| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Update Neutron if mac address of the port changed https://review.openstack.org/77939 | 00:02 |
|---|---|---|
| *** Hefeweizen has joined #openstack-ironic | 00:02 | |
| lifeless | NobodyCam: agrgghhgh | 00:02 |
| devananda | ok, I need to step away for a few hours... will try to be back late tonight to do a few more reviews | 00:02 |
| NobodyCam | lifeless: ahh looks like we need to add compute_manager to nova.conf in the nova elelmet too :) | 00:02 |
| lifeless | NobodyCam: yes I have a patch for that | 00:02 |
| NobodyCam | :) | 00:03 |
| openstackgerrit | lifeless proposed a change to openstack/ironic: Provide a new ComputeManager for Ironic https://review.openstack.org/82637 | 00:03 |
| lifeless | I was just testing it but it build nova-bm for some reason | 00:03 |
| lifeless | let me push up | 00:03 |
| openstackgerrit | Jay Faulkner proposed a change to openstack/ironic: Set good defaults for heartbeat interval & timeout https://review.openstack.org/82615 | 00:04 |
| NobodyCam | lifeless: are building with DIB_REPOLOCATION_nova or _ironic to test these? | 00:04 |
| NobodyCam | im using ironic for the compute_manager paatch | 00:05 |
| jroll | hey, is there a documented way to mock out the waiting done in looping calls? | 00:07 |
| jroll | I'd like to speed up these 8 second seamicro tests :) | 00:08 |
| jroll | going to see what mocking out eventlet.greenthread.sleep does for me... | 00:10 |
| *** epim has quit IRC | 00:10 | |
| lifeless | NobodyCam: DIB_REPOLOCATION_ironic=https://review.openstack.org/openstack/ironic | 00:14 |
| lifeless | DIB_REPOREF_ironic=refs/changes/37/82637/3 | 00:14 |
| NobodyCam | :) | 00:15 |
| *** openstack has joined #openstack-ironic | 00:19 | |
| lifeless | perhaps /4 now :) | 00:21 |
| NobodyCam | :) | 00:21 |
| lifeless | yeah | 00:21 |
| jroll | NobodyCam: I've got a couple fixes to speed tests up - should those go in separate reviews or is one ok? | 00:22 |
| jroll | it's 3 changes, 4 lines total | 00:22 |
| jroll | but those 3 are totally unrelated | 00:22 |
| lifeless | jroll: 3 changes :) | 00:22 |
| NobodyCam | jroll: they are all tests? | 00:22 |
| lifeless | jroll: GitCommitMessages is fairly clear on this | 00:23 |
| jroll | NobodyCam: yes, all in tests | 00:23 |
| jroll | I'll put up three reviews | 00:24 |
| NobodyCam | lifeless: is correct, but four lines, I would accept a single "Speeed up ironic tests." review | 00:24 |
| NobodyCam | :-p | 00:24 |
| jroll | ok, one review it is :) | 00:25 |
| jroll | ~30s -> ~12.5s | 00:25 |
| * NobodyCam notes he ended his example with a . which would not pass pep8 | 00:25 | |
| jroll | per environment | 00:25 |
| jroll | heh, yeah | 00:25 |
| jroll | oops, it's -2 +5, I underestimated :P | 00:26 |
| jroll | I'll put it up and if someone prefers, I'll break it out | 00:26 |
| *** matsuhashi has joined #openstack-ironic | 00:26 | |
| NobodyCam | :) jroll nice :) oh no over the limit 3 reviews | 00:26 |
| NobodyCam | lol | 00:26 |
| jroll | lol | 00:26 |
| NobodyCam | hehehe j/k | 00:26 |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic: Speed up ironic tests https://review.openstack.org/82680 | 00:30 |
| jroll | NobodyCam: ^ | 00:30 |
| jroll | and on that note, I'm off. have a good night all :) | 00:30 |
| openstackgerrit | A change was merged to openstack/ironic: Pass no arguments to _wait_for_provision_state() https://review.openstack.org/82669 | 00:46 |
| *** krtaylor has quit IRC | 01:07 | |
| *** krtaylor has joined #openstack-ironic | 01:11 | |
| *** krtaylor has quit IRC | 01:20 | |
| *** krtaylor has joined #openstack-ironic | 01:23 | |
| *** rwsu has quit IRC | 01:26 | |
| *** nosnos has joined #openstack-ironic | 01:29 | |
| *** matsuhashi has quit IRC | 01:30 | |
| *** matsuhashi has joined #openstack-ironic | 01:30 | |
| *** rwsu has joined #openstack-ironic | 01:42 | |
| openstackgerrit | lifeless proposed a change to openstack/ironic: Provide a new ComputeManager for Ironic https://review.openstack.org/82637 | 01:44 |
| *** yonglihe_ has quit IRC | 01:56 | |
| *** yonglihe_ has joined #openstack-ironic | 01:59 | |
| *** rloo has quit IRC | 02:26 | |
| *** rloo has joined #openstack-ironic | 02:27 | |
| *** rloo has quit IRC | 02:28 | |
| *** rloo has joined #openstack-ironic | 02:29 | |
| lifeless | dtantsur: I've seen HTTPInternalServerError' object has no attribute '__name__ now | 02:43 |
| *** rloo has quit IRC | 02:51 | |
| openstackgerrit | David Shrewsbury proposed a change to openstack/ironic: Fix missed exception raise in _add_driver_fields https://review.openstack.org/82701 | 02:57 |
| openstackgerrit | lifeless proposed a change to openstack/ironic: Provide a new ComputeManager for Ironic https://review.openstack.org/82637 | 03:00 |
| *** matsuhashi has quit IRC | 03:12 | |
| *** nosnos has quit IRC | 03:29 | |
| *** matsuhashi has joined #openstack-ironic | 03:30 | |
| *** vkozhukalov has joined #openstack-ironic | 03:35 | |
| *** matsuhashi has quit IRC | 03:36 | |
| *** harlowja_ is now known as harlowja_away | 03:49 | |
| *** killer_prince is now known as lazy_prince | 04:06 | |
| *** matsuhashi has joined #openstack-ironic | 04:23 | |
| *** nosnos has joined #openstack-ironic | 04:25 | |
| *** lazy_prince2 has joined #openstack-ironic | 04:41 | |
| *** lazy_prince2 is now known as killer_prince | 04:41 | |
| *** rameshg87 has joined #openstack-ironic | 04:47 | |
| *** eghobo has joined #openstack-ironic | 05:01 | |
| *** tzumainn has quit IRC | 05:03 | |
| *** vkozhukalov has quit IRC | 05:05 | |
| *** matsuhashi has quit IRC | 05:14 | |
| *** vkozhukalov has joined #openstack-ironic | 05:15 | |
| *** matsuhashi has joined #openstack-ironic | 05:18 | |
| *** Manishanker has joined #openstack-ironic | 05:26 | |
| *** rameshg87 has quit IRC | 05:28 | |
| *** vkozhukalov has quit IRC | 05:39 | |
| *** pradipta_away is now known as pradipta | 06:00 | |
| *** mrda is now known as mrda_away | 06:08 | |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/82720 | 06:09 |
| *** eghobo has quit IRC | 06:46 | |
| *** lsmola_ has quit IRC | 07:02 | |
| *** lsmola_ has joined #openstack-ironic | 07:03 | |
| *** vkozhukalov has joined #openstack-ironic | 07:03 | |
| *** eghobo has joined #openstack-ironic | 07:05 | |
| vkozhukalov | morning, all | 07:06 |
| *** rameshg87 has joined #openstack-ironic | 07:08 | |
| *** romcheg has joined #openstack-ironic | 07:13 | |
| *** romcheg has quit IRC | 07:15 | |
| *** romcheg has joined #openstack-ironic | 07:17 | |
| *** saju_m has joined #openstack-ironic | 07:30 | |
| *** romcheg has quit IRC | 07:41 | |
| *** max_lobur has joined #openstack-ironic | 07:42 | |
| *** romcheg has joined #openstack-ironic | 07:44 | |
| *** ifarkas has joined #openstack-ironic | 07:48 | |
| *** eghobo has quit IRC | 07:54 | |
| *** romcheg has quit IRC | 07:57 | |
| *** killer_prince has quit IRC | 08:30 | |
| dtantsur | Morning Ironic | 08:36 |
| *** killer_prince has joined #openstack-ironic | 08:37 | |
| vkozhukalov | dtantsur: morning | 08:40 |
| openstackgerrit | Yuriy Zveryanskyy proposed a change to openstack/ironic: Remove logging of exceptions from controller's methods https://review.openstack.org/81267 | 08:47 |
| *** matsuhashi has quit IRC | 08:49 | |
| *** jistr has joined #openstack-ironic | 08:49 | |
| *** saju_m has quit IRC | 08:50 | |
| *** athomas has joined #openstack-ironic | 08:50 | |
| *** saju_m has joined #openstack-ironic | 08:52 | |
| *** jistr is now known as jistr|training | 08:54 | |
| openstackgerrit | Yuriy Zveryanskyy proposed a change to openstack/ironic: Decrease running time of SeaMicro driver tests https://review.openstack.org/81763 | 08:55 |
| *** matsuhashi has joined #openstack-ironic | 08:56 | |
| *** matsuhashi has quit IRC | 09:14 | |
| *** derekh has joined #openstack-ironic | 09:15 | |
| *** lucas-dinner is now known as lucasagomes | 09:17 | |
| *** matsuhashi has joined #openstack-ironic | 09:18 | |
| openstackgerrit | Yuriy Zveryanskyy proposed a change to openstack/ironic: Update Neutron if mac address of the port changed https://review.openstack.org/77939 | 09:21 |
| *** pradipta is now known as pradipta_away | 09:24 | |
| *** pradipta_away is now known as pradipta | 09:24 | |
| *** tzumainn has joined #openstack-ironic | 09:30 | |
| dtantsur | agordeev, hi! After "Remove any lib/baremetal usage for ironic deploy" do we need to update localrc, i.e. BM_BUILD_DEPLOY_RAMDISK=True and BM_DEPLOY_FLAVOR=...? | 09:32 |
| *** saju_m has quit IRC | 09:38 | |
| *** nosnos has quit IRC | 09:39 | |
| *** killer_prince has quit IRC | 09:49 | |
| agordeev | dtantsur: hi, only if you need non-default setting. Otherwise there's no reason for updating them. | 09:55 |
| *** tatyana has joined #openstack-ironic | 09:55 | |
| dtantsur | agordeev, well, currently I have them as stated here: http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html (except for s/ubuntu/fedora/) | 09:55 |
| *** saju_m has joined #openstack-ironic | 09:56 | |
| *** max_lobur has quit IRC | 10:00 | |
| agordeev | dtantsur: there's no need of explisit IRONIC_VM_SPECS_RAM=512 (since it's default value). BM_BUILD_DEPLOY_RAMDISK, BM_DEPLOY_FLAVOR were renamed to IRONIC_BUILD_DEPLOY_RAMDISK (defaults to True) and IRONIC_DEPLOY_FLAVOR (defaults to "ubuntu deploy-ironic") respectively. | 10:01 |
| dtantsur | agordeev, thanks. I think quickstart should be updated as well.. | 10:02 |
| agordeev | dtantsur: also, "-a amd64" is the default value too, so it was ommited as well | 10:02 |
| dtantsur | agordeev, what about IRONIC_VM_SPECS_DISK=10 ? | 10:03 |
| agordeev | dtantsur: yup, at least it should reference to actual environment variable, not old BM_* | 10:04 |
| agordeev | dtantsur: yeah, it defaults to 10 | 10:05 |
| *** matsuhashi has quit IRC | 10:08 | |
| *** matsuhashi has joined #openstack-ironic | 10:10 | |
| *** killer_prince has joined #openstack-ironic | 10:18 | |
| *** killer_prince has quit IRC | 10:29 | |
| *** pradipta is now known as pradipta_away | 10:32 | |
| *** athomas has quit IRC | 10:36 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Supports filtering port by address https://review.openstack.org/82773 | 10:42 |
| *** romcheg has joined #openstack-ironic | 10:44 | |
| *** romcheg has left #openstack-ironic | 10:50 | |
| *** killer_prince has joined #openstack-ironic | 10:50 | |
| *** max_lobur has joined #openstack-ironic | 10:52 | |
| *** max_lobur1 has joined #openstack-ironic | 10:53 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Replace fixtures with mock in test_keystone.py https://review.openstack.org/79702 | 10:57 |
| *** max_lobur has quit IRC | 10:57 | |
| *** saju_m has quit IRC | 11:07 | |
| *** saju_m has joined #openstack-ironic | 11:08 | |
| *** saju_m has quit IRC | 11:09 | |
| *** saju_m has joined #openstack-ironic | 11:15 | |
| *** athomas has joined #openstack-ironic | 11:22 | |
| *** athomas_ has joined #openstack-ironic | 11:26 | |
| *** athomas has quit IRC | 11:29 | |
| *** rameshg87 has left #openstack-ironic | 11:36 | |
| *** saju_m has quit IRC | 11:38 | |
| *** saju_m has joined #openstack-ironic | 11:38 | |
| *** jistr|training has quit IRC | 11:40 | |
| *** athomas has joined #openstack-ironic | 11:40 | |
| *** athomas_ has quit IRC | 11:41 | |
| *** jistr has joined #openstack-ironic | 11:41 | |
| *** jistr is now known as jistr|training | 11:41 | |
| *** ifarkas has quit IRC | 11:43 | |
| *** ifarkas_ has joined #openstack-ironic | 11:48 | |
| *** ifarkas_ has quit IRC | 11:50 | |
| *** ifarkas has joined #openstack-ironic | 11:51 | |
| *** harlowja_away has quit IRC | 12:03 | |
| *** matsuhashi has quit IRC | 12:16 | |
| *** linggao has joined #openstack-ironic | 12:20 | |
| *** matsuhashi has joined #openstack-ironic | 12:23 | |
| *** romcheg has joined #openstack-ironic | 12:30 | |
| *** zul has quit IRC | 12:38 | |
| *** zul has joined #openstack-ironic | 12:41 | |
| *** saju_m has quit IRC | 12:41 | |
| *** rloo has joined #openstack-ironic | 12:50 | |
| *** lucasagomes is now known as lucas-hungry | 12:51 | |
| *** saju_m has joined #openstack-ironic | 12:55 | |
| *** saju_m has quit IRC | 12:59 | |
| *** saju_m has joined #openstack-ironic | 13:00 | |
| *** matsuhashi has quit IRC | 13:03 | |
| *** saju_m has quit IRC | 13:04 | |
| *** saju_m has joined #openstack-ironic | 13:05 | |
| *** jbjohnso_ has joined #openstack-ironic | 13:05 | |
| *** saju_m has quit IRC | 13:09 | |
| *** saju_m has joined #openstack-ironic | 13:10 | |
| *** todd_dsm has joined #openstack-ironic | 13:20 | |
| *** ChanServ changes topic to "the gerrit event stream is currently hung, blocking all testing. troubleshooting is in progress (next update at 14:00 utc)" | 13:21 | |
| *** lazy_prince has quit IRC | 13:22 | |
| *** lazy_prince has joined #openstack-ironic | 13:23 | |
| *** matty_dubs|gone is now known as matty_dubs | 13:24 | |
| *** ChanServ changes topic to "OpenStack Bare Metal Provisioning | Docs: http://docs.openstack.org/developer/ironic/ | Bugs: https://bugs.launchpad.net/ironic | Status: https://etherpad.openstack.org/p/IronicWhiteBoard" | 13:29 | |
| *** lucas-hungry is now known as lucasagomes | 13:33 | |
| *** killer_prince has quit IRC | 13:35 | |
| *** lazy_prince is now known as killer_prince | 13:35 | |
| NobodyCam | good morning Ironic | 13:58 |
| agordeev | NobodyCam: good morning! | 13:59 |
| mdurnosvistov | Good morning/afternoon NobodyCam and all! :) | 14:00 |
| lucasagomes | morning all :) | 14:01 |
| lucasagomes | NobodyCam, http://lists.openstack.org/pipermail/openstack-dev/2014-March/030964.html | 14:01 |
| NobodyCam | morning agordeev mdurnosvistov lucasagomes | 14:03 |
| *** jbjohnso_ has quit IRC | 14:05 | |
| NobodyCam | brb morning walkies | 14:06 |
| yuriyz | morning all | 14:09 |
| lucasagomes | morning yuriyz | 14:10 |
| NobodyCam | morning yuriyz | 14:10 |
| dtantsur | Hi Ironic. Getting to yesterdays problems: has anyone tried devstack + ironic patch on IPv6-enabled box (ubuntu, fedora, whatever)? | 14:16 |
| dtantsur | I'm trying to figure out the scope of problems I'm getting | 14:16 |
| *** jbjohnso_ has joined #openstack-ironic | 14:20 | |
| jroll | morning all :) | 14:23 |
| NobodyCam | morning jroll | 14:24 |
| NobodyCam | dtantsur: I have been running dib built images which i beleieve have not turned off ip v6, but need to dbl chck that | 14:25 |
| jroll | this looks like a "recheck no bug" to me, yeah? https://review.openstack.org/#/c/82680/ | 14:27 |
| jroll | down to only 8 rc1 bugs left :) | 14:32 |
| romcheg | Morning NobodyCam, lucasagomes | 14:32 |
| lucasagomes | morning romcheg :) | 14:33 |
| romcheg | and jroll | 14:33 |
| NobodyCam | morning romcheg | 14:33 |
| jroll | hey romcheg | 14:33 |
| NobodyCam | jroll: awesome | 14:33 |
| romcheg | The review session is planned to 3PM GMT isn't it? | 14:33 |
| NobodyCam | 8:00 pst... | 14:33 |
| NobodyCam | i think thats 1500 gmt | 14:34 |
| jroll | yep | 14:34 |
| romcheg | time zones are pain in the ass... | 14:34 |
| NobodyCam | romcheg: http://www.worldtimebuddy.com | 14:35 |
| *** todd_dsm has quit IRC | 14:37 | |
| Shrews | NobodyCam: ooh, i like that | 14:37 |
| *** todd_dsm has joined #openstack-ironic | 14:37 | |
| romcheg | wow, cool | 14:37 |
| NobodyCam | Shrews: lol its the only way I can keep the time zone straght with the daylight saving always moving around | 14:38 |
| NobodyCam | :-p | 14:38 |
| NobodyCam | open floor question: would it be bad to hyjack the horizon apache server for ipxe use? could it done, or nor even worth looking in to? | 14:40 |
| *** todd_dsm has quit IRC | 14:40 | |
| romcheg | NobodyCam: i oppose that | 14:40 |
| romcheg | A lot of big customers as I know do not use horizon | 14:41 |
| romcheg | They use their own web UI | 14:41 |
| jroll | vkozhukalov / agordeev : did you two see my wiki page updates? https://wiki.openstack.org/wiki/Ironic-python-agent | 14:41 |
| jroll | NobodyCam: -1 on that also. I would almost rather just document "set up an http server for ipxe use" | 14:42 |
| jroll | and if needed "image needs to be at $SOME_PATH" | 14:42 |
| NobodyCam | ok... thats why I asked :) | 14:42 |
| romcheg | Even is there's no custom ui that replaces horizon, sticking to it is not the best way to go | 14:43 |
| NobodyCam | how would we get a image on to the http server? | 14:43 |
| NobodyCam | I agree it would have been a total hack | 14:43 |
| jroll | NobodyCam: our current method is wget from the build machine :P | 14:43 |
| NobodyCam | ya so the conductor will require a http server to for ipxe? | 14:44 |
| NobodyCam | we pull the image from glance | 14:45 |
| romcheg | Maybe torrent tracker? | 14:45 |
| NobodyCam | if the http server is not on the conductor how will the image get to a location it can be served from | 14:45 |
| NobodyCam | the conductor gets the image id to pull from glance | 14:46 |
| devananda | morning, all | 14:46 |
| NobodyCam | good morning devananda :) | 14:46 |
| jroll | morning devananda | 14:46 |
| romcheg | Morning devananda! | 14:46 |
| romcheg | NobodyCam: there are two ways: we can either push the image to the web server (over scp for instance) | 14:47 |
| jroll | NobodyCam: in our environment, we just have our DHCP server also running apache, serving the agent image | 14:47 |
| romcheg | NobodyCam: or we can give that node a command to put it on it's own | 14:48 |
| jroll | is the deploy ramdisk that ironic currently uses static? | 14:48 |
| devananda | jroll: yes | 14:48 |
| jroll | right - so it doesn't necessarily need to be served from glance | 14:48 |
| devananda | jroll: it's configurable per-node. but in practice, it's static | 14:48 |
| jroll | I understand that's optimal, but a standalone apache server would scale better than "download image to conductor and serve from there" | 14:49 |
| devananda | jroll: the image is cached and symlinked on conductor(s) | 14:49 |
| devananda | jroll: and it's very small today (~30MB) so it is almost guaranteed to be served from FS cache | 14:49 |
| agordeev | jroll: yup. Well done, great job on the wiki updates! | 14:49 |
| devananda | jroll: from a machine in some reasonable proximity to the node requesting it (rather than a potentially distant centralized apache service) | 14:50 |
| NobodyCam | devananda: I started this as I was asked about ipxe serving images over http | 14:50 |
| devananda | jroll: and for concurrent deploys, it will scale as much as the conductors do. no need to scale something else (apache) too | 14:50 |
| devananda | NobodyCam: never admit when you started it :p | 14:51 |
| jroll | devananda: fair | 14:52 |
| jroll | agordeev: :) | 14:52 |
| devananda | jroll: i'm fairly resistant to add an external operational requirement to ironic when there isn't a really compeling reason to do so | 14:52 |
| lucasagomes | morning devananda | 14:52 |
| NobodyCam | lol | 14:52 |
| jroll | devananda: I know :) | 14:53 |
| jroll | just making a suggestion based on what we're doing today for the much larger agent image | 14:53 |
| devananda | jroll: *nod* | 14:54 |
| devananda | jroll: fwiw, chaining to ipxe and pulling the larger image over HTTP(S) seems totally reasonable to me | 14:54 |
| JoshNang | morning everyone! | 14:54 |
| NobodyCam | morning JoshNang | 14:55 |
| *** Manishanker has left #openstack-ironic | 14:55 | |
| Shrews | yuriyz: vote changed! thanks for the explanation :) | 14:55 |
| *** jbjohnso_ has quit IRC | 14:57 | |
| devananda | I have several emails I should respond to ... and i'm still working on first cup'o'coffee... | 14:58 |
| Shrews | devananda: coffee has higher priority | 14:59 |
| devananda | lucasagomes: romcheg: max_lobur1: ya'll up for starting soon? | 14:59 |
| NobodyCam | I have to run to doctors this morning ... so I will miss the jam session | 14:59 |
| romcheg | devananda: +1 | 14:59 |
| NobodyCam | I will be back around 10 ish | 14:59 |
| lucasagomes | devananda, yup | 15:00 |
| jroll | I'm ready whenever, currently mucking around with sound cards in debian :| | 15:00 |
| devananda | jroll: PM me your google hangout info | 15:01 |
| JoshNang | devananda: can I join in as well? | 15:01 |
| jroll | oh, hangouts | 15:01 |
| * jroll gets his laptop | 15:01 | |
| russell_h | morning | 15:02 |
| devananda | jroll: or not, up to you | 15:02 |
| jroll | nah, that's fine | 15:02 |
| jroll | morning russell_h | 15:03 |
| lucasagomes | yeah, so if we start serving larger images via http (e.g ramdisk agent) we would need the http server, but idk if Ironic should do something about it, like starting/stopping the http server sounds like somehting out of our plate | 15:03 |
| devananda | https://plus.google.com/hangouts/_/7ecpiv2qebg86vb37ils8gcrb8 | 15:04 |
| devananda | https://etherpad.openstack.org/p/IronicReviewDay <-- Scratch Pad | 15:05 |
| *** jbjohnso_ has joined #openstack-ironic | 15:10 | |
| openstackgerrit | Yuriy Zveryanskyy proposed a change to openstack/ironic: Update Neutron if mac address of the port changed https://review.openstack.org/77939 | 15:11 |
| *** saju_m has quit IRC | 15:13 | |
| jroll | Shrews: are you working on https://bugs.launchpad.net/ironic/+bug/1295870 ? | 15:13 |
| Shrews | jroll: i am | 15:14 |
| jroll | cool | 15:14 |
| devananda | https://review.openstack.org/#/c/80400/ vs https://review.openstack.org/82668 | 15:15 |
| pquerna | lucasagomes: re: http server, we've been playing with using swift direct URLs w/ the temp sign key so they don't need additional auth; keeps ironic out of the http server business. | 15:16 |
| lucasagomes | pquerna, heh ack! thanks | 15:17 |
| pquerna | and as a bonus its a decent place to store 300mb images :x | 15:17 |
| pquerna | (and yeah, as devananda mentioned, we chain iPXE first, then have http) | 15:18 |
| devananda | so this bug does not seem to have any progress | 15:20 |
| devananda | https://bugs.launchpad.net/ironic/+bug/1291420 | 15:20 |
| devananda | anyone know the scheduler and want to take it? | 15:21 |
| devananda | right, i'll toss a patch up in a minute | 15:22 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Run ipmi power status less aggressively https://review.openstack.org/82668 | 15:34 |
| *** krtaylor has quit IRC | 15:42 | |
| mdurnosvistov | devananda: I can try to deal with this bug https://bugs.launchpad.net/ironic/+bug/1291420 | 15:42 |
| rloo | devananda: wrt https://review.openstack.org/#/c/82668/. don't know if you saw or not, but yuriy had a comment. | 15:43 |
| devananda | mdurnosvistov: I just replied on the bug -- if you want to work on it, great! | 15:45 |
| devananda | I thought it would be simpler than it actually will be | 15:45 |
| *** krtaylor has joined #openstack-ironic | 15:50 | |
| mdurnosvistov | devananda: If I have any questions I ask you or someone else :) | 15:50 |
| *** eghobo has joined #openstack-ironic | 15:52 | |
| *** eghobo has quit IRC | 15:52 | |
| *** eghobo has joined #openstack-ironic | 15:53 | |
| *** romcheg1 has joined #openstack-ironic | 15:57 | |
| *** killer_prince is now known as lazy_prince | 15:57 | |
| *** romcheg has quit IRC | 15:59 | |
| *** matty_dubs is now known as matty_dubs|lunch | 16:02 | |
| devananda | adam_g: around? | 16:04 |
| notq | we're going to be working on deploying an openstack ironic cloud, going to start with deploying it with devstack to become more familiar with it. seems like a good place to start | 16:05 |
| devananda | notq: ++ | 16:05 |
| notq | i have some emails and cordination to do with monty taylor | 16:05 |
| devananda | notq: see PM | 16:05 |
| *** hemna_ has joined #openstack-ironic | 16:10 | |
| *** max_lobur1 has quit IRC | 16:11 | |
| devananda | notq: monty == mordred | 16:12 |
| notq | thanks | 16:13 |
| devananda | notq: also https://etherpad.openstack.org/p/IronicAndDevstackAgain is our current walkthrough for using devstack | 16:13 |
| notq | great, i went to a similar link ot that and it was empty, i'll chcek that one | 16:13 |
| devananda | notq: and lots of info here - http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html#deploying-ironic-with-devstack | 16:13 |
| mordred | what did I do? | 16:13 |
| *** ifarkas has quit IRC | 16:22 | |
| *** vkozhukalov has quit IRC | 16:23 | |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic: Set good defaults for heartbeat interval & timeout https://review.openstack.org/82615 | 16:23 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Run ipmi power status less aggressively https://review.openstack.org/82668 | 16:40 |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic: Avoid API root controller dependency on v1 dir https://review.openstack.org/78912 | 16:41 |
| jroll | g/b 63 | 16:42 |
| jroll | oops | 16:42 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Run ipmi power status less aggressively https://review.openstack.org/82668 | 16:46 |
| devananda | https://review.openstack.org/#/c/81524 -- jroll, Shrews: either of you test that this actually loads those opts from the CONF file now? | 16:56 |
| *** jistr|training has quit IRC | 16:58 | |
| rloo | devananda: ^^ wrt 81524, i tested it. | 17:00 |
| *** matty_dubs|lunch is now known as matty_dubs | 17:02 | |
| *** ifarkas has joined #openstack-ironic | 17:03 | |
| *** harlowja has joined #openstack-ironic | 17:03 | |
| *** romcheg has joined #openstack-ironic | 17:05 | |
| Shrews | devananda: yes, though the periodic timer confuses me a bit, but that's my issue | 17:07 |
| russell_h | Shrews: I think JayF has better fix for this that he's working on right now | 17:07 |
| *** romcheg1 has quit IRC | 17:08 | |
| *** romcheg has quit IRC | 17:10 | |
| openstackgerrit | Jay Faulkner proposed a change to openstack/ironic: Run periodic_task in a with a dynamic timer https://review.openstack.org/82615 | 17:10 |
| JayF | devananda: https://review.openstack.org/82615 is the new fix to that heartbeat_timeout bug. I found the root cause of the bug, and it was in ironic common instead of oslo. See what you think. | 17:11 |
| JayF | JoshNang: ^ also relevant | 17:11 |
| devananda | JayF: nice | 17:13 |
| devananda | JayF: i'm looking at syncing the periodic_task.py code from oslo-incubator right now | 17:13 |
| JayF | Yeah this is actually a really nasty bug, I'm kinda surprised it hasn't broken anyone's deployments. There's no way it would work reliably with that bug + the default configuration | 17:14 |
| JayF | the intermittant failure that was documented for tests would almost certainly happen in a production environment occassionally as well, especially in periods of high load | 17:14 |
| jroll | JayF: it didn't break anyone's prod environment because nobody has a prod environment :) | 17:15 |
| JayF | /yet/. | 17:16 |
| jroll | right | 17:16 |
| jroll | I'm not saying it would not have broken it | 17:16 |
| devananda | yea, i think it would have definitely broken in prod | 17:16 |
| devananda | looking at syncing olso-incubator periodic_task.py, but it's changed more than I want to sync right now | 17:17 |
| devananda | and it sounds like with JayF 's fix, it's not needed to sync to fix that bug now | 17:17 |
| *** ifarkas_ has joined #openstack-ironic | 17:20 | |
| openstackgerrit | A change was merged to openstack/ironic: Update Neutron if mac address of the port changed https://review.openstack.org/77939 | 17:25 |
| *** dwalleck has joined #openstack-ironic | 17:35 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Run ipmi power status less aggressively https://review.openstack.org/82668 | 17:37 |
| openstackgerrit | A change was merged to openstack/ironic: Updated from global requirements https://review.openstack.org/79334 | 17:45 |
| *** epim has joined #openstack-ironic | 17:48 | |
| *** athomas has quit IRC | 17:49 | |
| *** tzumainn has quit IRC | 17:52 | |
| devananda | lucasagomes: quick one for you: https://review.openstack.org/#/c/81763/ | 17:53 |
| lucasagomes | devananda, will take a look | 17:53 |
| lucasagomes | devananda, btw | 17:53 |
| jbjohnso_ | devananda, fyi, I'm currently cleaning up my console server for apache2 release | 17:53 |
| adam_g | devananda, hey | 17:54 |
| jbjohnso_ | well, publishing as apache2, it's still under development (needs a few corners addressed) | 17:55 |
| *** derekh has quit IRC | 17:55 | |
| lucasagomes | jbjohnso_, https://review.openstack.org/#/c/82146/ | 17:56 |
| jbjohnso_ | lucasagomes, hmm, need to look at that on pyghmi side.. | 17:57 |
| lucasagomes | jbjohnso_, looks like a bug in pyghmi right? | 17:57 |
| lucasagomes | jbjohnso_, btw, forgot to show u: $yum search pyghmi... python-pyghmi.noarch : Python General Hardware Management Initiative (IPMI and others) :) | 17:58 |
| jbjohnso_ | lucasagomes, so interesting, the root issue there is that HMAC | 17:58 |
| openstackgerrit | A change was merged to openstack/ironic: Fix 'spacing' parameters for periodic tasks https://review.openstack.org/81524 | 17:59 |
| jbjohnso_ | so pycrypto hmac is passed that value directly.. | 17:59 |
| jbjohnso_ | currently pyghmi does not process it.. | 17:59 |
| openstackgerrit | Ghe Rivero proposed a change to openstack/ironic: Update SeaMicro to use MixinVendorInterface https://review.openstack.org/79903 | 18:00 |
| lucasagomes | jbjohnso_, right /me go look into the code | 18:00 |
| jbjohnso_ | looking up the HMAC documentation | 18:00 |
| jbjohnso_ | key (byte string) | 18:02 |
| jbjohnso_ | so pycrypto is explicit that it should be a byte array... | 18:02 |
| *** eguz has joined #openstack-ironic | 18:02 | |
| jbjohnso_ | hmm.. this is non trivial to address in a perfect way... | 18:05 |
| lucasagomes | jbjohnso_, ah eww | 18:05 |
| jbjohnso_ | I'd not want the limitation tehy have, for example, where only ASCII is allowed... | 18:05 |
| lucasagomes | +1 | 18:05 |
| lucasagomes | things will get yet more hairy with py3 | 18:05 |
| lucasagomes | where the default is unicode | 18:05 |
| *** eghobo has quit IRC | 18:05 | |
| jbjohnso_ | actually, less perhaps | 18:06 |
| jbjohnso_ | with py2, unicode strings are like aliens | 18:06 |
| lucasagomes | heh | 18:07 |
| devananda | JayF: hi! Im' looking at https://review.openstack.org/#/c/82615/5/ironic/common/service.py | 18:07 |
| jbjohnso_ | I have to ponder.... | 18:07 |
| devananda | JayF: it seems to me we dont need to change default periodic_interval at all | 18:07 |
| devananda | JayF: using a dynamic timer (good catch!) it's actually doing what it should -- sleeping only until the next timer needs t orun,r ather than 60sec | 18:08 |
| JayF | devananda: I believe we do, because otherwise the first run of a heartbeat will have to wait 60s before the timer is updated | 18:08 |
| devananda | nope | 18:08 |
| JayF | devananda: i.e. tick just occurred, 59s left in the timer before next run, heartbeat is registered, it'll have to wait 59s before run_periodic_task is run again, and when that's run again, the next interval will be 10s | 18:09 |
| devananda | JayF: so, i rebased on master. antoher patch addressed a similar issue. try it | 18:09 |
| JayF | devananda: then what triggers the initial run_periodic_Task? | 18:09 |
| devananda | no, it's settign the timer to min(periodic_interval, time-to-next-job) | 18:09 |
| devananda | when I start the service, I see this: | 18:09 |
| devananda | 2014-03-25 11:08:40.472 8468 DEBUG ironic.openstack.common.loopingcall [-] Dynamic looping call sleeping for 14.77 seconds _inner /opt/source/ironic/ironic/openstack/common/loopingcall.py:132 | 18:09 |
| devananda | so, i've set 474 heartbeat_interval=15 | 18:10 |
| devananda | in my config file | 18:10 |
| JayF | hmm | 18:10 |
| JayF | are all the periodic tasks being registered before the periodic_task runs then? | 18:10 |
| *** vkozhukalov has joined #openstack-ironic | 18:10 | |
| JayF | er, before run_periodic_task is called | 18:10 |
| devananda | and since 93b042c0be landed, it's honoured now | 18:10 |
| devananda | JayF: i think they are _now_ :( | 18:11 |
| devananda | * :) | 18:11 |
| devananda | JayF: your patch was probably correct prior to 93b042c0be landing | 18:11 |
| jbjohnso_ | so unicode to byte array.. picking an encoding... | 18:11 |
| JayF | I can't find 93b042c0be? | 18:11 |
| jbjohnso_ | perhaps if it is unicode, I'll just do UTF-8 | 18:12 |
| devananda | JayF: https://review.openstack.org/#/c/81524/ | 18:12 |
| devananda | merged 13 min ago | 18:12 |
| devananda | hehe | 18:12 |
| lucasagomes | jbjohnso_, can't u use hmac from python lib? | 18:12 |
| devananda | JayF: if I set periodic_interval really low, like 1s, it's unnecessarily chatty -- but it works. so I think we can safely leave teh default periodic_interval at 60s, and if someone chooses to set any periodic task's config'd interval less than that, it works now | 18:13 |
| JayF | Okay, I did see that was posted for review but don't understand how it related? | 18:13 |
| devananda | it changed the order in which options are loaded for the periodic_task decorators | 18:13 |
| devananda | CONF is read before those modules are imported now, so they are created with the right values | 18:14 |
| jbjohnso_ | lucasagomes, let me see why... I might have needed pycrpto for the crypt, but maybe hmac is different | 18:14 |
| lucasagomes | jbjohnso_, ack | 18:14 |
| JayF | devananda: so the periodic_tasks are all configured and loaded /before/ the first run_periodic_task after that patch? | 18:14 |
| JoshNang | jbjohnso_: unrelated to the current discussion, but is it possible to drop the username/password requirement in pyghmi? it doesn't seem to be a strict requirement in the cli ipmitool | 18:14 |
| jbjohnso_ | JoshNang, guess I could default the value to blank strings... | 18:15 |
| devananda | JayF: afaict, yes | 18:16 |
| JayF | devananda: okay cool, me and russell_h were talking about how to make that work and didn't have an answer. So nice that someone fixed up the other half. I'll remove the default change and the warning in the help right now | 18:17 |
| JoshNang | jbjohnso_: i think it is possible that some devices may not have a password initially | 18:17 |
| devananda | JayF: ty | 18:17 |
| jbjohnso_ | JoshNang, well at a low level, the value if not supplied is a byte string of zero length | 18:17 |
| russell_h | devananda: wait, are you running JayF's patch when that log message is generated, or how are you getting a DynamicLoopingCall? | 18:17 |
| jbjohnso_ | JoshNang, for network at least, KCS is a different beast | 18:18 |
| jbjohnso_ | but I have nothing to really recommend pyghmi for over ipmitool or ipmicmd in the KCS case... | 18:18 |
| devananda | russell_h: yes. I rebased his patch on current master | 18:18 |
| russell_h | ah, gotcha, +1 | 18:18 |
| JoshNang | jbjohnso_: interesting. makes sense for the low level. | 18:18 |
| openstackgerrit | Jay Faulkner proposed a change to openstack/ironic: Run periodic_task in a with a dynamic timer https://review.openstack.org/82615 | 18:19 |
| jbjohnso_ | lucasagomes, looks like I might be able to convert to hmac built in, but still need PyCrypto Cipher at the moment.. but no user/pass data goes direct into Cipher anyhow | 18:20 |
| JayF | devananda: ^ there's the update, should be g2g | 18:20 |
| lucasagomes | jbjohnso_, nice! do you want me to open a bug for that? | 18:20 |
| lucasagomes | I think it's nice to keep track of the problem | 18:20 |
| jbjohnso_ | lucasagomes, of course, I don't know if it will fix anything mind you | 18:20 |
| lucasagomes | jbjohnso_, ack | 18:20 |
| jbjohnso_ | lucasagomes, let me try a unicode string in there real quick | 18:20 |
| lucasagomes | ack | 18:21 |
| jbjohnso_ | lucasagomes, python's built in hmac is no happier with unicode strings | 18:22 |
| lucasagomes | :( | 18:22 |
| jbjohnso_ | lucasagomes, and you are right, it gets more fun in python 3 | 18:23 |
| jbjohnso_ | >>> hmac.new('bob') | 18:23 |
| jbjohnso_ | that fails in python 3 | 18:23 |
| lucasagomes | hah | 18:23 |
| lucasagomes | yeah because the default str type is unicode in py3 | 18:23 |
| jbjohnso_ | >>> hmac.new(b'bob') | 18:24 |
| jbjohnso_ | of course is all happy | 18:24 |
| lucasagomes | jbjohnso_, maybe we should make pyghmi more friendly and try to convert it to byte string | 18:24 |
| lucasagomes | and raise an exception if that fails | 18:24 |
| lucasagomes | other than ask for the user to convert it | 18:24 |
| rloo | JayF: wrt 82615, deva had a comment about not changing the doc string (I think he means the help=) | 18:25 |
| *** shakamunyi has joined #openstack-ironic | 18:25 | |
| *** shakamunyi has quit IRC | 18:25 | |
| *** shakamunyi has joined #openstack-ironic | 18:26 | |
| *** shakamunyi has quit IRC | 18:26 | |
| NobodyCam | did I miss all the fun? | 18:26 |
| rloo | NobodyCam. Yawn. No :-) | 18:26 |
| NobodyCam | ahh :-p | 18:26 |
| NobodyCam | morning rloo | 18:26 |
| jbjohnso_ | NobodyCam, so I'll do... password.encode('utf-8') | 18:26 |
| jbjohnso_ | unless byte string already.. | 18:27 |
| rloo | NobodyCam: should almost be bagel time for you ;) | 18:27 |
| NobodyCam | lol almost... just got back from doctors :-p | 18:27 |
| jbjohnso_ | lucasagomes, I meant | 18:27 |
| lucasagomes | jbjohnso_, right, sounds good to me :) thanks | 18:27 |
| NobodyCam | morning jbjohnso_ :) | 18:28 |
| jbjohnso_ | NobodyCam, morning | 18:28 |
| jbjohnso_ | lucasagomes, cool, works on python 2.6 | 18:29 |
| lucasagomes | jbjohnso_, :D awesome | 18:29 |
| lucasagomes | jbjohnso_, I don't know if u use six, but if you do u can also convert with six.binary_type() | 18:30 |
| devananda | rloo: correct. JayF -- I'd prefer you not change the doc / help string there | 18:31 |
| jbjohnso_ | lucasagomes, yeah, actually, the only problem I can find is that bytearray didn't have it, so I'll just catch the attributeerror | 18:31 |
| lucasagomes | jbjohnso_, right | 18:32 |
| rloo | devananda: is that cuz of getting the strings translated? | 18:34 |
| openstackgerrit | A change was merged to openstack/ironic: Fix missed exception raise in _add_driver_fields https://review.openstack.org/82701 | 18:34 |
| devananda | rloo: in principle, yes | 18:34 |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Tolerate unicode strings as username and password https://review.openstack.org/82892 | 18:35 |
| jbjohnso_ | lucasagomes, ^^ | 18:35 |
| jbjohnso_ | now to actually test it myself | 18:35 |
| *** ifarkas_ has quit IRC | 18:35 | |
| lucasagomes | jbjohnso_, great! | 18:36 |
| lucasagomes | thanks for that | 18:36 |
| devananda | rloo: https://www.transifex.com/projects/p/ironic/ | 18:36 |
| jbjohnso_ | lucasagomes, seems to work, could you have the submitter of the other patch try out the change to see if it makes that change superfluous? | 18:37 |
| openstackgerrit | A change was merged to openstack/ironic: Decrease running time of SeaMicro driver tests https://review.openstack.org/81763 | 18:37 |
| devananda | rloo: as you can see, several languages actually have partial translations already. let's not make their jobs harder :) | 18:37 |
| devananda | JayF: ^ | 18:37 |
| lucasagomes | jbjohnso_, sure will ask him to give it another go with the fix in place | 18:37 |
| devananda | JayF: going to do another iteration of your patch, or shall I? other than help=, I'm ready to approve it | 18:38 |
| rloo | devananda: cool. thx! | 18:39 |
| NobodyCam | brb | 18:44 |
| lucasagomes | aight! I think I'm done for today | 18:47 |
| lucasagomes | have a good night everybody | 18:47 |
| devananda | g'night lucasagomes ! | 18:48 |
| *** lucasagomes is now known as lucas-afk | 18:48 | |
| jbjohnso_ | realizing I'll need to move my launchpad account | 18:49 |
| jbjohnso_ | lenovo me will not have that email address | 18:50 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Run periodic_task in a with a dynamic timer https://review.openstack.org/82615 | 18:50 |
| devananda | jbjohnso_: gmail :) | 18:50 |
| NobodyCam | night lucas-afk | 18:50 |
| *** ifarkas_ has joined #openstack-ironic | 18:53 | |
| openstackgerrit | A change was merged to stackforge/pyghmi: Tolerate unicode strings as username and password https://review.openstack.org/82892 | 18:55 |
| jbjohnso_ | devananda, btw, you interested in seeing the web widget for my console server? sadly the only screen sharing solution I am allowed to use is a java one... the console widgets do not require java of course | 18:55 |
| devananda | jbjohnso_: yes - but not today :) | 18:56 |
| Shrews | devananda, NobodyCam: any idea why, in the nova ironic driver, HTTPBadRequest is specifically ignored in some instances? One example here: https://github.com/openstack/ironic/blob/master/ironic/nova/virt/ironic/driver.py#L289 | 18:56 |
| *** mdurnosvistov_ has joined #openstack-ironic | 18:58 | |
| *** derekh has joined #openstack-ironic | 18:59 | |
| *** jistr has joined #openstack-ironic | 18:59 | |
| adam_g | devananda, missed your ping earlier this AM, sup? | 19:00 |
| devananda | adam_g: it looks like neither nova.virt.ironic nor ironic.conductor.manager send any msg to neutron when deploying an instance | 19:04 |
| devananda | adam_g: i haven't looked deeper yet but wondering if you knew off hand why networking works :) | 19:04 |
| adam_g | uhm | 19:05 |
| adam_g | :) | 19:05 |
| adam_g | one sec | 19:05 |
| devananda | also, these could use a few more reviews // a core to ++ : https://review.openstack.org/#/c/82680/ and https://review.openstack.org/#/c/82668/ and https://review.openstack.org/#/c/81267/ and https://review.openstack.org/#/c/82615/ | 19:06 |
| *** tatyana has quit IRC | 19:07 | |
| openstackgerrit | A change was merged to stackforge/pyghmi: Fix missing delay_xmit argument breaking power wait requests https://review.openstack.org/82569 | 19:07 |
| adam_g | devananda, AIUI, nova is doing the neutron orchestration and just passing the resulting infos via network_info to the driver, which is then patching the ironic port | 19:07 |
| *** max_lobur has joined #openstack-ironic | 19:08 | |
| JayF | devananda: sorry about that, I intended to revert that docstring too and just missed it | 19:08 |
| devananda | JayF: np | 19:09 |
| adam_g | devananda, the pxe driver posts back to neutron to update its dhcp according to port | 19:09 |
| adam_g | (during deploy) | 19:09 |
| devananda | adam_g: ah right. outside of the driver. so when the MAC changes, in theory, do we need to do more than tell neutron about that? | 19:10 |
| devananda | hopeflly not cause we landed it, heh | 19:10 |
| adam_g | devananda, hmm | 19:11 |
| adam_g | devananda, what was landed? | 19:11 |
| devananda | adam_g: https://review.openstack.org/#/c/77939/ | 19:12 |
| adam_g | devananda, i'm not sure what is required to update an the mac addr of neutron port. i know it / OVS can be pretty strict about what traffic is allowed on a namespace | 19:13 |
| adam_g | devananda, hmm. i would have thought neutron integration would be at the driver level? | 19:15 |
| devananda | adam_g: considering that the VIF<->MAC relationship is set up by Nova today (even above nova.virt.ironic) .... | 19:16 |
| *** dwalleck has quit IRC | 19:16 | |
| devananda | ideally this should expose something back to nova, except, nova has no concept of changign the MAC of a VM | 19:17 |
| devananda | adam_g: there should probably be a driver hook for it | 19:17 |
| *** openstackgerrit has quit IRC | 19:18 | |
| devananda | Shrews: hi! how's 1295870 going? | 19:18 |
| *** max_lobur has quit IRC | 19:18 | |
| *** openstackgerrit has joined #openstack-ironic | 19:18 | |
| Shrews | devananda: trying to resolve the various exception handling, thus the previous question | 19:18 |
| Shrews | devananda: but I don't even see HTTPBadRequest used *at all* in the ironicclient code, so I'm dismissing that exception's use | 19:19 |
| Shrews | devananda: testing will be the most difficult part :( | 19:19 |
| adam_g | devananda, i guess i dont fully grok the other driver use cases beyond the pxe_ssh stuff i've been poking at in devstack. in terms of OVS, it only works because the VMs are booting on a bridge that has already been plugged into the correct namespace/VS/neutron net. i dont see how neutron DHCP fits into more real world setups where we're booting real metal | 19:20 |
| devananda | Shrews: yea... | 19:20 |
| devananda | adam_g: k. so there's two integration points for neutron | 19:21 |
| devananda | Shrews: i suspect that exception handling has something to do with: dont fail during _cleanup if someone else already ran it | 19:21 |
| devananda | Shrews: but IMBW | 19:21 |
| devananda | Shrews: copy into nova tree, run nova unit tests? :( | 19:22 |
| openstackgerrit | A change was merged to openstack/ironic: Adds max retry limit to sync_power_state task https://review.openstack.org/77420 | 19:22 |
| devananda | adam_g: 1) nova tells neutron about the node's MAC, associates with tenant network, neutron assigns IP, etc | 19:22 |
| devananda | adam_g: 2) ironic tells neutron about the DHCP BOOTP options | 19:22 |
| Shrews | devananda: that's what I suspected as well, but if that exception is not even thrown... meh, testing will reveal more hopefully | 19:22 |
| devananda | adam_g: what we've got here is a third case, which IIRC came from lifeless originally, where the physical NIC has been changed and we need to inform neutron | 19:23 |
| devananda | adam_g: nova lacks this concept because you dont change the NIC of a VM. so Ironic needed to tell neutron directly | 19:23 |
| adam_g | devananda, right | 19:24 |
| devananda | adam_g: but this is unrelated to whether the specific deploy driver uses DHCP BOOTP or not | 19:24 |
| Shrews | devananda: what's the cutoff date for I-release bug fixes? | 19:24 |
| lifeless | which BTW we had to do about 2 weeks ago to a real server | 19:24 |
| devananda | it should, in theory, trigger neutron's back-end to update that VIF to associate to the new MAC | 19:24 |
| lifeless | we did it by DB surgery | 19:24 |
| devananda | lifeless: so the fix for that landed but we haven't really tested it ... | 19:25 |
| lifeless | I believe SpamapS will confirm (he did it) Was Not Fun | 19:25 |
| devananda | Shrews: RC is as-soon-as-i-say-go, basically | 19:25 |
| devananda | Shrews: but we can wait another week or two if really needed | 19:25 |
| devananda | Shrews: or we can back-port things | 19:25 |
| *** zul has quit IRC | 19:26 | |
| adam_g | i guess i need to be schooled on some more fundamental questions unrelated to the new feature, about DHCP BOOTP + PXE + neutron | 19:27 |
| devananda | Shrews: i didn't really answer your question. I'm aiming for RC1 by friday, at which point we'll open development for Juno. we can do back-ports until the actual release is cut a few weeks later. | 19:27 |
| devananda | lifeless: i saw your msg from ~2am that "it worked". we just landed a bunch of bug fixes today, and another 6 or so are still in flight | 19:28 |
| adam_g | Shrews, i think the client needs an exception to handle that, which can be caught by the driver. see note @ http://git.openstack.org/cgit/openstack/ironic/tree/ironic/nova/virt/ironic/driver.py#n498 i was hitting a traceback /w a HTTPBadRequest from neutron API | 19:28 |
| devananda | lifeless: i'm seeing all the tripleo-ci stuff green on the patches, so i'm assuming taht means these fixes aren't breakign tripleo :) | 19:29 |
| Shrews | devananda: ok, that's what i wanted to know. this is turning out to be a significant enough change that friday may be rushing it | 19:29 |
| devananda | Shrews: that's what I suspected | 19:29 |
| lifeless | devananda: ironic isn't fully tested in tripleo yet | 19:29 |
| adam_g | Shrews, https://bugs.launchpad.net/ironic/+bug/1297063 | 19:29 |
| lifeless | devananda: until we've landed the tie, tht, t-i patches and the ClusterComputeManager in ironic | 19:30 |
| devananda | Shrews: can you subdivide it meaningfully? | 19:30 |
| lifeless | devananda: after *that* you'll get useful results from toci. | 19:30 |
| *** zul has joined #openstack-ironic | 19:30 | |
| devananda | lifeless: think we can do that today? :) | 19:30 |
| lifeless | ask in the meeting in open chat | 19:30 |
| devananda | k | 19:30 |
| Shrews | adam_g: i think you're referencing a different thing than me | 19:31 |
| Shrews | devananda: maybe. let me ponder over it | 19:32 |
| lifeless | Shrews: got my slive battery this morning | 19:33 |
| lifeless | Shrews: will run unplugged and drain them both to get a baseline today | 19:34 |
| Shrews | lifeless: awesome. please let me know how that works | 19:34 |
| *** sphoorti has joined #openstack-ironic | 19:39 | |
| *** sphoorti has left #openstack-ironic | 19:40 | |
| Shrews | devananda: So, one thing we can do is just cherry pick a few methods of IronicDriver to use the new code, and everything else remains the same. Then we can just test the crap out of those couple of code paths rather than the entire driver. | 19:46 |
| Shrews | It will be a mixed style, but slower integration may be prudent here. | 19:46 |
| openstackgerrit | A change was merged to openstack/ironic: Avoid API root controller dependency on v1 dir https://review.openstack.org/78912 | 19:48 |
| openstackgerrit | A change was merged to openstack/ironic: Update SeaMicro to use MixinVendorInterface https://review.openstack.org/79903 | 19:48 |
| openstackgerrit | A change was merged to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/82720 | 19:48 |
| *** shakayumi has joined #openstack-ironic | 19:54 | |
| *** romcheg has joined #openstack-ironic | 19:56 | |
| *** shakamunyi has joined #openstack-ironic | 19:56 | |
| *** romcheg has left #openstack-ironic | 19:57 | |
| *** shakayumi has quit IRC | 19:58 | |
| jbjohnso_ | devananda, once I get console working, my next stop fyi was removing any requirement of needing to know mac address in order to pxe boot a node | 19:59 |
| jbjohnso_ | devananda, also, support applying static ip config that need not require dhcp to give 'the' address, just 'a' address | 19:59 |
| devananda | jbjohnso_: a "default pxe image" would address that. the RS folks are working on it | 20:01 |
| devananda | jroll: ^ | 20:01 |
| devananda | jbjohnso_: again, ditto for not requiring DHCP BOOTP to be updated dynamically. tehre's even a bug alraedy open about it (wishlist'd for the moment) | 20:02 |
| jbjohnso_ | devananda, well, I'd be supporting UUID indicated boot parameters with or without cooperation of a dnsmasq or dhcp (unless we are on the same IP) | 20:03 |
| jbjohnso_ | if on same IP, it would have to set a static config.. | 20:03 |
| *** shakamunyi has quit IRC | 20:03 | |
| jbjohnso_ | a static option to say 'let this guy handle it' | 20:03 |
| jbjohnso_ | but first, console server | 20:04 |
| *** derekh has quit IRC | 20:05 | |
| *** jistr has quit IRC | 20:06 | |
| *** shakamunyi has joined #openstack-ironic | 20:06 | |
| *** vkozhukalov has quit IRC | 20:28 | |
| notq | ran through until a connection error max retries exceeded for nova flavor-create baremetal | 20:28 |
| NobodyCam | notq: devtest or devstack? | 20:29 |
| notq | devstack | 20:30 |
| lifeless | jbjohnso_: BTW - lots of environments *really* don't want wildcard DHCP. | 20:30 |
| lifeless | jbjohnso_: its a huge feature for Ironic that it doesn't boot anything that plugs in | 20:30 |
| lifeless | jbjohnso_: so please don't make wildcard the default. | 20:30 |
| NobodyCam | ack.. I'm prob more helpful with devtest.. but I can try and help | 20:30 |
| NobodyCam | whats in n-compute | 20:31 |
| jbjohnso_ | lifeless, well, historically for xCAT, we only 'went to town' if a 'discovery pool' was added | 20:31 |
| lifeless | jbjohnso_: sure, not trying to constraint the impl; just saying make it opt in | 20:31 |
| jbjohnso_ | lifeless, but in this particular case, I was thinking the setup would default to PXE snooping only | 20:31 |
| lifeless | literally the first question every DC ops folk who've deployed tripleo have asked me is 'will this break other DHCPing devices' | 20:31 |
| jbjohnso_ | lifeless, so if you have a PXE attempt that is not defined, it is snooped, but no response sent | 20:31 |
| jbjohnso_ | lifeless, if defined, then a dhcp response is still not sent | 20:32 |
| jbjohnso_ | lifeless, anyway, I need to actually implement that, but I need to make console session logging configurable and have it auto-respond to cfg changes before I'm ready | 20:33 |
| jbjohnso_ | and actually have it out with some copyright data | 20:33 |
| linggao | Hi devananda, I'd like to try this bug unless you have given it to someone else. https://bugs.launchpad.net/ironic/+bug/1279331 | 20:34 |
| notq | does devstack work where running stack.sh again does a reasonable resume? | 20:34 |
| notq | going to try to run the command manually and troubleshoot it | 20:35 |
| JayF | https://review.openstack.org/#/c/82680/ only needs one more core review to get tests sped up in Ironic, if someone has a moment or two | 20:35 |
| adam_g | notq, is nova-api running? | 20:35 |
| JayF | linggao: btw; that bug that you were inquiring about in the meeting --> https://review.openstack.org/#/c/82615/ | 20:36 |
| JayF | linggao: that's the patchset if you wanted to have a look. It's already got one +2 | 20:36 |
| linggao | JayF thanks. | 20:36 |
| openstackgerrit | Russell Haering proposed a change to openstack/ironic: Re-use hash rings across requests https://review.openstack.org/82915 | 20:40 |
| notq | i don't see any nova processes | 20:41 |
| devananda | notq: you should unstack before calling stack a second time | 20:41 |
| notq | devananda: thanks. | 20:42 |
| linggao | JayF, good catch on the dynamic timer. | 20:48 |
| JayF | linggao: ty, russell_h helped with finding it too | 20:49 |
| openstackgerrit | lifeless proposed a change to openstack/ironic: Provide a new ComputeManager for Ironic https://review.openstack.org/82637 | 20:51 |
| NobodyCam | JayF: .2 long enough? | 20:52 |
| *** mrda_away is now known as mrda | 20:52 | |
| notq | this time it says stack.sh completed. so good times. | 20:53 |
| NobodyCam | notq: :) | 20:53 |
| linggao | devananda, I'll take a look at bug 1279331. I have assined it to myself. | 20:54 |
| openstackgerrit | lifeless proposed a change to openstack/ironic: Provide a new ComputeManager for Ironic https://review.openstack.org/82637 | 20:54 |
| openstackgerrit | lifeless proposed a change to openstack/ironic: Provide a new ComputeManager for Ironic https://review.openstack.org/82637 | 20:55 |
| devananda | linggao: this is going to be a substantial amount of work | 20:55 |
| linggao | oo | 20:55 |
| linggao | Can you give me some insight? | 20:56 |
| devananda | i'm in meetings at the moment, so will respond sporadically | 20:56 |
| devananda | linggao: the bug report has a lengty explanation of the involved components | 20:57 |
| devananda | linggao: it will touch on the hash ring, conductor.manager, and driver code | 20:57 |
| linggao | devananda, yes. I red it and looks like a periodic task needs to be added. | 20:58 |
| linggao | red/read | 20:58 |
| devananda | yep | 20:59 |
| linggao | devananda, it looks like there are a lot of things needs to be done for that periodic task. I need to get myself familiar with those and get your advice before coding. | 21:07 |
| devananda | linggao: taking on a bug that complex may not be the best way to get familiar with it | 21:08 |
| *** shakayumi has joined #openstack-ironic | 21:08 | |
| *** shakayumi has quit IRC | 21:09 | |
| linggao | :-), I thought fixing bugs is a good way to learn the code. Well, I will give it back if you are not comfortable with it. | 21:11 |
| *** shakamunyi has quit IRC | 21:11 | |
| devananda | linggao: fixing bugs is a great way to learn code -- i'm just pointing out that that particular bug might be a bit much to take on at once by yourself | 21:14 |
| *** max_lobur has joined #openstack-ironic | 21:14 | |
| linggao | devananda, I understand. :) | 21:14 |
| *** ifarkas has quit IRC | 21:15 | |
| *** ifarkas_ has quit IRC | 21:16 | |
| *** max_lobur has quit IRC | 21:20 | |
| *** Shalmazar has joined #openstack-ironic | 21:20 | |
| NobodyCam | RuntimeError: Broken toolchain: cannot link a simple C program | 21:22 |
| NobodyCam | i wish it would sa which one! | 21:23 |
| *** linggao has quit IRC | 21:23 | |
| NobodyCam | s/sa/say/ | 21:24 |
| *** Madasi has quit IRC | 21:29 | |
| *** Shalmazar is now known as Madasi | 21:32 | |
| *** Madasi is now known as Guest2116 | 21:33 | |
| *** jbjohnso_ has quit IRC | 21:33 | |
| *** matty_du1s has joined #openstack-ironic | 21:42 | |
| *** matty_dubs has quit IRC | 21:45 | |
| *** notq has quit IRC | 21:46 | |
| *** russellb has quit IRC | 21:47 | |
| *** Guest2116 has quit IRC | 21:47 | |
| *** zigo has quit IRC | 21:47 | |
| *** wendar has quit IRC | 21:47 | |
| *** Shalmazar has joined #openstack-ironic | 21:47 | |
| *** russellb has joined #openstack-ironic | 21:48 | |
| *** notq has joined #openstack-ironic | 21:49 | |
| *** wendar has joined #openstack-ironic | 21:49 | |
| NobodyCam | devananda: do we have any test on the class PeriodicService?? | 21:49 |
| *** zigo has joined #openstack-ironic | 21:49 | |
| devananda | possibly not | 21:50 |
| devananda | ? | 21:50 |
| NobodyCam | I not seeing any :-p | 21:50 |
| NobodyCam | virsh list | 21:51 |
| NobodyCam | ww | 21:51 |
| JayF | With blueprints, when you subscribe, what does it mean when you check 'participation essential' | 22:02 |
| devananda | JayF: afaik, nothing really | 22:05 |
| JayF | cool then I'm checking it | 22:05 |
| JayF | never seems like a good idea to be nonessential | 22:06 |
| JayF | except on snow days ;) | 22:06 |
| NobodyCam | JayF: spam you until you reply, maybe? | 22:06 |
| openstackgerrit | A change was merged to openstack/ironic: Run ipmi power status less aggressively https://review.openstack.org/82668 | 22:18 |
| NobodyCam | devananda: ^^^ first of the four.. just +a the last one | 22:20 |
| NobodyCam | appears to be afternoon walkies time brb | 22:24 |
| devananda | awesome, ty | 22:24 |
| JayF | NobodyCam: what about https://review.openstack.org/#/c/82615/ ? | 22:25 |
| openstackgerrit | A change was merged to openstack/ironic: Remove logging of exceptions from controller's methods https://review.openstack.org/81267 | 22:25 |
| NobodyCam | JayF: check zuul | 22:25 |
| *** matty_du1s is now known as matty_dubs|gone | 22:26 | |
| JayF | NobodyCam: I meant more that it needed an additional +2/+A... was fishing for a review ;) | 22:26 |
| *** Shalmazar is now known as Madasi | 22:27 | |
| JayF | devananda: any further bugs I can assist with? I see all the ones from https://launchpad.net/ironic/+milestone/icehouse-rc1 are in progress or assigned to others | 22:34 |
| devananda | JayF: hm.. take a look through the open bugs list then | 22:35 |
| devananda | JayF: the rc1 list are the ones i've seen,a nd i try to go through the full list about once a week. another pair of eyes triaging bugs certainly wouldn't hurt | 22:36 |
| devananda | it's quite possible i missed something | 22:36 |
| NobodyCam | JayF: I +a'd it :) | 22:39 |
| JayF | woo | 22:41 |
| JayF | NobodyCam: why doesn't it show up in gerrit then? | 22:41 |
| NobodyCam | checks | 22:42 |
| devananda | https://review.openstack.org/#/c/82615/ doesn't show a +A | 22:43 |
| JayF | I guess I would've just expected to see your name up in the reviewers with a check | 22:43 |
| devananda | woot - our review queue is down to a single page again! | 22:43 |
| devananda | i'm almost not sure if that's a good thing, lol | 22:44 |
| JayF | just wait, once you get the floodgates opened for juno it'll start flowing :) | 22:44 |
| JayF | I know we'll have the agent driver coming in soon | 22:44 |
| NobodyCam | oh my bad... I had it left open because I couldnot believe a change that didn't break a test somewhere | 22:44 |
| NobodyCam | i was side tracked looking for tests | 22:45 |
| JayF | NobodyCam: if anything that might be a sign of bad tests :x | 22:45 |
| JayF | or no tests at all | 22:45 |
| NobodyCam | JayF: saddly its the latter | 22:45 |
| JayF | Yeah I def. ran all tests locally :) had to make a case-senstive mac partition to do it, haha | 22:46 |
| NobodyCam | you've got our tests running on mac? | 22:48 |
| NobodyCam | how you do it | 22:48 |
| NobodyCam | i keep running into raise DistributionNotFound(req) # XXX put more info here | 22:48 |
| JayF | Not /on/ mac directly | 22:48 |
| JayF | but in a VM with my mac dir mounted via NFS | 22:49 |
| NobodyCam | oh :-p | 22:49 |
| JayF | but I suspect it's possible on mac | 22:49 |
| JayF | you just have to make a partition with Case Sensitivity | 22:49 |
| JayF | because there are apparently some libraries that differ only by capitalization, so without case senstivity, the imports are fubar'd | 22:49 |
| NobodyCam | ok I approved that one | 22:49 |
| NobodyCam | :-p | 22:49 |
| NobodyCam | thought I had | 22:50 |
| devananda | i was about to ask, NobodyCam, how you planned on testing https://review.openstack.org/#/c/82615/7 anyway | 22:50 |
| devananda | our unit tests shouldn't be testing functionality within an external lib (eg, dynamic_timer) | 22:50 |
| devananda | I would be inclined to have a functional test (eg in tempest) for it | 22:51 |
| NobodyCam | yea, think they do need to be tested | 22:51 |
| *** mdurnosvistov_ has quit IRC | 22:51 | |
| devananda | JayF: oh, this one https://bugs.launchpad.net/ironic/+bug/1265587 // https://review.openstack.org/#/c/64711/ | 22:52 |
| devananda | JayF: apparently assigned to me and i dropped the ball on the fix. | 22:52 |
| JayF | Sure I'll see if I can pick it up | 22:53 |
| devananda | thanks | 22:53 |
| * devananda steps away to find food and stretch legs | 22:53 | |
| NobodyCam | :) | 22:58 |
| * NobodyCam remembers to set DIB_REPOLOCATION this time | 22:59 | |
| openstackgerrit | A change was merged to openstack/ironic: Speed up ironic tests https://review.openstack.org/82680 | 22:59 |
| *** epim has quit IRC | 23:01 | |
| * JayF watches his tests run gloriously fast on master | 23:01 | |
| JayF | jroll++ | 23:01 |
| *** krtaylor has quit IRC | 23:23 | |
| lifeless | comstud: https://review.openstack.org/#/c/82637/ answered you | 23:24 |
| lifeless | devananda: ^ is one of the now three patches left for direct Ironic support in tripleo-ci | 23:24 |
| *** jgrimm has quit IRC | 23:25 | |
| devananda | ack | 23:25 |
| devananda | lifeless: admittedly i dont quite follow your comment re: db-less scheduler | 23:26 |
| devananda | lifeless: pending comstud's reply, it lgtm | 23:27 |
| openstackgerrit | A change was merged to openstack/ironic: Run periodic_task in a with a dynamic timer https://review.openstack.org/82615 | 23:27 |
| NobodyCam | lifeless: is my DIB_REPOREF_ironic=refs/changes/37/82637/9 still valid :) | 23:29 |
| openstackgerrit | Adam Gandelman proposed a change to openstack/ironic: Raise unexpected exceptions during destroy() https://review.openstack.org/82948 | 23:50 |
| *** lucas-afk has quit IRC | 23:52 | |
| NobodyCam | ahh nice chatch adam_g :) | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!