** sacharya has quit IRC | 00:09 | |
| *** Sukhdev has joined #openstack-ironic | 00:16 | |
| harlowja | dtantsur where is the code in ironic that converts the old periodic format to the new (i'm gonna see about what a layer in oslo.service that uses the older stuff would look like) | 00:19 |
| *** ayoung has joined #openstack-ironic | 00:20 | |
| *** rcernin has quit IRC | 00:24 | |
| openstackgerrit | Merged openstack/ironic: devstack: Fetch tarball images via https https://review.openstack.org/315663 | 00:27 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/315057 | 00:27 |
| *** catintheroof has quit IRC | 00:27 | |
| *** jvgrant__ has joined #openstack-ironic | 00:32 | |
| *** jvgrant_ has quit IRC | 00:34 | |
| *** mbound has quit IRC | 00:44 | |
| *** vishwanathj has quit IRC | 00:45 | |
| *** YiboCai has joined #openstack-ironic | 00:49 | |
| *** shu-mutou has joined #openstack-ironic | 00:54 | |
| *** hoangcx has joined #openstack-ironic | 00:55 | |
| *** shu-mutou has left #openstack-ironic | 00:55 | |
| *** openstackgerrit has quit IRC | 01:02 | |
| *** openstackgerrit has joined #openstack-ironic | 01:03 | |
| *** mtanino has quit IRC | 01:05 | |
| *** rbudden has quit IRC | 01:07 | |
| *** sacharya has joined #openstack-ironic | 01:10 | |
| *** chlong has joined #openstack-ironic | 01:13 | |
| *** sacharya has quit IRC | 01:15 | |
| *** vishwanathj has joined #openstack-ironic | 01:18 | |
| *** alex_xu_ has joined #openstack-ironic | 01:30 | |
| *** chlong has quit IRC | 01:31 | |
| *** ijw has quit IRC | 01:33 | |
| *** M4g1c5t0rM has joined #openstack-ironic | 01:40 | |
| *** mbound has joined #openstack-ironic | 01:45 | |
| *** rama_y has joined #openstack-ironic | 01:45 | |
| *** mbound has quit IRC | 01:49 | |
| *** alexpilotti has quit IRC | 02:01 | |
| *** alexpilotti has joined #openstack-ironic | 02:03 | |
| openstackgerrit | watanabe isao proposed openstack/ironic: Update neutron services' phrase https://review.openstack.org/315837 | 02:03 |
| *** mbound has joined #openstack-ironic | 02:46 | |
| *** mbound has quit IRC | 02:50 | |
| *** sivaramakrishna has joined #openstack-ironic | 02:51 | |
| *** mtanino has joined #openstack-ironic | 02:53 | |
| *** thrash is now known as thrash|g0ne | 02:53 | |
| *** harlowja_at_home has joined #openstack-ironic | 02:56 | |
| *** absubram has joined #openstack-ironic | 02:57 | |
| *** absubram_ has joined #openstack-ironic | 02:59 | |
| *** absubram has quit IRC | 03:02 | |
| *** absubram_ is now known as absubram | 03:02 | |
| *** chlong has joined #openstack-ironic | 03:08 | |
| *** afaranha has quit IRC | 03:08 | |
| *** sacharya has joined #openstack-ironic | 03:12 | |
| *** sacharya has quit IRC | 03:16 | |
| *** rama_y has quit IRC | 03:26 | |
| *** rama_y has joined #openstack-ironic | 03:26 | |
| *** sacharya has joined #openstack-ironic | 03:28 | |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Add portgroups to support LAG interfaces - API https://review.openstack.org/206244 | 03:38 |
| *** absubram has quit IRC | 03:39 | |
| *** gcb has joined #openstack-ironic | 03:41 | |
| *** szhsong has joined #openstack-ironic | 03:43 | |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack/bifrost: Config drive spec uses network_data.json https://review.openstack.org/315862 | 03:49 |
| *** rama_y has quit IRC | 03:58 | |
| *** rama_y has joined #openstack-ironic | 04:07 | |
| *** ayoung has quit IRC | 04:07 | |
| *** vmud213 has joined #openstack-ironic | 04:07 | |
| *** vmud213 has quit IRC | 04:15 | |
| *** vmud213 has joined #openstack-ironic | 04:16 | |
| *** mtanino has quit IRC | 04:19 | |
| *** bodepd has quit IRC | 04:21 | |
| *** links has joined #openstack-ironic | 04:23 | |
| *** gcb has quit IRC | 04:39 | |
| lazy_prince | Hi.. Can someone from core +A https://review.openstack.org/#/c/315623/ | 04:45 |
| *** vishwanathj has quit IRC | 04:45 | |
| *** rama_y has quit IRC | 04:48 | |
| zer0c00l | pip 8.1.2 crap! | 04:50 |
| zer0c00l | is it fixed yet? | 04:51 |
| *** gcb has joined #openstack-ironic | 04:52 | |
| *** joprovost has quit IRC | 04:56 | |
| *** gcb has quit IRC | 04:58 | |
| *** alexpilotti has quit IRC | 05:00 | |
| *** alexpilotti has joined #openstack-ironic | 05:01 | |
| *** fragatina has quit IRC | 05:10 | |
| *** ChubYann has quit IRC | 05:17 | |
| jlvillal | zer0c00l, What about pip 8.1.2? | 05:18 |
| *** hamzy has joined #openstack-ironic | 05:19 | |
| *** gcb has joined #openstack-ironic | 05:20 | |
| *** amotoki has quit IRC | 05:23 | |
| *** chlong has quit IRC | 05:29 | |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Add network drivers https://review.openstack.org/285852 | 05:30 |
| *** harlowja_at_home has quit IRC | 05:34 | |
| *** rcernin has joined #openstack-ironic | 05:50 | |
| *** vmud213 has quit IRC | 05:50 | |
| *** vmud213 has joined #openstack-ironic | 05:51 | |
| *** amotoki has joined #openstack-ironic | 05:52 | |
| *** fragatina has joined #openstack-ironic | 05:54 | |
| *** szhsong has quit IRC | 05:56 | |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Add network drivers https://review.openstack.org/285852 | 05:56 |
| *** fragatina has quit IRC | 05:58 | |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Update the deploy drivers with network flipping logic https://review.openstack.org/213262 | 06:03 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Added operator documentation for ironic portgroups https://review.openstack.org/228496 | 06:05 |
| *** fragatin_ has joined #openstack-ironic | 06:11 | |
| *** fragatin_ has quit IRC | 06:11 | |
| *** fragatin_ has joined #openstack-ironic | 06:12 | |
| *** mjura has joined #openstack-ironic | 06:14 | |
| *** e0ne has joined #openstack-ironic | 06:15 | |
| *** wajdi has quit IRC | 06:15 | |
| *** fragatin_ has quit IRC | 06:16 | |
| *** mjura has quit IRC | 06:20 | |
| *** mjura has joined #openstack-ironic | 06:21 | |
| *** parlos has joined #openstack-ironic | 06:27 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: TEST Checking Gates https://review.openstack.org/311109 | 06:27 |
| *** fragatina has joined #openstack-ironic | 06:32 | |
| *** fragatina has quit IRC | 06:33 | |
| *** fragatina has joined #openstack-ironic | 06:33 | |
| *** wajdi has joined #openstack-ironic | 06:33 | |
| parlos | Good Morning | 06:45 |
| *** parlos has quit IRC | 06:45 | |
| *** amotoki has quit IRC | 06:51 | |
| *** sacharya has quit IRC | 06:54 | |
| *** Sukhdev has quit IRC | 06:55 | |
| *** tesseract has joined #openstack-ironic | 06:56 | |
| *** amotoki has joined #openstack-ironic | 07:01 | |
| *** amotoki has quit IRC | 07:01 | |
| *** amotoki has joined #openstack-ironic | 07:07 | |
| *** amotoki has quit IRC | 07:07 | |
| openstackgerrit | vinay kumar muddu proposed openstack/ironic-python-agent: Wait for all interfaces to be up before node lookup https://review.openstack.org/303503 | 07:16 |
| *** amotoki has joined #openstack-ironic | 07:17 | |
| *** athomas has joined #openstack-ironic | 07:23 | |
| *** hoangcx has quit IRC | 07:25 | |
| *** jcoufal has joined #openstack-ironic | 07:44 | |
| *** szhsong has joined #openstack-ironic | 07:53 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** zzzeek has joined #openstack-ironic | 08:02 | |
| *** ohamada has joined #openstack-ironic | 08:06 | |
| *** ohamada has quit IRC | 08:08 | |
| *** ohamada has joined #openstack-ironic | 08:09 | |
| *** kromanenko_ has joined #openstack-ironic | 08:19 | |
| *** lucas-dinner is now known as lucasagomes | 08:31 | |
| lucasagomes | good morning all | 08:32 |
| *** vmud213 has quit IRC | 08:32 | |
| *** vmud213 has joined #openstack-ironic | 08:32 | |
| *** lintan_ has quit IRC | 08:33 | |
| *** alex_xu_ has quit IRC | 08:33 | |
| *** lintan_ has joined #openstack-ironic | 08:34 | |
| *** alex_xu has joined #openstack-ironic | 08:37 | |
| aarefiev | morning all! | 08:39 |
| vdrok | morning ironic, lucasagomes aarefiev | 08:39 |
| aarefiev | morning vdrok | 08:40 |
| *** parlos has joined #openstack-ironic | 08:40 | |
| *** jtomasek_ has joined #openstack-ironic | 08:46 | |
| pas-ha | morning aarefiev vdrok lucasagomes | 08:46 |
| *** jistr has joined #openstack-ironic | 08:46 | |
| vsaienko | morning Ironic'ers | 08:48 |
| *** jtomasek_ has quit IRC | 08:51 | |
| *** e0ne has quit IRC | 08:52 | |
| *** sacharya has joined #openstack-ironic | 08:54 | |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Add network drivers https://review.openstack.org/285852 | 08:55 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Update the deploy drivers with network flipping logic https://review.openstack.org/213262 | 08:55 |
| sivaramakrishna | morning all! | 08:57 |
| *** sacharya has quit IRC | 08:59 | |
| *** electrofelix has joined #openstack-ironic | 09:07 | |
| *** dmk0202 has joined #openstack-ironic | 09:09 | |
| *** oberezovskyi has quit IRC | 09:10 | |
| jroll | welp, it's friday alright | 09:11 |
| jroll | http://logs.openstack.org/35/313035/4/check/gate-tempest-dsvm-ironic-agent_ssh/b98e41d/logs/ironic-bm-logs/node-0_console_12-05-2016-22:24:40.txt.gz | 09:11 |
| jroll | http://logs.openstack.org/35/313035/4/check/gate-tempest-dsvm-ironic-agent_ssh/b98e41d/logs/screen-ir-cond.txt.gz?level=WARNING | 09:11 |
| jroll | :| | 09:11 |
| jroll | sambetts: something's suddenly up with tinyipa | 09:13 |
| jroll | I think just whole disk images? it's unclear to me | 09:13 |
| sambetts | jroll: ?! | 09:15 |
| jroll | right? | 09:15 |
| jroll | so this does bring up a point about tinyipa | 09:15 |
| jroll | if this was ubuntu, I'd be like "oh apt-get install lib-whatever" | 09:15 |
| jroll | nfi what to do here, but then again haven't really looked around | 09:16 |
| sambetts | jroll: you just need to add it to build_files/finalreqs.lst if it exists in the tiny core repo | 09:16 |
| sambetts | jroll: which I'm pretty sure this does, however it should actually be being pulled in by gdisk as a requirement... | 09:17 |
| jroll | ya | 09:17 |
| sambetts | huh... well... ok, so when they updated gdisk for us to remove the requirement on icu they also removed the whole dep file, because unless you use sgdisk its an optional requirement | 09:18 |
| jroll | $ curl -s distro.ibiblio.org/tinycorelinux/7.x/x86/tcz/gdisk.tcz.dep | grep pop 1 ↵ | 09:18 |
| jroll | popt.tcz | 09:18 |
| jroll | hm | 09:18 |
| sambetts | jroll: thats the x86 deps | 09:19 |
| sambetts | you need x86_64 | 09:19 |
| jroll | so we need to add popt-dev? | 09:19 |
| jroll | ah | 09:19 |
| sambetts | gdisk.tcz.dep doesn't exist in x86_64 :/ | 09:19 |
| jroll | yep :| | 09:19 |
| sambetts | not sure why it wasn't caught by the gate when we merged the patch... | 09:20 |
| jroll | yeah, that's interesting | 09:21 |
| jroll | unless they removed it between merge and post job that built it | 09:21 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-python-agent: tinyipa: add popt to deps https://review.openstack.org/315982 | 09:22 |
| jroll | that should do it | 09:22 |
| sambetts | I guess because we weren't testing agent deploy with tinyIPA? | 09:22 |
| sambetts | jroll: ha you beat me to it! | 09:22 |
| *** vmud213 has quit IRC | 09:22 | |
| jroll | that could be too? | 09:22 |
| jroll | I think this is just agent driver whole disk images | 09:22 |
| sambetts | jroll: did JayF's patch merge switching to tinyIPA for all things and thats why we now see it | 09:23 |
| jroll | so you might be right | 09:23 |
| lucasagomes | morning sambetts jroll | 09:23 |
| jroll | but post queues are long lately | 09:23 |
| jroll | um | 09:23 |
| jroll | oh no, that's where I saw the break https://review.openstack.org/#/c/313035/ | 09:23 |
| jroll | so maybe rest of gate is fine til that lands :) | 09:23 |
| sambetts | ooooh yeah post queue might have delayed the build | 09:23 |
| sambetts | so the change to 7.x wasn't picked up until that was done | 09:24 |
| *** vmud213 has joined #openstack-ironic | 09:24 | |
| jroll | http://status.openstack.org/openstack-health/#/?searchProject=ironic&groupKey=project&resolutionKey=hour&end=2016-05-13T09:24:08.130Z | 09:24 |
| jroll | doesn't look like we're totally broken | 09:24 |
| openstackgerrit | Sivaramakrishna Garimella proposed openstack/ironic: Added operator documentation for ironic portgroups https://review.openstack.org/228496 | 09:25 |
| jroll | oh, well maybe http://status.openstack.org/openstack-health/#/?searchProject=ironic&groupKey=project&resolutionKey=hour&end=2016-05-13T09:24:08.130Z&duration=P1D | 09:25 |
| jroll | there http://status.openstack.org/openstack-health/#/g/project/openstack~2Fironic?groupKey=project&resolutionKey=hour&duration=P1D&end=2016-05-13T09:24:08.130Z | 09:25 |
| sambetts | :/ yeah we should get that landed ASAP while the post queue is short | 09:25 |
| jroll | not enough load to tell | 09:25 |
| jroll | yep, lucasagomes mind hitting 315982? | 09:26 |
| sambetts | hey lucasagomes | 09:26 |
| jroll | and morning lucasagomes :) | 09:26 |
| * lucasagomes looks | 09:26 | |
| lucasagomes | jroll, done, +2+a | 09:27 |
| jroll | thanks man | 09:27 |
| sambetts | your up super early aren't you? | 09:27 |
| jroll | so TIL openstack-health is suddenly awesome | 09:27 |
| jroll | also, we're failing oslo master | 09:27 |
| jroll | https://github.com/openstack/oslo.log/commit/7c671ea77f4d7701d0def9a9e17b6c49310ca438 | 09:27 |
| jroll | HA | 09:27 |
| lucasagomes | yeah I didn't know about it | 09:28 |
| jroll | sambetts: ish, couldn't sleep | 09:28 |
| sambetts | :( | 09:28 |
| openstackgerrit | Merged openstack/python-ironic-inspector-client: Bump supported API version to 1.5 https://review.openstack.org/315047 | 09:28 |
| * lucasagomes likes the failopotamus project for specific jobs failures as well | 09:28 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Make iPXE + TinyIPA the defaults for devstack https://review.openstack.org/313035 | 09:28 |
| jroll | oh ugh, depends-on won't help there | 09:29 |
| jroll | whoops >.> | 09:29 |
| sambetts | lucasagomes: failopotamus? | 09:29 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Make iPXE + TinyIPA the defaults for devstack https://review.openstack.org/313035 | 09:29 |
| lucasagomes | sambetts, yeah, it's a tiny script https://github.com/sdague/failopotamus | 09:29 |
| sambetts | jroll: because its not a src job right? | 09:30 |
| lucasagomes | it greats a graph with the failures of a specific job in gate | 09:30 |
| * lucasagomes internally it just calls graphite | 09:30 | |
| lucasagomes | but yeah, it's nice | 09:30 |
| sambetts | ha thats cool! | 09:30 |
| jroll | sambetts: yep | 09:31 |
| jroll | oh wow, gerrit recognized I put it back correctly and reapplied the votes | 09:31 |
| jroll | O_o | 09:31 |
| sambetts | woah?! I didn't know that was a thing! | 09:32 |
| *** e0ne has joined #openstack-ironic | 09:33 | |
| jroll | right? | 09:33 |
| *** zhihaisong has joined #openstack-ironic | 09:37 | |
| lucasagomes | aarefiev, the nova bug seems tricky, I'm not sure yet, but it seems that since we create all the nodes with the same amount of cpus, ram and disk they all have the same resource_class_id in the inventories table | 09:38 |
| lucasagomes | so once one is changed to 0, all the rest is updated | 09:38 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Remove verbose option https://review.openstack.org/315990 | 09:38 |
| jroll | ^ this should unbreak us on oslo master | 09:38 |
| sambetts | wow... our tests actual test a URL we don't support http://logs.openstack.org/66/315766/2/check/gate-tempest-dsvm-ironic-pxe_ssh/695c53b/console.html | 09:39 |
| sambetts | actually* | 09:39 |
| sambetts | according to our API ref anyway... | 09:39 |
| *** jcoufal has quit IRC | 09:39 | |
| jroll | sambetts: O_o | 09:40 |
| jroll | that feels like a bug / another broken gate thing | 09:40 |
| *** szhsong has quit IRC | 09:40 | |
| sambetts | thats on my patch https://review.openstack.org/#/c/315766/ | 09:40 |
| jroll | like, clearly something happened there, we didn't magically get a new broken test | 09:40 |
| jroll | sure you didn't break that? :) | 09:40 |
| jroll | I don't quite trust our api ref yet :/ | 09:41 |
| sambetts | I broke it if the test is trying to do /v1/nodes/states/console, because the URL we define and I made it support in my patch is v1/nodes/<node_id>/states/console | 09:42 |
| jroll | right, apparently we test that | 09:42 |
| * jroll wonders what that does | 09:42 | |
| *** zhihais has joined #openstack-ironic | 09:42 | |
| sambetts | woah?! no please no ... | 09:44 |
| sambetts | we're actually testing /v1/nodes/states/console/<node_id> | 09:44 |
| jroll | hahahahahaha | 09:44 |
| jroll | T_T | 09:45 |
| sambetts | ewwwwwwwwwwwwwwwwwwwwwwwwwwwww!!!!! | 09:45 |
| jroll | (that does totally work) | 09:45 |
| *** zhihaisong has quit IRC | 09:45 | |
| jroll | /v1/nodes/states/console gives "error_message": "{\"debuginfo\": null, \"faultcode\": \"Client\", \"faultstring\": \"Missing argument: \\\"node_ident\\\"\"}" | 09:45 |
| * sambetts claws his eyes out | 09:45 | |
| * sambetts also thinks that is a bug | 09:45 | |
| *** hoangcx has joined #openstack-ironic | 09:45 | |
| aarefiev | lucasagomes: hey, I have found the root cause, it's issue in db code while updating inventories | 09:46 |
| jroll | fwiw v1/nodes/states/power/uuid doesn't work | 09:46 |
| lucasagomes | aarefiev, yeah that's what I was heading to | 09:46 |
| lucasagomes | aarefiev, thanks for that | 09:46 |
| lucasagomes | aarefiev, do you have a patch for it already? | 09:46 |
| aarefiev | lucasagomes: I'm working on it now, it's one line fix | 09:47 |
| sambetts | jroll: .... oh man, these controllers really need somework... | 09:48 |
| jroll | sambetts: curl -v -X PUT http://localhost:6385/v1/nodes/maintenance/45a6b4e2-fb9f-4ab6-8760-74e231d441b3 -> 202 | 09:48 |
| * jroll cries | 09:48 | |
| *** chopmann has joined #openstack-ironic | 09:48 | |
| sambetts | yeah... that should be fixed by my patch too | 09:49 |
| jroll | this is a really interesting case of breaking an api | 09:49 |
| jroll | technically we shouldn't do that, but... | 09:49 |
| sambetts | I say if we never supported it in the first place its a bug | 09:49 |
| lucasagomes | aarefiev, cool | 09:50 |
| jroll | well, how do you define supported | 09:50 |
| sambetts | documented support | 09:50 |
| jroll | well than all of our apis are a bug ;) | 09:50 |
| jroll | I do get what you mean | 09:50 |
| sambetts | heh, http://docs.openstack.org/developer/ironic/webapi/v1.html#get--v1-nodes-(node_ident)-states-console <- supported | 09:50 |
| lucasagomes | yeah, it's too many magic with controllers | 09:51 |
| lucasagomes | parameters can be inputed as part of the URL or body request | 09:51 |
| jroll | yeah, it's a bit of a dumpster fire | 09:51 |
| sambetts | lucasagomes: having worked with pecan and RoR a lot in the past, we've used and abused the controllers really badly | 09:52 |
| lucasagomes | sambetts, right, the controllers way AFAIUI is the wsme way of doing it right? | 09:52 |
| jroll | nah, controllers are a well-understood thing in most frameworks | 09:52 |
| jroll | model-view-controller | 09:53 |
| sambetts | not as far as know, the controllers come from pecan, wsme just fudges some validation stuff in there | 09:53 |
| *** chopmann has quit IRC | 09:53 | |
| lucasagomes | gotcha | 09:53 |
| lucasagomes | yeah apparently we did it wrong from the start then :-/ | 09:53 |
| jroll | model == internal (object) representation, view == transformation to the API representation, controller == api logic | 09:54 |
| jroll | imo the magic routing and the way pecan exposes parameters, as well as unfamiliarity with the things, is what made us shoot ourselver in the foot | 09:54 |
| jroll | ourselves* | 09:54 |
| lucasagomes | yeah | 09:55 |
| lucasagomes | yeah, it's funny that this problem may affect other OS projects using wsme | 09:55 |
| *** sacharya has joined #openstack-ironic | 09:55 | |
| lucasagomes | cause we looked at them to make our api | 09:56 |
| jroll | I kind of want to rewrite it in $other_framework, but we don't have great api testing so we'll almost certainly break something | 09:56 |
| sambetts | the thing that breaks the pecan RestController magic is that sub-resource functions are expecting to receive idents from all its parents, e.g. def get_one(node_id, state_id): ... but we've written weird functions that only receive one parameter in e.g. def raid(node_id): and so when it tries to do its matching its like oh I've got enough info to call that function so I must be able too | 09:57 |
| *** amotoki has quit IRC | 09:58 | |
| lucasagomes | jroll, yeah agreed, loads of corner cases there | 09:58 |
| jroll | I so want to get the network stuff done, make our testign amazing, and then just do tech debt cleanup for a cycle | 09:59 |
| jroll | if I was a dictator, I think that's what I would do | 10:00 |
| *** sacharya has quit IRC | 10:00 | |
| lucasagomes | jroll, heh don't need to be one, I mean, it's arguable | 10:00 |
| lucasagomes | and in fairness we do need to do it yes | 10:01 |
| jroll | lucasagomes: yeah, just saying not everyone would agree and I can't force it | 10:01 |
| lucasagomes | yeah... there's a bunch of spaghetti code (/me thinks about the pxe config generation bits) | 10:02 |
| openstackgerrit | Aparna proposed openstack/ironic: Modify doc for RAID clean steps in manual cleaning https://review.openstack.org/315998 | 10:02 |
| sambetts | jroll: wow... just wow... https://github.com/openstack/ironic/blob/master/ironic_tempest_plugin/services/baremetal/v1/json/baremetal_client.py#L331 does it one way, then the function below it does it the right way ... | 10:06 |
| *** gcb has quit IRC | 10:06 | |
| jroll | sambetts: yeah, so we intend for /v1/nodes/states/console?node_id=foo to work | 10:06 |
| jroll | (apparently) | 10:06 |
| sambetts | do we? | 10:07 |
| *** e0ne has quit IRC | 10:07 | |
| jroll | well | 10:08 |
| jroll | if we have a test for it... | 10:09 |
| sambetts | that call creates a URL that looks like /v1/nodes/states/console/<node_id> | 10:10 |
| *** YiboCai has quit IRC | 10:11 | |
| sambetts | also I'm not sure if that file is actually a test is it? its just used by the tests? and its the only one of its kind looking through the fiel | 10:11 |
| jroll | aha, I see | 10:12 |
| jroll | hrm | 10:12 |
| sambetts | is the OrphanedObjectError thing fixed yet? I still see it in the n-cpu logs for this patches tests | 10:14 |
| sambetts | python-ironiclient does it /v1/nodes/<node id>/states/console | 10:15 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Fix all possible node subresources routing collisions https://review.openstack.org/315766 | 10:15 |
| *** amotoki has joined #openstack-ironic | 10:18 | |
| jroll | sambetts: should be as of like... 2200 utc? | 10:19 |
| *** hoangcx has quit IRC | 10:19 | |
| sambetts | jroll: hmmm that test run posted results at 22:29 UTC so it might have been the last one to run without the fix | 10:20 |
| *** d0ugal has quit IRC | 10:20 | |
| *** d0ugal has joined #openstack-ironic | 10:21 | |
| jroll | sambetts: that time is totally an estimate fwiw :) | 10:22 |
| jroll | 21:46:41 JayF | yep just landed | 10:22 |
| jroll | to be exact | 10:22 |
| sambetts | :-P the new run I've just pushed should pass then, cool | 10:23 |
| *** kromanenko__ has joined #openstack-ironic | 10:26 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Test post don't upvote https://review.openstack.org/311865 | 10:29 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: TEST Checking Gates https://review.openstack.org/311109 | 10:30 |
| *** jtomasek_ has joined #openstack-ironic | 10:33 | |
| jroll | very interesting http://lists.openstack.org/pipermail/openstack-dev/2016-May/094867.html | 10:34 |
| *** amotoki has quit IRC | 10:36 | |
| jroll | bbiab | 10:36 |
| *** jtomasek_ has quit IRC | 10:42 | |
| *** amotoki has joined #openstack-ironic | 10:43 | |
| *** amotoki has quit IRC | 10:44 | |
| openstackgerrit | Merged openstack/ironic-python-agent: tinyipa: add popt to deps https://review.openstack.org/315982 | 10:52 |
| sambetts | woooop!! ^ | 10:52 |
| lucasagomes | sambetts, w00t | 10:57 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/virtualbmc: Add unittests for the config.py module https://review.openstack.org/316021 | 10:57 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/virtualbmc: Set IPMI session timeout for the virtual BMCs https://review.openstack.org/316022 | 10:57 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/virtualbmc: Set IPMI session timeout for the virtual BMCs https://review.openstack.org/316022 | 11:00 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/virtualbmc: Bump pyghmi version to 1.0.3 https://review.openstack.org/316023 | 11:00 |
| *** alexpilotti has quit IRC | 11:00 | |
| *** mkovacik__ has joined #openstack-ironic | 11:02 | |
| openstackgerrit | Yuriy Zveryanskyy proposed openstack/ironic: [WIP] [POC] Callback from ramdisk support for Ansible driver https://review.openstack.org/310816 | 11:02 |
| *** mkovacik__ is now known as _milan_ | 11:03 | |
| *** devananda has quit IRC | 11:03 | |
| *** baoli has joined #openstack-ironic | 11:04 | |
| sivaramakrishna | vsaienko: can you rebase the hwinfo patch https://review.openstack.org/#/c/256364/61 ? | 11:04 |
| *** sivaramakrishna has quit IRC | 11:05 | |
| *** alexpilotti has joined #openstack-ironic | 11:07 | |
| *** baoli has quit IRC | 11:09 | |
| *** mgould has joined #openstack-ironic | 11:14 | |
| *** devananda has joined #openstack-ironic | 11:14 | |
| openstackgerrit | vinay kumar muddu proposed openstack/ironic-python-agent: Fix local boot issue with fedora in uefi mode https://review.openstack.org/302143 | 11:14 |
| mgould | morning Ironic | 11:14 |
| *** blinky_ghost has joined #openstack-ironic | 11:17 | |
| lucasagomes | mgould, g'morning | 11:17 |
| sambetts | jroll: right now we make DELETE v1/chassis/<chassis id>/nodes/<node id> an operation not permitted, I assume this was done because right now the routing would make v1/chassis/<chassis id>/nodes pass the chassis id into the delete function, with my changes the right IDs are passed into the right places, e.g. if you do DELETE v1/chassis/<chassis id>/nodes it complains you havn't pass in a node | 11:18 |
| sambetts | id, which is the correct behavior, can we make that action valid now? | 11:18 |
| sambetts | lucasagomes: ^^ | 11:18 |
| mgould | morning lucasagomes sambetts jrist | 11:19 |
| mgould | *jroll | 11:19 |
| sambetts | hey mgould | 11:19 |
| lucasagomes | sambetts, yeah the old api has a bunch of from_<resource> preventing certain actions | 11:19 |
| *** thrash|g0ne is now known as thrash | 11:20 | |
| lucasagomes | due the way it was programmed (aka wrongly) :-/ | 11:20 |
| blinky_ghost | hi all, can anybody tell me if ironic supports vmware esxi based vms? I'm trying to test tripleo on my lab u using pxe_ssh driver and virt_type=vmware but I have this errors when deploying the overcloud: File "/usr/lib/python2.7/site-packages/ironic/drivers/modules/ssh.py", line 275, in _set_boot_device ironic-conductor[1216]: raise NotImplementedError() ironic-conductor[1216]: NotImplementedError | 11:20 |
| sambetts | lucasagomes: yeah, with my changes that we can remove a lot of them :/ | 11:20 |
| lucasagomes | now problem is that, by changing this behaviors we also change the API :-/ not sure if we now should allow things like DELETE /chassis/<uuid>/nodes/<uuid> | 11:20 |
| sambetts | blinky_ghost: I think all supported drivers are here -> http://docs.openstack.org/developer/ironic/deploy/drivers.html | 11:21 |
| lucasagomes | I mean, I wouldn't mind, but that's something that wasn't allowed for older versions and now it is allowed | 11:21 |
| lucasagomes | sounds a bit inconsistent | 11:21 |
| sambetts | lucasagomes: additional things should be fine though right? as long as we're not removing anything that was supported/ | 11:21 |
| lucasagomes | blinky_ghost, I don't think we support vmware esxi | 11:21 |
| sambetts | ? | 11:21 |
| lucasagomes | blinky_ghost, BUT, if you use the libvirt driver form ironic-staging-drivers | 11:22 |
| lucasagomes | or virtualBMC that will work | 11:22 |
| lucasagomes | beucase the use python-libvirt underneath and AFAIK, libvirt works with vmware esxi | 11:22 |
| * lucasagomes checks | 11:22 | |
| lucasagomes | blinky_ghost, yeah, seems so http://libvirt.org/ | 11:22 |
| blinky_ghost | lucasgomes: I'm using pxe_ssh driver, which should I use for libvirt? | 11:23 |
| lucasagomes | sambetts, right, it won't break anyone, but that discussion may fall into the microversion-ish realm | 11:23 |
| *** jaoh has joined #openstack-ironic | 11:24 | |
| lucasagomes | where APIs of the same version (if we don't bump) will be inconsistent across deployments | 11:24 |
| lucasagomes | it's a fine line tho... | 11:24 |
| lucasagomes | but we also can argue that it's was a bug and should be fixed across all versions... so I don't know, I'm good with it but we need to have a consensus I guess | 11:24 |
| sambetts | blinky_ghost: we're transfering over to virtualBMC, which is a translation layer between IPMI and libvirt, so you can test the real pxe_ipmitool driver | 11:24 |
| lucasagomes | blinky_ghost, right, so the libvirt driver is not in the ironic tree | 11:25 |
| lucasagomes | blinky_ghost, you have to do a "pip install ironic-staging-drivers" | 11:25 |
| sambetts | lucasagomes: yeah, its a hard one :/ | 11:25 |
| lucasagomes | blinky_ghost, when then you can import "pxe_libvirt_iscsi" or "pxe_libvirt_agent" depending on ur deploy method | 11:26 |
| aarefiev | lucasagomes: nova fix here https://review.openstack.org/#/c/316031/ | 11:26 |
| lucasagomes | unfortunately we lack documentation for this driver :-/ | 11:26 |
| blinky_ghost | lucasgomes I can use pxe_ipmitool also? for vmware? | 11:26 |
| lucasagomes | blinky_ghost, you after you enroll a node with it, you can then look at "ironic node-get-properties" command | 11:27 |
| lucasagomes | blinky_ghost, yes, if you use VirtualBMC to translate the ipmitool commands to libvirt you can | 11:27 |
| lucasagomes | aarefiev, thanks | 11:27 |
| lucasagomes | sambetts, yeah, it's just politics. But totally arguable I mean | 11:27 |
| lucasagomes | we should build the api in the correctly! | 11:28 |
| sambetts | lucasagomes: yeah, why do I always managed to get into the political parts of OpenStack XD | 11:28 |
| lucasagomes | sambetts, heh there are many | 11:28 |
| lucasagomes | in the correct way* | 11:28 |
| sambetts | lucasagomes: I'll leave the not permitted exceptions in place for now | 11:29 |
| lucasagomes | sambetts, ++ | 11:29 |
| *** dprince has quit IRC | 11:29 | |
| sambetts | lucasagomes: although my changes still end up changing DELETE v1/chassis/<uuid>/nodes from operation not permitted to bad request missing node id, because the expose validate catches it before it even makes it into the function | 11:30 |
| lucasagomes | :-/ yeah some corner cases | 11:32 |
| lucasagomes | I wonder if it worth to send an email ML about it, with the api folks. Or at least bring these corner cases up to the next meeting ? | 11:33 |
| lucasagomes | see what people think about it | 11:33 |
| sambetts | lucasagomes: right now POST v1/chassis/<uuid>/nodes is operation not permitted, and I think that one would be super useful to enable | 11:33 |
| sambetts | lucasagomes: for creating a node belonging to a chassis | 11:33 |
| lucasagomes | yeah to create a node that belongs to a chassis, right now the only way is to pass chassis_uuid in the body request when creating the node directly | 11:34 |
| sambetts | yeah | 11:34 |
| jroll | sambetts: well, I think we decided /v1/chassis/<chassis id>/nodes was duplicate and we didn't want duplicate things | 11:37 |
| blinky_ghost | lucasagomes: to test the virtual bmc, I just run pip install vrtualbmc? | 11:38 |
| jroll | sambetts: lucasagomes: I think the right thing to do here is make a list of bad/inconsistent things, and propose a spec with what to do to fix them | 11:38 |
| sambetts | is it a dup though? because it should autopopulate chassis_id into the node right? | 11:38 |
| jroll | and then we have lots of bikeshedding there | 11:38 |
| jroll | sambetts: I meant with e.g. DELETE | 11:38 |
| lucasagomes | blinky_ghost, yeah, "pip install virtualbmc" that will gives you a "vbmc" command which you can use to register the vms | 11:38 |
| lucasagomes | blinky_ghost, https://github.com/umago/virtualbmc#usage | 11:39 |
| sambetts | jroll: ah, right, well ideally it shouldn't be a dup because it should validation the node your trying to delete belongs to that chassis | 11:39 |
| sambetts | validate* | 11:39 |
| lucasagomes | check vbmc --help as well for additional parameters such as "--port" (which should be unique for each virtual bmc) etc... | 11:39 |
| sambetts | man ... I'm back to Friday typing hands... | 11:39 |
| lucasagomes | jroll, yeah, that works too | 11:39 |
| jroll | sambetts: sure, I guess | 11:39 |
| *** jaoh has quit IRC | 11:41 | |
| *** derekh has joined #openstack-ironic | 11:42 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/virtualbmc: Set IPMI session timeout for the virtual BMCs https://review.openstack.org/316022 | 11:43 |
| * lucasagomes goes grab a quick lunch | 11:45 | |
| *** lucasagomes is now known as lucas-hungry | 11:45 | |
| *** baoli has joined #openstack-ironic | 11:45 | |
| *** baoli has quit IRC | 11:47 | |
| *** baoli has joined #openstack-ironic | 11:47 | |
| *** baoli_ has joined #openstack-ironic | 11:48 | |
| blinky_ghost | lucasagomes: how do I integrate this with ironic? | 11:48 |
| *** baoli has quit IRC | 11:52 | |
| *** cdearborn has joined #openstack-ironic | 11:54 | |
| *** sacharya has joined #openstack-ironic | 11:56 | |
| jroll | JayF: jcook: you'll be interested in https://bugs.launchpad.net/ironic/+bug/1580931 | 11:59 |
| openstack | Launchpad bug 1580931 in Ironic "[RFE] Tooling for recovering from -WAIT and -ING state" [Wishlist,Incomplete] - Assigned to Tan Lin (tan-lin-good) | 11:59 |
| *** sacharya has quit IRC | 12:01 | |
| *** tloeb has joined #openstack-ironic | 12:07 | |
| mat128 | jroll: that'd be great ^ | 12:09 |
| jroll | yeah | 12:10 |
| *** links has quit IRC | 12:14 | |
| *** mbound has joined #openstack-ironic | 12:16 | |
| *** clif_h has quit IRC | 12:17 | |
| *** jroll has quit IRC | 12:17 | |
| zhihais | Hi, guys, i am using liberty, and trying to build a cloud that have both kvm compute nodes and bare metal nodes. But for bare metal, nova scheduler has its own config like 'host_manager'. How should i setup nova-scheduler to support the co-existance of kvm vms and bare metal node? | 12:19 |
| *** aarefiev has quit IRC | 12:20 | |
| *** amotoki has joined #openstack-ironic | 12:20 | |
| *** aarefiev has joined #openstack-ironic | 12:21 | |
| *** raildo-afk is now known as raildo | 12:22 | |
| *** jroll has joined #openstack-ironic | 12:22 | |
| *** jroll has quit IRC | 12:22 | |
| *** jroll has joined #openstack-ironic | 12:22 | |
| *** raildo has left #openstack-ironic | 12:22 | |
| *** joprovost has joined #openstack-ironic | 12:26 | |
| *** joprovost has quit IRC | 12:26 | |
| *** alaski is now known as lascii | 12:29 | |
| *** clif_h has joined #openstack-ironic | 12:30 | |
| *** dprince has joined #openstack-ironic | 12:34 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: WIP: grenade checking https://review.openstack.org/311109 | 12:38 |
| *** fellypefca has joined #openstack-ironic | 12:39 | |
| *** mtanino has joined #openstack-ironic | 12:41 | |
| *** Goneri has joined #openstack-ironic | 12:41 | |
| *** rcernin_ has joined #openstack-ironic | 12:42 | |
| *** rcernin has quit IRC | 12:42 | |
| *** jaypipes is now known as leakypipes | 12:42 | |
| openstackgerrit | Thiago Paiva Brito proposed openstack/python-oneviewclient: OneView client with managers https://review.openstack.org/286803 | 12:46 |
| *** mtanino has quit IRC | 12:49 | |
| *** lucas-hungry is now known as lucasagomes | 12:53 | |
| *** tloeb has quit IRC | 12:54 | |
| *** e0ne has joined #openstack-ironic | 12:57 | |
| *** jaoh has joined #openstack-ironic | 12:57 | |
| lucasagomes | zhihais, hi there, the host manager is fine because it will check for the hypervisor type when doing the sched things | 12:58 |
| lucasagomes | zhihais, but other stuff such as tweaking the {ram,disk,cpu}_allocation_ratio etc will really depend on ur use case | 12:58 |
| lucasagomes | if you ended up seeing that using one n-sch does not fit into ur model, I believe you will have to use nova cells to separate the virtual machines and the baremetal ones | 12:59 |
| lucasagomes | and then you configure the scheduler differently on each cell | 12:59 |
| *** ayoung has joined #openstack-ironic | 13:00 | |
| *** amotoki has quit IRC | 13:00 | |
| *** alexpilotti has quit IRC | 13:02 | |
| *** piet has joined #openstack-ironic | 13:02 | |
| *** chopmann has joined #openstack-ironic | 13:04 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add tests for cleaning https://review.openstack.org/304469 | 13:06 |
| *** alexpilotti has joined #openstack-ironic | 13:06 | |
| *** tloeb has joined #openstack-ironic | 13:09 | |
| *** rbudden has joined #openstack-ironic | 13:10 | |
| *** dprince has quit IRC | 13:11 | |
| *** dprince has joined #openstack-ironic | 13:11 | |
| *** rcernin_ has quit IRC | 13:16 | |
| mgould | lucasagomes, I thought you needed cells to have VM and BM in the same installation? | 13:16 |
| lucasagomes | mgould, I don't think you required it, you can use host aggregates to separate both types and use the same scheduler | 13:17 |
| lucasagomes | I believe it's mostly up to the business logic, say for example, that for baremetal you must use the exact filters | 13:18 |
| vsaienko | morning dtantsur, lucasagomes, jroll | 13:19 |
| jroll | \o vsaienko | 13:20 |
| lucasagomes | mgould, then you will need a dedicated scheduler for the baremetal loads (since the vms will use different filters) | 13:20 |
| vsaienko | jroll once you have a time could you please review https://review.openstack.org/#/c/304473/ | 13:20 |
| vsaienko | the patch adds experimental job to test cleaning at the gates | 13:21 |
| *** odyssey4me has quit IRC | 13:21 | |
| *** odyssey4me has joined #openstack-ironic | 13:21 | |
| NobodyCam | good morning Ironicers | 13:22 |
| jroll | vsaienko: one comment, else lgtm | 13:22 |
| jroll | hey NobodyCam | 13:22 |
| lucasagomes | vsaienko, IRONIC_VM_SPEC_DISK=1 works?! Casue IPA will look for >=4 GB no? | 13:22 |
| lucasagomes | I might be wrong | 13:22 |
| jroll | oh good point | 13:22 |
| NobodyCam | hey hey jroll :) | 13:22 |
| jroll | we could lay down root device hints no? | 13:22 |
| NobodyCam | oh ya TGIF!!! | 13:22 |
| lucasagomes | jroll, yup | 13:23 |
| lucasagomes | checking our devstack/lib/ironic | 13:23 |
| vsaienko | jroll, lucasagomes it has been already fixed https://github.com/openstack/ironic/commit/3295a7e8ec0312b84694a02aaf4f1d6a0228d6e3 | 13:23 |
| mgould | NobodyCam, morning! | 13:23 |
| lucasagomes | vsaienko, a-ha! | 13:24 |
| lucasagomes | ok | 13:24 |
| NobodyCam | good morning mgould :) | 13:24 |
| jroll | vsaienko: oh neat | 13:24 |
| *** amotoki has joined #openstack-ironic | 13:24 | |
| openstackgerrit | Merged openstack/virtualbmc: Add unittests for the config.py module https://review.openstack.org/316021 | 13:24 |
| jroll | vsaienko: yeah just the one comment then | 13:24 |
| NobodyCam | and lucasagomes too | 13:24 |
| persia | NobodyCam: So, this is the kind of Friday that is not usually considered ideal | 13:24 |
| lucasagomes | NobodyCam, hi there | 13:24 |
| vsaienko | jroll fixed | 13:25 |
| jroll | thanks | 13:25 |
| NobodyCam | persia: ??? may I ask why.. (and good morning) | 13:25 |
| *** vmud213 has quit IRC | 13:26 | |
| * lucasagomes remember to update the RFE about root device hints... | 13:26 | |
| * lucasagomes does that | 13:26 | |
| *** rcernin has joined #openstack-ironic | 13:27 | |
| lucasagomes | btw folks, if you have some spare time here's a quick patch https://review.openstack.org/#/c/315593/ | 13:27 |
| lucasagomes | plumbing work for gating the ironic job against the virtualbmc gate | 13:27 |
| persia | NobodyCam: I don't actually know. WIkipedia claims it is a 19th century conflation of the misfortune of the numeric value of the day of the month and the day of the week, surrounding some obscure christian myths, but there are other origin stories going back to the 14th century. | 13:27 |
| lucasagomes | to make sure we won't break anything | 13:27 |
| mat128 | lucasagomes: not that I'm stalking you, but reviewed it already :) | 13:28 |
| lucasagomes | mat128, a-ha! awesome thank you a lot! | 13:28 |
| *** absubram has joined #openstack-ironic | 13:28 | |
| NobodyCam | persia: oh wow. I had no idea. | 13:29 |
| NobodyCam | mat128: there is always https://review.openstack.org/#/c/272658 which could use s few eyes :p | 13:30 |
| jroll | persia: hehehe | 13:30 |
| persia | For me, it's a fun coincidence that happens every six or seven months (usually), but reading "TGIF" today was particularly amusing. | 13:30 |
| NobodyCam | :p | 13:31 |
| *** absubram_ has joined #openstack-ironic | 13:32 | |
| *** jcoufal has joined #openstack-ironic | 13:32 | |
| *** absubram has quit IRC | 13:32 | |
| *** absubram_ is now known as absubram | 13:32 | |
| *** wajdi has quit IRC | 13:33 | |
| *** wajdi has joined #openstack-ironic | 13:33 | |
| NobodyCam | lucasagomes: +2 | 13:34 |
| chopmann | hi hi folks, someone from Ironic Projetc here? We (tloeb, chopmann, jaoh) have some question regarding ironic development... Anyone got around 10-20 Min time? :-) | 13:34 |
| lucasagomes | NobodyCam, cheerios | 13:35 |
| *** wajdi_ has joined #openstack-ironic | 13:39 | |
| *** athomas has quit IRC | 13:41 | |
| *** wajdi has quit IRC | 13:43 | |
| *** openstack has joined #openstack-ironic | 13:46 | |
| *** jcoufal has quit IRC | 13:47 | |
| chopmann | we just saw the team meeting is on monday. The "Open Discussion" Block is for that sort of things right? | 13:47 |
| *** athomas has joined #openstack-ironic | 13:48 | |
| *** parlos has quit IRC | 13:48 | |
| mgould | chopmann, kinda | 13:48 |
| chopmann | mgould: we are planning to build a Network Devices Provisioner. Most Switches/Routers use dhcp+tftp for Zero Touch Config/ Initial Provisioning | 13:48 |
| mgould | OK | 13:49 |
| chopmann | we dont know if ironic is the right place for such a feature | 13:49 |
| *** mtanino has joined #openstack-ironic | 13:49 | |
| mgould | chopmann, that sounds like more of a Neutron feature to me | 13:49 |
| mgould | but I'm not very familiar with Neutron, so I could be totally wrong | 13:49 |
| *** joprovost has joined #openstack-ironic | 13:50 | |
| mgould | generally Ironic only provisions baremetal compute nodes | 13:50 |
| mgould | but sure, I guess you could treat your switches as baremetal nodes and provision them with Ironic | 13:50 |
| chopmann | it's less about setting up networks...we want to do the initial config on some OpenFlow switches so they can connect to the Controller | 13:50 |
| mgould | oh, OK | 13:50 |
| mat128 | chopmann: it could be done | 13:51 |
| jroll | I would love for ironic to provision network gear | 13:51 |
| mat128 | chopmann: in our internal backlog of things we have something similar to this | 13:51 |
| jroll | switches, firewalls, etc | 13:51 |
| mat128 | yup | 13:51 |
| chopmann | yup something like it | 13:51 |
| sambetts | ++ for deploying network gear | 13:51 |
| jaoh | cool | 13:51 |
| mgould | does OpenStack currently have anything for deploying network gear, or does it have to be done manually? | 13:52 |
| chopmann | nothing we know of | 13:52 |
| mat128 | mgould: I don't think theres anything to *deploy* network gear | 13:52 |
| sambetts | Ironic + https://en.wikipedia.org/wiki/Converged_infrastructure == Win | 13:52 |
| *** openstack has joined #openstack-ironic | 13:54 | |
| mgould | sounds like chopmann already has network gear that can boot from an image over the network | 13:55 |
| chopmann | sort of | 13:55 |
| *** ChrisAusten has joined #openstack-ironic | 13:55 | |
| chopmann | you can boot them up, and the get the config from the tftp server | 13:55 |
| mgould | ah, OK | 13:55 |
| chopmann | or you can "tell them" to update the switch. | 13:56 |
| mgould | can we treat "switch config file" as an image format in Glance? | 13:56 |
| chopmann | that sort of stuff. BlackBox Switches | 13:56 |
| mat128 | hook that into -inspector and you have datacenter profiling | 13:56 |
| mat128 | mgould: I dont think you would want to have a static config in glance | 13:56 |
| mat128 | depending on what is connected | 13:56 |
| *** ChrisAusten has quit IRC | 13:57 | |
| mgould | mat128, oh, good point | 13:57 |
| mat128 | you might want to use something like https://github.com/internap/netman to control the switch once it's up and running | 13:57 |
| chopmann | something more like a template based --> say something like a C***O Switch Family | 13:57 |
| *** sacharya has joined #openstack-ironic | 13:57 | |
| mgould | does neutron assume you've already configured your physical network infrastructure? | 13:57 |
| sambetts | tbh, unless your deploying an OS image onto a blank switch I don't think Ironic is the right tool for the job | 13:57 |
| chopmann | for whitebox switches, Ironic fits the bill | 13:58 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: WIP: grenade checking https://review.openstack.org/311109 | 13:58 |
| chopmann | (sort of) | 13:58 |
| mat128 | mgould: https://github.com/openstack/networking-generic-switch/ | 13:58 |
| mat128 | would make neutron use the switch | 13:58 |
| mat128 | but for initial provisioning | 13:58 |
| mgould | mat128, thanks! | 13:58 |
| sambetts | mat128, mgould | 13:58 |
| mat128 | I planned to integrate netman in networking-generic-switch | 13:58 |
| mgould | chopmann, first time I've heard "whitebox" used like that, but I like it :-) | 13:59 |
| *** derekh has quit IRC | 13:59 | |
| sambetts | or you could use any of our (Cisco) ml2 drivers | 13:59 |
| sambetts | for our switches | 13:59 |
| mgould | or other vendor-specific drivers, if you'd bought someone else's gear in a moment of craziness? :-) | 13:59 |
| sambetts | its an interesting problem, because it depends if your wanting to configure undercloud/OOB switches or overcloud switches | 13:59 |
| mgould | presumably "I've just plugged a new rack into my datacenter, please add it to my overcloud" is an important use case | 14:00 |
| sambetts | not sure if you can run neutron in a tripleo undercloud | 14:00 |
| sambetts | I assume you can | 14:00 |
| mat128 | sambetts: Oh true, you're from Cisco :) | 14:00 |
| sambetts | ;) | 14:00 |
| mat128 | but yeah, undercloud vs overcloud | 14:01 |
| mgould | sambetts, I think you can run neutron in the undercloud, let me check... | 14:01 |
| *** dansmith is now known as superdan | 14:02 | |
| *** joprovost has quit IRC | 14:02 | |
| *** joprovost1 has joined #openstack-ironic | 14:02 | |
| sambetts | if you can run neutron in your undercloud, then you can use your undercloud to bootstrap everything into your overcloud via ml2 I would expect | 14:02 |
| *** sacharya has quit IRC | 14:02 | |
| *** mjura has quit IRC | 14:03 | |
| sambetts | (not tried personally) | 14:03 |
| *** ametts has joined #openstack-ironic | 14:04 | |
| * mgould spins up a tripleo box | 14:04 | |
| *** joprovost1 is now known as joprovost | 14:05 | |
| chopmann | we want to host some SDN Controllers in OpenStack and use "Ironic" to provision/configure the HW-Switches to talk to the SDN controllers inside openstack | 14:05 |
| chopmann | :-) | 14:05 |
| mat128 | thats good | 14:05 |
| mgould | sambetts, don't you need neutron in the undercloud for deploying Ironic nodes? | 14:06 |
| * mgould thought we relied on neutron's dhcp service for that | 14:06 | |
| sambetts | mgould: good point! | 14:06 |
| mat128 | chopmann: I think it fits the bill | 14:06 |
| mariojv | good morning yuriyz - i had a question about your comment on https://review.openstack.org/#/c/226058/15/ironic/conductor/manager.py about the "or get driver status" | 14:06 |
| *** absubram has quit IRC | 14:06 | |
| mariojv | it looks like current master has that particular comment: https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L282 | 14:06 |
| *** absubram has joined #openstack-ironic | 14:06 | |
| yuriyz | hi mariojv | 14:06 |
| mariojv | so i am not sure whether i should remove it | 14:07 |
| mgould | mat128, chopmann so the only problem is that the "image file" is actually a config file for the switch, which we might want to generate dynamically? | 14:07 |
| chopmann | jep | 14:07 |
| mat128 | chopmann: if you switches are on PDUs and/or have IPMI, you will be able to power control them | 14:07 |
| mat128 | then you can add them a port, for the management MAC address and get them to fetch their config via TFTP | 14:07 |
| yuriyz | this is not directly related to your change, but looks like artefact of old code | 14:07 |
| jroll | honestly how I would do it is a base "kick" config file that ironic puts down, and then use neutron from there | 14:08 |
| mat128 | mgould: I think the "image file" would only be the base setup for it to be reachable | 14:08 |
| mariojv | ok - i may as well remove it then since i'm editing the docstring | 14:08 |
| mariojv | thank you | 14:08 |
| mgould | jroll, so just enough for the switch to talk to Neutron? | 14:08 |
| mat128 | mgould: then they enroll the switch in their SDN controller *or* neutron | 14:08 |
| mgould | nice | 14:08 |
| sambetts | so I think a custom deploy driver in Ironic would work that doesn't do image deployment | 14:08 |
| jroll | mgould: enough for neutron to configure the switch correctly, yeah | 14:08 |
| mgould | yeah, that shouldn't change too often | 14:08 |
| sambetts | but it wouldn't be able to be driven from nova | 14:08 |
| yuriyz | mariojv yes better remove it | 14:08 |
| mat128 | sambetts: thats correct | 14:08 |
| jroll | sambetts: why not? | 14:09 |
| jroll | config file could be a raw image | 14:09 |
| mat128 | sambetts: nova boot --flavor=Cisco3750 --image=iOS ? | 14:09 |
| mgould | sambetts, personally I'm happy to consider a config file as code for a *very* high level instruction set architecture :-) | 14:09 |
| mat128 | ;) | 14:09 |
| jroll | (which is actually just text) | 14:09 |
| mariojv | i will put the word "synchronously" back again as well, since that's the default behavior | 14:09 |
| sambetts | mat128: for that I think it would work, but for a config file? | 14:09 |
| mat128 | sambetts: --image=my_basic_template ? | 14:10 |
| jroll | sambetts: mv my_config.txt my_config.img && glance create --file my_config.img | 14:10 |
| mat128 | then maybe use config drive data to assign the management ip? | 14:10 |
| mat128 | jroll: oh god | 14:10 |
| mat128 | so simple | 14:10 |
| chopmann | the base config file tells the switch --> go to IP xxx.xxx.xxx.xxx (SDN Controller), wich OpenFlow Version it should speak and so on... | 14:10 |
| mat128 | why not | 14:10 |
| jroll | mat128: :) | 14:10 |
| sambetts | ewwww, but ok :-P | 14:10 |
| mat128 | well | 14:10 |
| mat128 | images don't have to be x86 images | 14:10 |
| mat128 | we already have whole disk images, partition images, etc | 14:11 |
| * mgould points you all at http://langsec.org/papers/Bratus.pdf, which is about treating input languages as code for implicit abstract "weird machines" | 14:11 | |
| mgould | very cool and rather scary stuff | 14:11 |
| mat128 | mgould: there's this one guy I was interviewing who considered literally *everything* as data transformation | 14:12 |
| mat128 | we didnt hire him,but I still think of this all the time | 14:12 |
| sambetts | then a custom deploy driver to set the right DHCP_OPTs and a weird config image should work today I guess | 14:12 |
| mgould | mat128, yeah, that can be a very powerful perspective | 14:13 |
| mgould | not always the most appropriate one, though :-) | 14:13 |
| chopmann | sambetts: we'll have a look at that | 14:13 |
| mat128 | mgould: I think he used to write switch/firewall firmwares, so I think he took it from there | 14:13 |
| mgould | mat128, right, I can see that | 14:14 |
| mat128 | mgould: He wasn't very comfortable with anything high level, unfortunately | 14:14 |
| mgould | multiple levels of interpretation | 14:14 |
| *** absubram has quit IRC | 14:14 | |
| mgould | heh, I have the opposite problem :-) | 14:14 |
| lucasagomes | the provisioning network gear makes me think of www.onie.org | 14:14 |
| mat128 | lucasagomes: interesting, goes back to "whitebox" switches that chopmann was mentioning | 14:15 |
| lucasagomes | yeah | 14:15 |
| jaoh | onie uses the same protocols | 14:15 |
| jaoh | it could also be done via tftp dhcp | 14:15 |
| mgould | lucasagomes, oh, cool | 14:15 |
| mat128 | lucasagomes: I remember seeing Cumulus Networks' demo in Vancouver where neutron-agent was simply running on their switches | 14:15 |
| sambetts | mat128: ^^ thats who I was reminded of by this convo | 14:16 |
| *** jcoufal has quit IRC | 14:16 | |
| lucasagomes | mat128, oh sweet | 14:16 |
| mat128 | lucasagomes: It was literally a linux box, running specialized hardware | 14:16 |
| mgould | mat128, cool! I guess that's a lot more efficient than VM -> physical switch -> virtual switch -> physical switch -> second VM... | 14:17 |
| mat128 | https://twitter.com/MathieuMitchell/status/600387166686982144 | 14:17 |
| mat128 | before I knew he was from Cumulus and he did a demo in the PM | 14:17 |
| *** vishwanathj has joined #openstack-ironic | 14:17 | |
| mgould | mat128, hahah | 14:18 |
| sambetts | hahaha brilliant! | 14:18 |
| mat128 | mgould: apparently it was very hard to get the box through customs | 14:18 |
| lucasagomes | lol | 14:18 |
| NobodyCam | mat128: lol | 14:19 |
| *** links has joined #openstack-ironic | 14:19 | |
| * sambetts feels sorry for the tech reviewers that got nivida gtx1080s for review at the press release and then had customs unbox them for them on the way home | 14:19 | |
| mgould | :-( | 14:20 |
| chopmann | we'll have a look at the info you guys gave us and probably come to the meeting and hopefully we get a slot on "Open Discussion" :-) | 14:20 |
| *** fellypefca has quit IRC | 14:20 | |
| chopmann | ty all! | 14:20 |
| mat128 | chopmann: make sure you stay around, I'm very interested by your project | 14:20 |
| openstackgerrit | Mario Villaplana proposed openstack/ironic: Allow vendor drivers to acquire shared locks https://review.openstack.org/226058 | 14:20 |
| sambetts | looking forward to it, /me interested too | 14:20 |
| mat128 | Have to leave for a few hours | 14:21 |
| *** mat128 is now known as mat128|away | 14:21 | |
| *** joprovost1 has joined #openstack-ironic | 14:21 | |
| sambetts | o/ mat128|away | 14:21 |
| JayF | jroll: ty for the bug link | 14:22 |
| wajdi_ | sambetts: "customs unbox them for them on the way home"....that's brutal! | 14:23 |
| *** joprovost has quit IRC | 14:23 | |
| *** joprovost1 is now known as joprovost | 14:23 | |
| vsaienko | jroll, lucasagomes could you please have look one more time https://review.openstack.org/#/c/304473/. I have missed to add some parameters after rebasing to master | 14:23 |
| * jlvillal reads the backlog and would be interested to see actual hardware for 'whitebox' switches. Or for ONIE stuff | 14:23 | |
| lucasagomes | vsaienko, sure | 14:24 |
| JayF | I would really like some ironicers to review https://review.openstack.org/#/c/313800/ -- as much as we've talked about wanting the gate to be tinyipa, this has been up two days with no reviews to make the change | 14:24 |
| *** jaybeale has joined #openstack-ironic | 14:25 | |
| *** piet has quit IRC | 14:25 | |
| *** piet has joined #openstack-ironic | 14:26 | |
| vsaienko | jroll, lucasagomes thanks | 14:26 |
| jaoh | hi jlvillal are you looking for something like this | 14:27 |
| jaoh | https://www.deltacomputer.com/edge-core-as4610-30p-poe-bms.html | 14:27 |
| jroll | krotscheck: lucasagomes: is it possible for one of you to invert this for me? it's going on a black background https://wiki.openstack.org/wiki/File:Bear_metal.svg | 14:27 |
| JayF | anyone else seen gate failures like http://logs.openstack.org/35/313035/6/check/gate-tempest-dsvm-ironic-agent_ssh/698f36c/logs/screen-ir-cond.txt.gz#_2016-05-13_10_04_54_582 | 14:27 |
| jroll | ("open up gimp and do X" is acceptable) | 14:27 |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Add support for the audit middleware https://review.openstack.org/272658 | 14:27 |
| jroll | JayF: yep https://github.com/openstack/ironic-python-agent/commit/085e622945e273cc759ff4bff7c08d15fc1e15c6 | 14:27 |
| lucasagomes | jroll, yup, 1 sec | 14:27 |
| lucasagomes | looking at JayF's patch | 14:27 |
| JayF | jroll: so that means my ramdisk defaults change is impacting the gate | 14:27 |
| JayF | https://review.openstack.org/#/c/313035/ | 14:28 |
| jroll | sambetts: hmm http://logs.openstack.org/08/085e622945e273cc759ff4bff7c08d15fc1e15c6/post/ironic-python-agent-buildimage-tinyipa/27deb9e/console.html | 14:28 |
| JayF | jroll: because agent_ssh shouldn't have been running with tinyipa | 14:28 |
| *** jaoh has quit IRC | 14:28 | |
| sambetts | jroll: yeah I was just trying to work out why that didn't build | 14:28 |
| jroll | JayF: are you sure? | 14:28 |
| JayF | I'll make sure it's tinyipa that's booting there | 14:28 |
| jroll | it is | 14:28 |
| JayF | then my patch is wrong | 14:28 |
| *** d0ugal has quit IRC | 14:29 | |
| *** links has quit IRC | 14:29 | |
| jroll | what should agent_ssh be using? | 14:29 |
| JayF | project-config should've isolated the defaults change from the gate | 14:29 |
| JayF | agent_ssh should be using coreos until my project-config change lands | 14:29 |
| jroll | ah | 14:29 |
| JayF | basically trying to split "devstack default" from "gate config" | 14:29 |
| jroll | JayF: yep, project-config matches what you say | 14:30 |
| JayF | but obviously that's not what's happening | 14:30 |
| JayF | maybe a devstack bug as much as anything else? | 14:30 |
| JayF | or perhaps it's like ipxe, with the default built in | 14:30 |
| JayF | I'll figure it out | 14:30 |
| *** d0ugal has joined #openstack-ironic | 14:30 | |
| jroll | p-c looks right | 14:30 |
| jroll | JayF: OH | 14:31 |
| openstackgerrit | Sam Betts proposed openstack/ironic-python-agent: Make sure tinyipa du takes into account branch https://review.openstack.org/316124 | 14:31 |
| jroll | you have export IRONIC_RAMDISK_TYPE={ramdisk-type} | 14:31 |
| sambetts | jroll: ^^^^^ | 14:31 |
| JayF | well *I* don't, lol | 14:31 |
| jroll | shoulc be export DEVSTACK_LOCAL_CONFIG+=$'\n'"..." | 14:31 |
| jroll | no never mind I'm an idiot | 14:32 |
| JayF | that's been configured that way in the gate | 14:32 |
| JayF | for forever maybe | 14:32 |
| jroll | at any rate, that isn't making it to the gate for the thing you linked | 14:32 |
| jroll | yeah I missed the next line | 14:32 |
| sambetts | jroll, JayF: we need this to pass post https://review.openstack.org/316124 :( | 14:32 |
| JayF | jroll: ^ I already +2 | 14:33 |
| jroll | +A | 14:33 |
| jroll | sigh | 14:33 |
| sambetts | mucho sigh... | 14:33 |
| sambetts | again not caught by the gate because that branch value is set in our normal gate | 14:34 |
| mgould | JayF, does agent_ssh require PXE? | 14:34 |
| mgould | also libvirt, those seem to be the two that you haven't changed ipxe-enabled on | 14:35 |
| JayF | mgould: agent_ssh is using the old bash ramdisk | 14:35 |
| mgould | ah, gotcha | 14:35 |
| JayF | mgould: also I was aksed to keep one job running normal-pxe | 14:35 |
| JayF | mgould: libvirt, I honestly had nfi what that job was or did so I left it be | 14:35 |
| mgould | yeah, that makes sense | 14:35 |
| lucasagomes | jroll, https://wiki.openstack.org/wiki/File:Bear_metal_white.svg | 14:36 |
| mgould | I think you could probably run tinyipa on that one too | 14:36 |
| lucasagomes | looks like a white square but the bear is there | 14:36 |
| *** jcoufal has joined #openstack-ironic | 14:37 | |
| jroll | lucasagomes: thanks man | 14:37 |
| *** d0ugal has quit IRC | 14:38 | |
| jlvillal | jaoh thanks! Though I think you left | 14:40 |
| jlvillal | lucasagomes, So is the movie Sing Street popular where you are? | 14:41 |
| jlvillal | vdrok, You have spent more time looking in the devstack code than me. Do we ever not set this to true? https://github.com/openstack/ironic/blob/master/devstack/lib/ironic#L92 | 14:43 |
| mgould | JayF, left some comments/questions | 14:43 |
| JayF | mgould: ty | 14:43 |
| mgould | most can probably be answered with "it's fine, you've just misunderstood how jobs are configured" | 14:43 |
| lucasagomes | jlvillal, hi there! hmmm never heard of it heh | 14:43 |
| * lucasagomes looks | 14:43 | |
| jlvillal | lucasagomes, Just wondering. It is supposed to be about a fictional Dublin band... | 14:44 |
| * jlvillal will try to drag his wife to go see it this weekend | 14:44 | |
| lucasagomes | jlvillal, oh right, yeah I don't know... I don't think it's very popular here | 14:44 |
| JayF | mgould: well, I'm not sure I understand how they work now, given my patch broke so bad (to change devstack defaults | 14:44 |
| jlvillal | lucasagomes, Thanks :) | 14:44 |
| mgould | JayF, heh | 14:44 |
| lucasagomes | jlvillal, OTOH the commitments is | 14:44 |
| lucasagomes | jlvillal, https://en.wikipedia.org/wiki/The_Commitments_%28film%29 | 14:45 |
| jlvillal | lucasagomes, That is a good movie :) | 14:45 |
| * mgould hasn't heard of Sing Street either, but it looks fun | 14:45 | |
| vdrok | jlvillal: setting to false will disable vm creation and network setup I guess | 14:45 |
| jlvillal | vdrok, Yeah I was just wondering when that would be a valid case. | 14:45 |
| lucasagomes | jlvillal, yeah... if you are looking for irish movie/tv stuff you may want to check "love/hate" | 14:45 |
| lucasagomes | or "the fall" | 14:46 |
| lucasagomes | jlvillal, https://en.wikipedia.org/wiki/Love/Hate_%28TV_series%29 | 14:46 |
| jlvillal | lucasagomes, Thanks. Or is that: go raibh maith agat | 14:46 |
| * jlvillal thinks it looks complicated to say 'thank you' in Irish. | 14:46 | |
| lucasagomes | heh yeah I don't even know how to pronounce that | 14:47 |
| jlvillal | Me either! :) | 14:47 |
| * lucasagomes knows very few irish words | 14:47 | |
| jlvillal | lucasagomes, beoir ? | 14:47 |
| jlvillal | = beer | 14:47 |
| vdrok | jlvillal: the only case I can recall right now is ironicclient functional tests job - https://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/ironic.yaml#L43 | 14:47 |
| *** mbound has quit IRC | 14:48 | |
| lucasagomes | jlvillal, hah yeah or slainte | 14:48 |
| vdrok | as we just do the api calls | 14:48 |
| jlvillal | vdrok, Ah so there is at least one case. That is good. | 14:48 |
| lucasagomes | madra rua = red dog = fox | 14:48 |
| lucasagomes | only random stuff heh | 14:48 |
| jlvillal | :) | 14:48 |
| * mgould has a friend who jokes that H in Gaelic has a silent ^ before it | 14:48 | |
| * jlvillal better get ready to go to work... | 14:48 | |
| mgould | so go raib^H mait^H agat | 14:49 |
| mgould | gets you reasonably close, AIUI | 14:49 |
| JayF | jroll: https://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/devstack-gate.yaml#L683 requires deploy-with-ipa to be true to set ramdisk type. https://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/devstack-gate.yaml#L808 has it set to false, as does agent_ssh and agent_ipmitool | 14:49 |
| jroll | O_o | 14:49 |
| JayF | jroll: I'm thinking we should unconditionally set the RAMDISK_TYPE | 14:49 |
| JayF | jroll: unless it causes harm for other setups? | 14:49 |
| jroll | JayF: curious what that does for bash ramdisk | 14:50 |
| jroll | but that's my only concern | 14:50 |
| JayF | which job is supposed to be bash ramdisk? | 14:50 |
| mgould | when are we EOLing the bash ramdisk? | 14:50 |
| JayF | pxe_ssh? | 14:50 |
| JayF | if so I'm 99% sure it's not running bash ramdisk today | 14:50 |
| jroll | yes, pxe_ssh | 14:50 |
| mgould | rename it to *_bash_ramdisk* for greater clarity, maybe | 14:50 |
| jroll | wat | 14:50 |
| jroll | since when | 14:50 |
| JayF | IDK; but my patch wouldn't have modified the behavior of the pxe_ssh job | 14:51 |
| JayF | if it had been running the bash ramdisk | 14:51 |
| JayF | because that's a completely diff codepath | 14:51 |
| JayF | going to doublecheck the logs | 14:51 |
| JayF | oh, agent_ssh was my failure | 14:51 |
| lucasagomes | JayF, in ur patch you may end up passing an empty IRONIC_RAMDISK_TYPE= no? | 14:51 |
| lucasagomes | for stable/{kilo, liberty} | 14:51 |
| JayF | lucasagomes: yes, and that's desired, so it uses the default | 14:51 |
| JayF | lucasagomes: since for k and l we only had 1 ramdisk | 14:51 |
| JayF | lucasagomes: otherwsie k and l jobs would fail post-change with inability to fetch ipa | 14:52 |
| lucasagomes | JayF, strange... I mean, yeah it should use the default but export an empty one seems odd | 14:52 |
| JayF | yeah jroll this isn't as bad as I thought, one character fix | 14:52 |
| lucasagomes | better to not export anything, right? | 14:52 |
| JayF | lucasagomes: I think that's what I'm doing, right? | 14:52 |
| JayF | if you think not please put it on the patch so the knowledge isn't lost | 14:52 |
| JayF | I'll have to rebase it on this fix anyway | 14:52 |
| jroll | JayF: cool | 14:53 |
| lucasagomes | JayF, ish yeah, https://review.openstack.org/#/c/313800/3/jenkins/jobs/devstack-gate.yaml (see L89) | 14:53 |
| lucasagomes | I would just put that inside the if conditional | 14:53 |
| lucasagomes | or get rid of it and make it a export DEVSTACK_LOCAL_CONFIG+=$'\n'"IRONIC_RAMDISK_TYPE=${ramdisk-type}" | 14:53 |
| lucasagomes | inside the if | 14:53 |
| lucasagomes | L689 | 14:54 |
| JayF | oh, gotcha | 14:54 |
| lucasagomes | what I mean is, $IRONIC_RAMDISK_TYPE may be unbonded | 14:54 |
| JayF | good catch | 14:54 |
| JayF | just put it on a comment and I'll fix it with the next patchset | 14:54 |
| JayF | I see what you mean sinec opening the code | 14:54 |
| lucasagomes | JayF, done | 14:54 |
| JayF | jroll: https://review.openstack.org/#/c/316138/ | 14:55 |
| *** rama_y has joined #openstack-ironic | 14:56 | |
| jroll | JayF: thanks | 14:57 |
| JayF | once tests pass I'll go beg for reviews from infra | 14:57 |
| JayF | and once this lands and takes effect, we can recheck my devstack defaults change patch | 14:57 |
| JayF | I thought that changing those defaults might unearth some craziness :) | 14:58 |
| *** sacharya has joined #openstack-ironic | 14:58 | |
| lucasagomes | jroll, btw, re root device hints operators... should we just wait for the oslo spec matcher thing and use it instead? | 14:58 |
| jroll | lucasagomes: possibly/probably? | 14:59 |
| lucasagomes | it's hard to map the other because eq , ==, s== ?! | 14:59 |
| jroll | yeah, insanity | 14:59 |
| lucasagomes | the type awareness of the nova spec match Vs. the non-type aware of inspector makes it crazy | 14:59 |
| lucasagomes | yeah | 14:59 |
| lucasagomes | dtantsur, ^ btw, re mapping the root device hints operators with inspector and nova | 15:01 |
| lucasagomes | seems like a overcomplex interface, I would rather chose 1 | 15:02 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Fix all possible node subresources routing collisions https://review.openstack.org/315766 | 15:03 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Fix invalid chassis/nodes/chassis_uuid API url https://review.openstack.org/316149 | 15:03 |
| *** d0ugal has joined #openstack-ironic | 15:05 | |
| *** dmk0202 has quit IRC | 15:13 | |
| *** openstackgerrit has quit IRC | 15:17 | |
| *** openstackgerrit has joined #openstack-ironic | 15:18 | |
| blinky_ghost | anyone can help me with this error when I'm trying to deploy overcloud: "No configuration found for any of the present NICs"? thanks | 15:31 |
| *** absubram has joined #openstack-ironic | 15:31 | |
| *** mgould has quit IRC | 15:38 | |
| *** yarkot has quit IRC | 15:43 | |
| *** anush has quit IRC | 15:44 | |
| *** absubram has quit IRC | 15:47 | |
| *** mbound has joined #openstack-ironic | 15:48 | |
| *** anush has joined #openstack-ironic | 15:49 | |
| *** mgould has joined #openstack-ironic | 15:49 | |
| *** ifarkas has quit IRC | 15:51 | |
| JayF | jroll: that project-config change is landing now, if you don't mind, recheck my devstack patch once it takes effect so it can land? I put explicit instructions on how to make sure the changes took effect in the gerrit comment | 15:51 |
| * JayF off the rest of the day, see you on on Monday o/ | 15:51 | |
| sambetts | o/ JayF | 15:51 |
| *** mbound has quit IRC | 15:53 | |
| *** MattMan has quit IRC | 15:54 | |
| *** MattMan has joined #openstack-ironic | 15:54 | |
| *** tesseract has quit IRC | 15:56 | |
| *** absubram has joined #openstack-ironic | 15:56 | |
| * leakypipes shakes fist at JayF for causing him many IRC highlights... :) | 15:57 | |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add tests for openstackclient plugin https://review.openstack.org/301687 | 15:57 |
| *** jcoufal has quit IRC | 16:00 | |
| blinky_ghost | lucasgomes: can I compile this https://github.com/umago/ironic-staging-drivers-packaging/blob/master/openstack-ironic-staging-drivers-0.1.0-1.src.rpm to tryout libvirt? thanks | 16:01 |
| *** alexpilotti has quit IRC | 16:01 | |
| openstackgerrit | greghaynes proposed openstack/bifrost: Add bonding support to bifrost https://review.openstack.org/316169 | 16:01 |
| *** rcernin has quit IRC | 16:02 | |
| *** alexpilotti has joined #openstack-ironic | 16:04 | |
| mgould | sambetts, just ran `openstack service list` on my undercloud VM, and neutron is indeed there :-) | 16:07 |
| sambetts | mgould: sweet :) | 16:07 |
| openstackgerrit | Merged openstack/virtualbmc: Bump pyghmi version to 1.0.3 https://review.openstack.org/316023 | 16:11 |
| sambetts | mgould: your still better at the englishing than I am XD | 16:14 |
| mgould | :-) | 16:15 |
| mgould | sambetts, short sentences Are Your Friends | 16:16 |
| sambetts | heh | 16:16 |
| lucasagomes | blinky_ghost, sorry missed ur ping | 16:17 |
| lucasagomes | blinky_ghost, so the libvirt driver is part of the 0.2.0 version | 16:17 |
| * lucasagomes needs to change his nick | 16:17 | |
| * mgould wishes he'd watched https://www.youtube.com/watch?v=C9-o1gLHHWo six months ago :-/ | 16:18 | |
| blinky_ghost | lucasagomes: how about this error? "No configuration found for any of the present NICs"? I'm testing with fake_pxe driver | 16:18 |
| lucasagomes | blinky_ghost, do you have any port registered in the node? | 16:19 |
| blinky_ghost | lucasagomes: how do I check that? ironic node-show? | 16:19 |
| lucasagomes | blinky_ghost, ironic port-list | 16:20 |
| blinky_ghost | lucasagomes: Yes I have 2 ports, I'm testing with 2 nodes. One of them, the compute node booted correctly, this one is stuck here, and the disk seems corrupted: "error: not a correct XFS inode". grub rescue | 16:22 |
| lucasagomes | blinky_ghost, right, because that error is coming from the node booting not being able to find the [i]PXE configuration option | 16:22 |
| lucasagomes | blinky_ghost, the mac address of the port is correct? Does it match with the node being booted? | 16:23 |
| blinky_ghost | lucasagomes: yes, the port is correct | 16:23 |
| blinky_ghost | lucasagomes: I did the exactly same test with OSP 8 and it works fine, could be a bug I guess | 16:24 |
| *** e0ne has quit IRC | 16:24 | |
| openstackgerrit | Merged openstack/ironic-python-agent: Make sure tinyipa du takes into account branch https://review.openstack.org/316124 | 16:25 |
| klindgren | So I am working on getting introspection to actually use lldp to report the switches that it is connected to vs's none. I am noticing some work currently in progress that it looks like ironic is trying to standardize on a dict for ports attached to switches. | 16:25 |
| klindgren | one of those items appears to be the switchid? which looks to be nothing more than the switches mac-address? | 16:26 |
| lucasagomes | blinky_ghost, right, when deploying can you check if the configuration exist under /httpboot/pxelinux.cfg/<mac addresses> ? (Or if you are using devstack /opt/stack/data/ironic/httpboot/pxelinux.cfg/...) | 16:26 |
| blinky_ghost | lucasagomes: it's empty | 16:26 |
| lucasagomes | blinky_ghost, right, but there's no node in "wait call-back" or "deploying" right? | 16:27 |
| lucasagomes | cause it will be created at that stage | 16:27 |
| lucasagomes | and cleaned up if the deployment fails etc... | 16:27 |
| klindgren | So if siwtch_id == switch mac addy(eg: xx:xx:xx:xx:xx:xx), switch_info == switch name(eg: switch.fqdn), port_id == switch port (eg eth1/101/5) | 16:28 |
| blinky_ghost | lucasagomes: I start the node by hand when it asks for wait call-back, it start's deploying then it shutdowns the nodes, I see a message "emergency shutdown". After this, the wait call-back never shows up again, so I start them by hand, and that's where I see the error. | 16:28 |
| *** fragatin_ has joined #openstack-ironic | 16:29 | |
| sambetts | klindgren: So the format is switch_id == mac_addr/openflow_id, switch_info == any string that can be customisable per vendor, and port_id == switch port | 16:30 |
| lucasagomes | blinky_ghost, oh ok, so deployment is failing... that's why after it failed | 16:30 |
| lucasagomes | blinky_ghost, ironic will remove the PXE configs, so after you restart it manually it will not boot anymore | 16:30 |
| lucasagomes | blinky_ghost, so, we need to figure out why it's failing in the first place. Any hints in the ironic-conductor log | 16:30 |
| lucasagomes | ? | 16:30 |
| sambetts | klindgren: if your planning to tackle this work, please look at the existing work, and also at the comment my colleage Robert Li left on the existing work | 16:30 |
| *** sabeen has joined #openstack-ironic | 16:31 | |
| blinky_ghost | lucasagomes: in OSP8 I don't see the emergency shutdown, actually it never shutdowns the node, I have to shut it down by hand as it's fake_pxe | 16:31 |
| blinky_ghost | lucasagomes: I'll redeploy and check the the ironic-conductor.log | 16:32 |
| *** fragatina has quit IRC | 16:32 | |
| lucasagomes | blinky_ghost, ok, and you are using the same image for both deployments ? | 16:32 |
| lucasagomes | ramdisks and tenant image? | 16:32 |
| sambetts | klindgren: we have this working down stream now, and switch_info is designed allow for vendor specific info, because of that the IPA side of it needs to do very little processing of the lldp packet, allowing plugging of vendor specific procssing hooks into inspector to handle different vendors switch processing | 16:33 |
| blinky_ghost | lucasagomes: no, I built one to Centos, from OSP8 I've installed a rpm package | 16:33 |
| lucasagomes | blinky_ghost, right so you built a new image for the tenant? Which is a CentOS based one? | 16:34 |
| blinky_ghost | lucasagomes: can I copy deploy-ramdisk-ironic.kernel deploy-ramdisk-ironic.initramfs from OSP8 and use from Centos? | 16:34 |
| *** athomas has quit IRC | 16:34 | |
| blinky_ghost | lucasagomes: Yes, correct | 16:35 |
| lucasagomes | blinky_ghost, you can yes | 16:35 |
| lucasagomes | blinky_ghost, so to build the tenant image, it may worth taking a look at http://docs.openstack.org/developer/ironic/deploy/install-guide.html#image-requirements | 16:35 |
| lucasagomes | e.g, if you want to use local boot, make sure grub2 is installed in the iamge | 16:35 |
| lucasagomes | image* | 16:35 |
| lucasagomes | same for EFI, make sure the modules are present in the image | 16:35 |
| lucasagomes | and things like that | 16:35 |
| blinky_ghost | Iucasagomes: Can I use everything from OSP8 except overcloud-full.qcow2 to test or is better to buld a new one? | 16:36 |
| klindgren | sambetts, I am curious what one would do with the switch_id? Which neutron mechanism driver uses that as identification for the switch that a change needs to be made on? | 16:38 |
| lucasagomes | blinky_ghost, its really up to you, for testing you can even use the cirros image | 16:39 |
| lucasagomes | which is tiny :-) | 16:39 |
| lucasagomes | (that's what we use in gate as well) | 16:39 |
| lucasagomes | blinky_ghost, what I mean is, if you want to build ur image just make sure it contains all the required bits (following http://docs.openstack.org/developer/ironic/deploy/install-guide.html#image-requirements) | 16:40 |
| sambetts | klindgren: I can't comment on which specific ml2 drivers use mac-addr or open-flow id, because ours doesn't it uses the mgnt ip, which we include in switch_info, which we have as a string encoded json | 16:40 |
| blinky_ghost | lucasagomes: OK, I'll test both ways thanks :) | 16:40 |
| lucasagomes | np :-) | 16:40 |
| sambetts | klindgren: I've campained to make it a generic string, but I was fought against | 16:41 |
| *** jistr has quit IRC | 16:42 | |
| *** Goneri has quit IRC | 16:44 | |
| lucasagomes | ok folks, I'm calling it a day | 16:44 |
| lucasagomes | have a great evening all, enjoy the weekend! | 16:44 |
| lucasagomes | talk to you all on monday :-) | 16:44 |
| *** daemontool_ has joined #openstack-ironic | 16:45 | |
| *** lucasagomes is now known as lucas-afk | 16:45 | |
| sambetts | o/ lucas-afk | 16:45 |
| lucas-afk | sambetts, see ya! | 16:45 |
| NobodyCam | have a good night | 16:45 |
| *** kromanenko_ has quit IRC | 16:47 | |
| *** kromanenko__ has quit IRC | 16:47 | |
| *** daemontool has quit IRC | 16:48 | |
| *** tloeb has quit IRC | 16:52 | |
| *** chopmann has quit IRC | 16:54 | |
| *** mgould has quit IRC | 16:56 | |
| *** romcheg has quit IRC | 16:56 | |
| *** kozhukalov has quit IRC | 16:56 | |
| *** agordeev has quit IRC | 16:56 | |
| *** yonglihe has quit IRC | 16:56 | |
| *** rajinir has quit IRC | 16:56 | |
| *** evgenyl has quit IRC | 16:56 | |
| *** sergek has quit IRC | 16:56 | |
| *** ChrisMead has quit IRC | 16:56 | |
| *** electrofelix has quit IRC | 16:59 | |
| *** catintheroof has joined #openstack-ironic | 16:59 | |
| *** catintheroof has quit IRC | 17:00 | |
| *** catintheroof has joined #openstack-ironic | 17:00 | |
| *** alexpilotti has quit IRC | 17:01 | |
| *** alexpilotti has joined #openstack-ironic | 17:03 | |
| *** fragatin_ has quit IRC | 17:07 | |
| *** ijw has joined #openstack-ironic | 17:08 | |
| blinky_ghost | lucasagomes: after copying ironic-python-agent.initramfs ironic-python-agent.kernel overcloud-full.initrd overcloud-full.vmlinuz from OSP8 | 17:10 |
| blinky_ghost | it works fine | 17:10 |
| *** ohamada has quit IRC | 17:10 | |
| blinky_ghost | lucasagomes: it uses overcloud-full.qcow2 from centos and the rest from OSP8 so it's some kind of bug on ramdisk or something | 17:11 |
| *** rajinir has joined #openstack-ironic | 17:12 | |
| *** kozhukalov has joined #openstack-ironic | 17:14 | |
| *** agordeev has joined #openstack-ironic | 17:14 | |
| *** evgenyl has joined #openstack-ironic | 17:15 | |
| *** amotoki has quit IRC | 17:21 | |
| *** romcheg has joined #openstack-ironic | 17:22 | |
| *** Sukhdev has joined #openstack-ironic | 17:25 | |
| openstackgerrit | Milan Kováčik proposed openstack/ironic-inspector: [WIP] Tempest: add basic test https://review.openstack.org/293926 | 17:25 |
| *** ijw has quit IRC | 17:25 | |
| *** _milan_ has quit IRC | 17:28 | |
| *** athomas has joined #openstack-ironic | 17:28 | |
| openstackgerrit | Sam Betts proposed openstack/ironic-specs: Add spec for VLAN Aware Baremetal Instances https://review.openstack.org/277853 | 17:28 |
| *** yonglihe has joined #openstack-ironic | 17:29 | |
| * sambetts is heading off, see everyone Monday | 17:31 | |
| *** sambetts is now known as sambetts|afk | 17:31 | |
| *** sergek has joined #openstack-ironic | 17:32 | |
| openstackgerrit | Jarrod Johnson proposed openstack/pyghmi: Disable futile attempts to correct uncorrectable time https://review.openstack.org/316218 | 17:32 |
| *** blinky_ghost has quit IRC | 17:34 | |
| *** ChrisMead has joined #openstack-ironic | 17:39 | |
| *** mat128|away is now known as mat128 | 17:42 | |
| *** piet has quit IRC | 17:47 | |
| openstackgerrit | John L. Villalovos proposed openstack/ironic: WIP: Create a early_create grenade phase https://review.openstack.org/316234 | 17:48 |
| openstackgerrit | Hugo Nicodemos proposed openstack/python-oneviewclient: Adding dynamic allocation operations https://review.openstack.org/258699 | 17:48 |
| openstackgerrit | Hugo Nicodemos proposed openstack/python-oneviewclient: Adding dynamic allocation operations https://review.openstack.org/258699 | 17:54 |
| *** alexpilotti has quit IRC | 18:00 | |
| openstackgerrit | Sam Betts proposed openstack/ironic: Fix invalid chassis/nodes/chassis_uuid API url https://review.openstack.org/316149 | 18:01 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Fix all possible node subresources routing collisions https://review.openstack.org/315766 | 18:01 |
| *** alexpilotti has joined #openstack-ironic | 18:06 | |
| *** keedya has joined #openstack-ironic | 18:13 | |
| *** linuxgeek has quit IRC | 18:17 | |
| *** spandhe has joined #openstack-ironic | 18:17 | |
| *** fragatina has joined #openstack-ironic | 18:31 | |
| *** jtomasek_ has joined #openstack-ironic | 18:37 | |
| *** ijw has joined #openstack-ironic | 18:38 | |
| *** jtomasek_ has quit IRC | 18:38 | |
| *** mbound has joined #openstack-ironic | 18:38 | |
| *** jaybeale has quit IRC | 18:39 | |
| *** jtomasek_ has joined #openstack-ironic | 18:39 | |
| *** ayoung has quit IRC | 18:42 | |
| *** e0ne has joined #openstack-ironic | 18:43 | |
| *** ChubYann has joined #openstack-ironic | 18:45 | |
| *** anush has quit IRC | 18:46 | |
| *** linuxgeek has joined #openstack-ironic | 18:46 | |
| *** Goneri has joined #openstack-ironic | 18:56 | |
| *** daemontool_ has quit IRC | 19:06 | |
| *** Sukhdev has quit IRC | 19:12 | |
| *** Sukhdev has joined #openstack-ironic | 19:15 | |
| *** yarkot has joined #openstack-ironic | 19:17 | |
| *** ijw has quit IRC | 19:24 | |
| *** jtomasek_ has quit IRC | 19:27 | |
| *** sabeen has quit IRC | 19:27 | |
| *** Sukhdev has quit IRC | 19:28 | |
| *** Sukhdev has joined #openstack-ironic | 19:28 | |
| *** sabeen has joined #openstack-ironic | 19:29 | |
| *** jaybeale has joined #openstack-ironic | 19:30 | |
| *** ayoung has joined #openstack-ironic | 19:30 | |
| openstackgerrit | Ramamani Yeleswarapu proposed openstack/ironic: Fix misformatted releasenotes of merged patch #314016 https://review.openstack.org/316249 | 19:34 |
| openstackgerrit | John L. Villalovos proposed openstack/ironic: Setup for using the Grenade 'early_create' phase https://review.openstack.org/316234 | 19:43 |
| *** rbudden has quit IRC | 19:44 | |
| jlvillal | cinerama: Thanks for the patch. You beat me to it :) | 19:47 |
| cinerama | heh, no problem | 19:47 |
| cinerama | just poking at this a bit on a quiet afternoon | 19:47 |
| jlvillal | cinerama: I just pushed a new change. | 19:52 |
| * jlvillal trying it out locally now | 19:53 | |
| *** e0ne has quit IRC | 19:59 | |
| cinerama | nice i should try it too | 20:00 |
| *** ijw has joined #openstack-ironic | 20:00 | |
| openstackgerrit | Ramamani Yeleswarapu proposed openstack/ironic: Fix formatting error in releasenotes https://review.openstack.org/316249 | 20:02 |
| *** alexpilo_ has joined #openstack-ironic | 20:02 | |
| *** alexpilotti has quit IRC | 20:02 | |
| klindgren | One final question re: ironic-python-agent + coreos build stuff and using it for on boarding of nodes via ironic_inspector. I am testing this out using kilo right now (but the IPA part is from master). But as currently packaged and release the IPA stuff flat out wont work on anything using IPMI? This is because the coreos image that is configured to use doesn't have the IPMI drivers and - atleast on my hardware. I needed t | 20:02 |
| klindgren | o load additional kernel modules to get the ipmi devices created so that ipmitool would function and find the BMC IP address. | 20:02 |
| *** dmk0202 has joined #openstack-ironic | 20:03 | |
| *** joprovost has quit IRC | 20:03 | |
| klindgren | is the idea that basically every deployment needs to have their own branch to customize the IPA agent to work with their hardware? Or is updating to newer versions of coreos mainly blocked due to the gate stuff? | 20:03 |
| mat128 | klindgren: interesting that you had to load additional modules | 20:05 |
| mat128 | I would file a bug | 20:05 |
| mat128 | We are not expecting operators to fork this to simply "get it working" | 20:06 |
| mat128 | I think we are fetching latest CoreOS nowadays | 20:06 |
| klindgren | I mean asside from the additional modules - that fact that its still using coreos 3xx - where it doesn't even have the ipmi modules built | 20:07 |
| klindgren | https://github.com/openstack/ironic-python-agent/blob/master/imagebuild/coreos/coreos-oem-inject.py#L24 | 20:07 |
| klindgren | still targeting 367.1.0 ? | 20:07 |
| mat128 | oh, the work I was thinking of got reverted | 20:08 |
| mat128 | https://github.com/openstack/ironic-python-agent/commit/943d2c00e94cff38c15bca5ecad40f9a9b27008f | 20:08 |
| mat128 | it was working on 5xx IIRC | 20:08 |
| mat128 | maybe we can pin to something newer, rather than completely rolling this back | 20:08 |
| klindgren | ok - so basically this in on hold until gate gets move to tinyipa | 20:08 |
| mat128 | we will still be supporting CoreOS | 20:09 |
| mat128 | it's just that TinyIPA is so much smaller that it gives us quicker feedback and allows us to run more instances on the same host | 20:09 |
| mat128 | are your modules part of 367 or you needed a fork to get a newer CoreOS? | 20:09 |
| klindgren | I moved to the latest stable just to get any ipmi modules (ipmi_si) - not sure when it was built - I didn't bother tracking it down. But inorder for ipmi0 to be created I also needed to do: | 20:12 |
| klindgren | write_files: | 20:12 |
| klindgren | - path: /etc/modules-load.d/ipmi.conf | 20:12 |
| klindgren | content: ipmi_devintf | 20:12 |
| klindgren | along with a: | 20:12 |
| klindgren | - name: systemd-modules-load.service | 20:12 |
| klindgren | command: restart | 20:12 |
| mat128 | Interesting | 20:13 |
| mat128 | klindgren: I would submit a review for using a newer (but pinned) CoreOS | 20:13 |
| mat128 | if it proves to work with the gate, we can at least have this | 20:13 |
| mat128 | or maybe we can re-revert the "autoroll" version we had in place | 20:13 |
| *** rbudden has joined #openstack-ironic | 20:14 | |
| mat128 | klindgren: we now have virtualBMC which we will use to test against IPMI in the gates | 20:14 |
| mat128 | but in-band IPMI will be harder | 20:14 |
| mat128 | :) | 20:14 |
| *** spandhe has quit IRC | 20:14 | |
| klindgren | yea - hardware :-/ | 20:14 |
| klindgren | standards? whats that? | 20:15 |
| mat128 | Haha | 20:15 |
| mat128 | We'd need IPMI support in QEMU | 20:15 |
| mat128 | everything is possible my friend :) | 20:15 |
| klindgren | one thing I was curious about - is how are people hanlding where for SOL one BMC needs to be set to com1 and another set to com2? | 20:15 |
| mat128 | I remember seeing an implementation of this, it also tried to provide out-of-band access (network) but it didnt really work well | 20:16 |
| klindgren | booting the deploymentr ram disk I can't set both ttys0 and ttys1 as only the first console of each type is used | 20:16 |
| mat128 | s/this/ ipmi in qemu/ | 20:16 |
| mat128 | klindgren: we (as an operator) aren't using serial over LAN at all so I don't know | 20:17 |
| mat128 | but for sure this means you can't hardcode anything in the template | 20:17 |
| mat128 | maybe you can pull this information from the node properties? | 20:17 |
| klindgren | yea - but it also means I need things changed in the images as well | 20:18 |
| klindgren | so I can't have a generic image | 20:18 |
| klindgren | I guess I can fixup after boot | 20:18 |
| klindgren | and get a patch in to pull console terminal from a node property | 20:18 |
| mat128 | that's ugly but yeah it would work | 20:19 |
| klindgren | so process for adding that feature within ironic. bug -> rfe? or will a spec be required (Pulling which console to use from the node properties) | 20:20 |
| mat128 | bug tagged "rfe" then if it requires more details / more design a spec will be requested | 20:20 |
| mat128 | http://lists.openstack.org/pipermail/openstack-dev/2016-May/094658.html | 20:21 |
| mat128 | one of the priorities for this cycle is nova-compatible serial console | 20:21 |
| mat128 | your point fits very well within that topic | 20:21 |
| openstackgerrit | John L. Villalovos proposed openstack/ironic: Setup for using the Grenade 'early_create' phase https://review.openstack.org/316234 | 20:21 |
| *** dmk0202 has quit IRC | 20:22 | |
| mat128 | make sure you also file a bug / open a review to switch to a newer CoreOS as I think this can be valuable to the group | 20:22 |
| mat128 | especially if you did the work | 20:24 |
| jlvillal | NobodyCam: jroll TheJulia: Pretty easy review here: https://review.openstack.org/#/c/316249/ Fixes weirdness seen in the bug-fixes section of: http://docs.openstack.org/releasenotes/ironic/current-series.html | 20:28 |
| jlvillal | devananda: ^^^ | 20:29 |
| *** e0ne has joined #openstack-ironic | 20:31 | |
| NobodyCam | Jlvillal +a :-) | 20:33 |
| jlvillal | Thanks NobodyCam :) | 20:34 |
| *** joprovost has joined #openstack-ironic | 20:35 | |
| *** ametts has quit IRC | 20:36 | |
| openstackgerrit | Merged openstack/ironic: Fix formatting error in releasenotes https://review.openstack.org/316249 | 20:40 |
| *** daemontool has joined #openstack-ironic | 20:46 | |
| *** e0ne has quit IRC | 20:51 | |
| *** daemontool has quit IRC | 20:52 | |
| *** joprovost has quit IRC | 20:53 | |
| jlvillal | jroll: Hmmm. I'm thinking releasenotes when doing stable/mitaka is pulling things from other git history. | 20:55 |
| jroll | jlvillal: yeah, there's known weirdness there afaik | 20:55 |
| jroll | you seeing anything super troubling? or rather what are you seeing? | 20:55 |
| jlvillal | jroll: Like this patch: https://review.openstack.org/#/c/316029/ | 20:56 |
| jlvillal | jroll: http://docs-draft.openstack.org/29/316029/1/check/gate-ironic-releasenotes/38ac413//releasenotes/build/html/mitaka.html | 20:56 |
| jlvillal | So there is breakage in master. But it is also breaking stable/mitaka | 20:56 |
| jroll | jlvillal: hm, published version looks okay http://docs.openstack.org/releasenotes/ironic/mitaka.html | 20:57 |
| jroll | weird | 20:57 |
| *** Sukhdev has quit IRC | 20:57 | |
| jlvillal | I reproduced locally just pulling stable/mitaka and nothing else. | 20:57 |
| devananda | klindgren: it sounds like you'd need to set different pxe append parameters per-node, instead of per-conductor | 20:57 |
| jroll | jlvillal: I think it may be something odd with like, what's in the local git tree, maybe the publish job only pulls in the branch | 20:58 |
| jlvillal | maybe | 20:58 |
| jroll | I'm running away right now though, feel free to file a bug if you track something down | 20:58 |
| jlvillal | jroll: Thanks. Have a great weekend | 20:59 |
| jroll | you too | 20:59 |
| mat128 | devananda: is this already possible? | 21:01 |
| klindgren | devananda, yea. I was hoping we could move more to an override per node model. | 21:01 |
| klindgren | Like setting the IPA image per conductor vs's per node. and setting the cleaning network per node vs's per conductor | 21:01 |
| klindgren | devananda, +1 to what mat128 asked. Is this already possible? I didn't see anything in the documentation that led me to believe it was possible. | 21:03 |
| mat128 | klindgren, devananda we're using a totally static pxe here, so I'd be missing out on the new dynamic templated pxe stuff :) | 21:04 |
| *** jaybeale_ has joined #openstack-ironic | 21:08 | |
| *** jaybeale has quit IRC | 21:08 | |
| mjturek1 | jlvillal: fyi I ran into this while running your grenade test today http://paste.openstack.org/show/497128/ wondering if you've seen it? | 21:09 |
| openstackgerrit | John L. Villalovos proposed openstack/ironic: Setup for using the Grenade 'early_create' phase https://review.openstack.org/316234 | 21:15 |
| jlvillal | mjturek1 Looks like a network issue to me. | 21:17 |
| jlvillal | mjturek1: As in it couldn't download something and then died. | 21:19 |
| mjturek1 | jlvillal: oh duh, I see know. | 21:19 |
| jlvillal | mjturek1: :) | 21:23 |
| *** sabeen has quit IRC | 21:25 | |
| *** lascii is now known as alaski | 21:28 | |
| *** thrash is now known as thrash|wknd | 21:41 | |
| *** keedya has quit IRC | 21:43 | |
| *** dprince has quit IRC | 21:45 | |
| *** jaybeale_ has quit IRC | 21:51 | |
| *** dmk0202 has joined #openstack-ironic | 21:53 | |
| *** keedya has joined #openstack-ironic | 21:56 | |
| *** Sukhdev has joined #openstack-ironic | 21:57 | |
| *** daemontool has joined #openstack-ironic | 21:59 | |
| *** alexpilo_ has quit IRC | 22:01 | |
| *** alexpilotti has joined #openstack-ironic | 22:04 | |
| *** dmk0202 has quit IRC | 22:06 | |
| *** fragatina has quit IRC | 22:11 | |
| openstackgerrit | John L. Villalovos proposed openstack/ironic: WIP: Setup for using the Grenade 'early_create' phase https://review.openstack.org/316234 | 22:19 |
| *** sacharya has quit IRC | 22:20 | |
| openstackgerrit | Kris Lindgren proposed openstack/ironic-python-agent: Changes to allow ipmi to work under the CoreOS IPA image https://review.openstack.org/316306 | 22:24 |
| *** fragatina has joined #openstack-ironic | 22:26 | |
| *** fragatina has quit IRC | 22:26 | |
| *** fragatina has joined #openstack-ironic | 22:27 | |
| *** cdearborn has quit IRC | 22:27 | |
| *** alex_xu has quit IRC | 22:30 | |
| *** lintan_ has quit IRC | 22:31 | |
| *** daemontool has quit IRC | 22:31 | |
| *** lintan_ has joined #openstack-ironic | 22:31 | |
| *** alex_xu has joined #openstack-ironic | 22:33 | |
| *** alexpilotti has quit IRC | 22:35 | |
| *** alexpilotti has joined #openstack-ironic | 22:38 | |
| *** keedya has quit IRC | 22:41 | |
| *** catintheroof has quit IRC | 22:49 | |
| *** rbudden has quit IRC | 22:51 | |
| openstackgerrit | Kris Lindgren proposed openstack/ironic-python-agent: Make ipmi to work under the CoreOS IPA image https://review.openstack.org/316306 | 22:54 |
| *** ppiela has quit IRC | 23:00 | |
| *** alexpilotti has quit IRC | 23:01 | |
| *** alexpilotti has joined #openstack-ironic | 23:02 | |
| *** lucas-afk has quit IRC | 23:05 | |
| *** lucasagomes has joined #openstack-ironic | 23:12 | |
| *** sacharya has joined #openstack-ironic | 23:20 | |
| *** baoli_ has quit IRC | 23:23 | |
| *** ekarlso has quit IRC | 23:23 | |
| *** sacharya has quit IRC | 23:25 | |
| *** absubram has quit IRC | 23:27 | |
| *** ekarlso has joined #openstack-ironic | 23:32 | |
| *** mtanino has quit IRC | 23:41 | |
| *** mbound has quit IRC | 23:49 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
