| comstud | that's probably not my ideal, but I'm not unhappy with it | 00:00 |
|---|---|---|
| comstud | the image builder code could be moved into oslo so it can be shared | 00:01 |
| comstud | even if its small enough to copy for now | 00:01 |
| russell_h | +1 to that too | 00:01 |
| comstud | this should allow you to avoid any nova changes, really, and get something working more quickly | 00:03 |
| comstud | which I think should be one of the goals:) | 00:03 |
| NobodyCam | comstud: ++ for getting things working quickly.. I like todo's / fixme: pointing out why we diverted from what would apper the "standard" method.. | 00:07 |
| comstud | nod | 00:08 |
| NobodyCam | :) | 00:08 |
| comstud | i'm dropping off (i'm technically off today :) | 00:08 |
| NobodyCam | :) after 5 here :) | 00:08 |
| NobodyCam | night comstud :) | 00:08 |
| comstud | same here | 00:09 |
| comstud | laters | 00:09 |
| russell_h | its definitely time to booze | 00:09 |
| comstud | good idea | 00:09 |
| * NobodyCam has no beer or cider... looks like i'm down (or up) to Rum & (diet) coke | 00:10 | |
| comstud | horrible problem to have | 00:10 |
| comstud | :) | 00:10 |
| * comstud & ok out | 00:11 | |
| * jroll throws a beer at NobodyCam | 00:11 | |
| * jroll walks away as well | 00:11 | |
| NobodyCam | thank you and night jroll | 00:11 |
| NobodyCam | :) | 00:11 |
| *** blamar has quit IRC | 00:17 | |
| *** eghobo has quit IRC | 00:18 | |
| *** matsuhashi has joined #openstack-ironic | 00:21 | |
| devananda | jroll: same bug tracker in LP is fine | 00:28 |
| *** devananda 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" | 00:34 | |
| *** blamar has joined #openstack-ironic | 00:34 | |
| NobodyCam | w00h00 Core 0: +98.0°C (high = +86.0°C, crit = +100.0°C) | 00:37 |
| devananda | NobodyCam: yikes | 00:42 |
| devananda | NobodyCam: white board cleaned up now | 00:42 |
| NobodyCam | devananda: ye brick after removing .cache dir | 00:43 |
| NobodyCam | devananda: awesome | 00:43 |
| *** lucasagomes has quit IRC | 00:53 | |
| *** jbjohnso has joined #openstack-ironic | 01:03 | |
| *** dwalleck has joined #openstack-ironic | 01:07 | |
| *** blamar has quit IRC | 01:23 | |
| openstackgerrit | Michael Davies proposed a change to openstack/ironic: Improve Ironic Conductor threading & locks https://review.openstack.org/78515 | 01:27 |
| *** dwalleck_ has joined #openstack-ironic | 01:29 | |
| *** dwalleck has quit IRC | 01:33 | |
| NobodyCam | fyi: removing snd recreating ~/.cache/image-create solved my build woes | 01:37 |
| *** nosnos has joined #openstack-ironic | 01:40 | |
| *** jbjohnso has quit IRC | 01:55 | |
| *** antonym has quit IRC | 02:03 | |
| *** antonym has joined #openstack-ironic | 02:03 | |
| *** dwalleck_ has quit IRC | 02:09 | |
| devananda | NobodyCam: are you working on / pushing up another rev of 72969? | 02:09 |
| NobodyCam | the under cloud bits yes | 02:16 |
| NobodyCam | as soon as I can test | 02:16 |
| NobodyCam | :) | 02:17 |
| rloo | devananda: do you have a few minutes to discuss https://bugs.launchpad.net/ironic/+bug/1290645 | 02:17 |
| *** lazy_prince is now known as killer_prince | 02:18 | |
| devananda | rloo: that's disconcerting | 02:20 |
| rloo | devananda. so the oslo code moved the configs from ../session.py to a new options.py, and in the move, they removed the default setting. | 02:20 |
| rloo | devananda: but they have a method that can be used to set the defaults. Assuming the method works, we can add code to call it. | 02:21 |
| rloo | we actually have code that calls it, in common/config.py | 02:22 |
| rloo | parse_args() | 02:22 |
| rloo | but I am wondering why common/config.py exists. I couldn't find any code that uses it. | 02:22 |
| rloo | devananda: and service.py, prepare_service() has similar code to config.parse_args(). So it seems like prepare_service() should call config.parse_args(), or get rid of config.parse_args() and move code to prepare_service()? | 02:24 |
| rloo | devananda: and/or I can just put up a review, and people can -X it. ha ha. | 02:24 |
| devananda | rloo: sounds like two issues | 02:25 |
| rloo | devananda. yes. Ok, I'll copy the default value from config.parse_args() for now, to service.py file. | 02:25 |
| devananda | rloo: thanks! | 02:27 |
| devananda | rloo: as for common.config vs service.prepare_service, i'm not sure. but they both seem to be in use | 02:27 |
| devananda | ironic/tests/conf_fixture.py:from ironic.common import config | 02:27 |
| rloo | devananda. ah, I thought I did a grep. must have missed that. thx. | 02:28 |
| *** derekh has quit IRC | 02:33 | |
| jroll | 17:28:44 @devananda | jroll: same bug tracker in LP is fine <- cool, thanks. /cc JayF | 02:40 |
| *** killer_prince is now known as lazy_prince | 02:44 | |
| *** lazy_prince is now known as killer_prince | 02:45 | |
| devananda | rloo: i've confirmeed the bug but lowered it to medium -- there's a simple work around (set the connection param) | 02:49 |
| rloo | devananda: yes, that's right. thx. | 02:50 |
| *** harlowja is now known as harlowja_away | 02:54 | |
| *** harlowja_away is now known as harlowja | 03:06 | |
| *** vkozhukalov has quit IRC | 03:09 | |
| *** dwalleck_ has joined #openstack-ironic | 03:14 | |
| *** dwalleck has joined #openstack-ironic | 03:16 | |
| *** dwalleck_ has quit IRC | 03:19 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Conductor logs better error if seamicroclient missing https://review.openstack.org/79520 | 03:24 |
| *** matsuhashi has quit IRC | 03:41 | |
| rloo | russell_h: sorry, forgot to mention. wrt database.connection, use connection=sqlite:///ironic/openstack/common/db/ironic.sqlite or set to something else in your config file. (bug 1290645) | 03:49 |
| *** rloo has quit IRC | 03:52 | |
| *** dwalleck has quit IRC | 04:19 | |
| *** killer_prince is now known as lazy_prince | 04:24 | |
| *** rameshg87 has joined #openstack-ironic | 04:27 | |
| *** matsuhashi has joined #openstack-ironic | 04:40 | |
| *** eghobo has joined #openstack-ironic | 04:56 | |
| *** matsuhashi has quit IRC | 06:05 | |
| *** eghobo has quit IRC | 06:07 | |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/78862 | 06:07 |
| *** eghobo has joined #openstack-ironic | 06:08 | |
| *** matsuhas_ has joined #openstack-ironic | 06:08 | |
| *** harlowja is now known as harlowja_away | 06:08 | |
| *** eguz has joined #openstack-ironic | 06:09 | |
| *** eghobo has quit IRC | 06:09 | |
| *** eguz has quit IRC | 06:11 | |
| *** romcheg has joined #openstack-ironic | 06:20 | |
| *** bashok has joined #openstack-ironic | 06:23 | |
| *** romcheg has quit IRC | 06:24 | |
| *** vkozhukalov has joined #openstack-ironic | 06:27 | |
| *** killer_prince has joined #openstack-ironic | 06:40 | |
| *** eghobo has joined #openstack-ironic | 06:41 | |
| *** rsacharya has joined #openstack-ironic | 07:14 | |
| *** max_lobur has joined #openstack-ironic | 07:35 | |
| *** romcheg has joined #openstack-ironic | 07:37 | |
| *** max_lobur has quit IRC | 07:50 | |
| *** vkozhukalov has quit IRC | 07:50 | |
| *** dshulyak has joined #openstack-ironic | 07:54 | |
| *** rameshg87 has left #openstack-ironic | 07:55 | |
| *** mrda is now known as mrda_away | 08:01 | |
| *** ifarkas has quit IRC | 08:04 | |
| *** Manishanker has joined #openstack-ironic | 08:09 | |
| *** eghobo has quit IRC | 08:18 | |
| *** jistr has joined #openstack-ironic | 08:33 | |
| openstackgerrit | Yuriy Zveryanskyy proposed a change to openstack/ironic: Add TaskManager lock on change port data https://review.openstack.org/77623 | 08:33 |
| *** ifarkas has joined #openstack-ironic | 08:34 | |
| agordeev | morning Ironic :) | 08:39 |
| GheRivero | morning | 08:41 |
| agordeev | GheRivero: morning ! | 08:42 |
| yuriyz | morning GheRivero, agordeev | 08:44 |
| agordeev | yuriyz: morning | 08:46 |
| *** max_lobur has joined #openstack-ironic | 08:49 | |
| *** max_lobur1 has joined #openstack-ironic | 09:00 | |
| *** derekh has joined #openstack-ironic | 09:01 | |
| *** rsacharya has quit IRC | 09:02 | |
| *** max_lobur has quit IRC | 09:03 | |
| *** martyntaylor has joined #openstack-ironic | 09:06 | |
| *** rsacharya has joined #openstack-ironic | 09:06 | |
| *** killer_prince has quit IRC | 09:07 | |
| *** jrist has quit IRC | 09:07 | |
| *** matsuhas_ has quit IRC | 09:08 | |
| *** tatyana has joined #openstack-ironic | 09:08 | |
| *** matsuhashi has joined #openstack-ironic | 09:09 | |
| *** martyntaylor has quit IRC | 09:10 | |
| agordeev | ironic.nova.driver still requires 'v2.0' at the of admin_url. Is it really necessary? Can be a bug? | 09:21 |
| *** jrist has joined #openstack-ironic | 09:21 | |
| *** lucasagomes has joined #openstack-ironic | 09:24 | |
| *** martyntaylor has joined #openstack-ironic | 09:28 | |
| *** romcheg has quit IRC | 09:31 | |
| *** athomas has joined #openstack-ironic | 09:35 | |
| *** martyntaylor has quit IRC | 09:49 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Move duplicated _get_node(s)_mac_addresses() https://review.openstack.org/78787 | 09:50 |
| *** martyntaylor has joined #openstack-ironic | 09:53 | |
| *** martyntaylor has quit IRC | 09:56 | |
| *** martyntaylor has joined #openstack-ironic | 09:57 | |
| *** romcheg has joined #openstack-ironic | 10:04 | |
| *** rsacharya has quit IRC | 10:05 | |
| *** martyntaylor has quit IRC | 10:05 | |
| *** saju_m has joined #openstack-ironic | 10:05 | |
| *** killer_prince has joined #openstack-ironic | 10:07 | |
| *** saju_m has quit IRC | 10:18 | |
| *** martyntaylor has joined #openstack-ironic | 10:19 | |
| *** ndipanov_gone is now known as ndipanov | 10:26 | |
| openstackgerrit | Yuriy Zveryanskyy proposed a change to openstack/ironic: Add TaskManager lock on change port data https://review.openstack.org/77623 | 10:28 |
| *** matsuhashi has quit IRC | 10:29 | |
| *** saju_m has joined #openstack-ironic | 10:30 | |
| mrda_away | Hi devananda, Just wondering whether you can take a look at https://review.openstack.org/#/c/78515/ as its ready for review and it's a bug marked High. It's the do_node_tear_down() pair for do_node_deploy(). | 10:33 |
| *** saju_m has quit IRC | 10:54 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: PXE validate() to fail if no Ironic API URL https://review.openstack.org/79579 | 11:00 |
| *** jrist has quit IRC | 11:58 | |
| *** linggao has joined #openstack-ironic | 12:03 | |
| *** killer_prince has quit IRC | 12:05 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: PXE validate() to fail if no Ironic API URL https://review.openstack.org/79579 | 12:19 |
| *** romcheg has quit IRC | 12:28 | |
| *** lucasagomes is now known as lucas-hungry | 12:31 | |
| *** jdob has joined #openstack-ironic | 12:31 | |
| *** romcheg has joined #openstack-ironic | 12:42 | |
| *** nosnos has quit IRC | 12:44 | |
| *** rloo has joined #openstack-ironic | 12:46 | |
| *** lazy_prince is now known as killer_prince | 12:56 | |
| *** jbjohnso has joined #openstack-ironic | 13:11 | |
| *** jrist has joined #openstack-ironic | 13:26 | |
| *** romcheg has quit IRC | 13:32 | |
| *** romcheg has joined #openstack-ironic | 13:32 | |
| *** lucas-hungry is now known as lucasagomes | 13:38 | |
| NobodyCam | good morning says the man making coffee | 13:54 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: PXE validate() to fail if no Ironic API URL https://review.openstack.org/79579 | 13:54 |
| rloo | mornin' NobodyCam | 13:55 |
| NobodyCam | morning rloo :) | 13:56 |
| lucasagomes | morning rloo NobodyCam | 13:59 |
| lucasagomes | :) | 13:59 |
| NobodyCam | morning lucasagomes :) | 13:59 |
| rloo | afternoon lucasagomes! | 13:59 |
| romcheg | Morning NobodyCam lucasagomes rloo! | 13:59 |
| jbjohnso | after digging so far into obscure terminal stuff, I can do some fun practical jokes | 13:59 |
| jbjohnso | like putting echo -e '\x1b)0\x0e' into someone's bashrc | 14:00 |
| rloo | hi romcheg! | 14:00 |
| rloo | jbjohnso: that scares me ;) | 14:00 |
| NobodyCam | hey hey romcheg :) morning morning jbjohnso | 14:00 |
| jbjohnso | try it out, it has a neat effect | 14:00 |
| jbjohnso | echo -e '\x0f' to undo | 14:01 |
| agordeev | jbjohnso: "reset" is also to undo | 14:02 |
| *** jbjohnso_ has joined #openstack-ironic | 14:03 | |
| jbjohnso_ | the things you learn when trying to figure out why ESXi's serial console does not really work well at all unless your terminal was watching throughout the whole boot process | 14:03 |
| *** jbjohnso has quit IRC | 14:06 | |
| *** dwalleck has joined #openstack-ironic | 14:09 | |
| *** dwalleck_ has joined #openstack-ironic | 14:11 | |
| *** dwalleck has quit IRC | 14:14 | |
| NobodyCam | brb | 14:20 |
| *** matty_dubs|gone is now known as matty_dubs | 14:28 | |
| GheRivero | morning all | 14:29 |
| NobodyCam | good morning GheRivero :) | 14:29 |
| *** dwalleck_ has quit IRC | 14:31 | |
| *** dwalleck_ has joined #openstack-ironic | 14:35 | |
| *** jrist has quit IRC | 14:43 | |
| *** jbjohnso_ has quit IRC | 14:55 | |
| *** blamar has joined #openstack-ironic | 15:02 | |
| *** dwalleck_ has quit IRC | 15:05 | |
| *** bashok has quit IRC | 15:05 | |
| devananda | good morning, all | 15:07 |
| NobodyCam | good orning devananda :) | 15:08 |
| *** jbjohnso_ has joined #openstack-ironic | 15:08 | |
| NobodyCam | morning even | 15:08 |
| NobodyCam | :-p | 15:08 |
| *** digambar has joined #openstack-ironic | 15:09 | |
| devananda | lucasagomes: i18n in a unit test? | 15:11 |
| *** jistr has quit IRC | 15:11 | |
| lucasagomes | devananda, morning | 15:12 |
| romcheg | Morning devananda | 15:12 |
| lucasagomes | devananda, yeah i was a bit unsure as well so I put the '?' | 15:12 |
| lucasagomes | i18n? | 15:12 |
| lucasagomes | but we can leave without it | 15:12 |
| devananda | aiui, i18n is meant for user-facing strings | 15:12 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Conductor logs better error if seamicroclient missing https://review.openstack.org/79520 | 15:13 |
| lucasagomes | devananda, right, but that skipTest will print the string for the user running the tests no? | 15:14 |
| devananda | lucasagomes: true ... | 15:14 |
| devananda | lucasagomes: though I would point out that's a developer, not a user of the product | 15:15 |
| lucasagomes | devananda, yeah, its kinda trivial, but I was unsure as well if in that case we should i18n it, so I left the comment | 15:15 |
| lucasagomes | devananda, yeah | 15:15 |
| lucasagomes | ok, let's leave it without it then | 15:15 |
| devananda | thanks for pointing it out | 15:16 |
| NobodyCam | speacking of SM did we file a bug for not including SMclient in requirments? | 15:18 |
| lucasagomes | devananda, np, thank u for the patch | 15:19 |
| lucasagomes | devananda, btw | 15:19 |
| devananda | NobodyCam: no. but that patch removes it | 15:19 |
| lucasagomes | devananda, we might want to do something similar wiht pyghmi as well no? | 15:19 |
| lucasagomes | I mean we haven't decided yet whether native ipmi will be the default driver for ironic, but right now I think it's better to keep ipmitool because people are working on other stuff for it like console access | 15:20 |
| devananda | lucasagomes: ++ | 15:20 |
| *** jistr has joined #openstack-ironic | 15:35 | |
| *** jistr is now known as jistr|afk | 15:36 | |
| *** harlowja_away is now known as harlowja | 15:38 | |
| *** harlowja has quit IRC | 15:43 | |
| *** harlowja has joined #openstack-ironic | 15:48 | |
| *** eghobo has joined #openstack-ironic | 15:53 | |
| *** hemna has joined #openstack-ironic | 16:03 | |
| *** dwalleck_ has joined #openstack-ironic | 16:08 | |
| NobodyCam | brb | 16:10 |
| *** ifarkas has quit IRC | 16:11 | |
| *** yuriyz has quit IRC | 16:13 | |
| *** vkozhukalov has joined #openstack-ironic | 16:16 | |
| *** jrist has joined #openstack-ironic | 16:19 | |
| openstackgerrit | Jim Rollenhagen proposed a change to openstack/ironic: Add Node.instance_info field https://review.openstack.org/79466 | 16:21 |
| *** matty_dubs is now known as matty_dubs|lunch | 16:22 | |
| *** jgrimm has joined #openstack-ironic | 16:26 | |
| *** zigo has joined #openstack-ironic | 16:29 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Make sure auth_url exists and is not versionless https://review.openstack.org/79666 | 16:30 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Make sure auth_url exists and is not versionless https://review.openstack.org/79666 | 16:32 |
| jroll | devananda: if we decide to upload the configdrive to swift and use a pointer, do you see the nova driver doing the upload, or ironic special-casing that key and doing it? (I'd prefer the former, I think) | 16:33 |
| NobodyCam | gah quick walkies brb | 16:34 |
| *** jistr|afk is now known as jistr | 16:36 | |
| *** killer_prince is now known as lazy_prince | 16:42 | |
| lucasagomes | devananda, ping re hash ring replicas | 16:46 |
| lucasagomes | devananda, I saw ur patch to make it experimental | 16:46 |
| lucasagomes | what's missing for that? | 16:46 |
| lucasagomes | I mean, what is missing to make it not experimental? more tests? | 16:47 |
| *** ifarkas has joined #openstack-ironic | 16:58 | |
| *** rsacharya has joined #openstack-ironic | 16:59 | |
| JoshNang | re: the seamicro patch, we talked about 3rd party testing for 3rd party drivers at the meetup. where do you guys see the line between included and 3rd party drivers? | 17:00 |
| devananda | lucasagomes: functional testing | 17:01 |
| lucasagomes | devananda, ack | 17:01 |
| lucasagomes | devananda, thanks | 17:01 |
| devananda | lucasagomes: that patch series should go a long way towards rebalancing when the ring changes | 17:06 |
| devananda | lucasagomes: but it doesn't cover any of the prepare-early-so-rebalance-is-faster ideas | 17:06 |
| lucasagomes | JoshNang, included in the requirements u mean? hmm I think that the libraries for drivers that only works for a specific driver shouldn't be included in the requirements.txt | 17:06 |
| *** romcheg1 has joined #openstack-ironic | 17:07 | |
| devananda | lucasagomes: we also need a way to trigger a rebalance, and i prefer swift's approach of making it manual | 17:07 |
| lucasagomes | devananda, ahh right yeah, that needs some work to cache images before and all | 17:07 |
| lucasagomes | devananda, idk the swift's approach | 17:07 |
| lucasagomes | devananda, but by ur comment, ur saying that in case a conductor join/leave the cluster | 17:08 |
| devananda | lucasagomes: operator runs something to tell the ring to rebalance, since that is a pretty heavy-weight operation | 17:08 |
| lucasagomes | the ring won't be automatically rebalanced? | 17:08 |
| devananda | so | 17:08 |
| devananda | right now two somewhat orthogonal things happen | 17:08 |
| devananda | RPC routing and who-owns-what changes as soon as the ironic.conductors table either gets udpated, or the record's updated_at is older than the timeout | 17:09 |
| devananda | the API service is actually SELECTing from that table upon every API request -- I consider this a bug | 17:09 |
| devananda | it should be cached and refreshed periodically | 17:09 |
| lucasagomes | right | 17:10 |
| devananda | so we have the RPC routing based on the current hash ring | 17:10 |
| devananda | and | 17:10 |
| devananda | we have the currently prepared deployment environments, console state (not yet implemented), etc | 17:11 |
| devananda | i added an RPC fanout-cast to tell all the conductors "go check the ring and update your local state (files, console, etc) based on it" | 17:11 |
| lucasagomes | right, yeah API always gets it from the DB, it might be easy to make a cache using a dict mapping + periodic task to update it. But in case we have many API services running we also want to make sure they all would redirect a request to one node to the specific conductor which is managing it | 17:12 |
| lucasagomes | so we have to make sure things r sync'd between the API service | 17:13 |
| lucasagomes | services* | 17:13 |
| lucasagomes | devananda, ah awesome! | 17:13 |
| devananda | lucasagomes: right. which gets tricky if we're caching it in memory in each API | 17:13 |
| lucasagomes | devananda, yes | 17:13 |
| devananda | this is where a distributed in-memory cache is compelling | 17:14 |
| devananda | or a static mapping and a "button" to rebalance it | 17:14 |
| devananda | eg. what swift does | 17:14 |
| lucasagomes | +1 yeah I would vote to use some dist mem cach mechnism for this type of situations | 17:14 |
| lucasagomes | memcacache idk | 17:14 |
| lucasagomes | ok | 17:15 |
| russell_h | could the conductor respond with some known error if it doesn't own the node the API thinks it does | 17:15 |
| russell_h | and that triggers a refresh in the API? | 17:15 |
| russell_h | the thing with a distributed cache is you need to be careful with how you write to it, you want a guarantee of strong consistency for hash rings / master election | 17:16 |
| lucasagomes | russell_h, hmm it's a good idea, but I don't think conductors explicity knows which node it "owns" | 17:17 |
| lucasagomes | all conductors with the drivers to manage that node is a potential owner | 17:17 |
| russell_h | I volunteer to make this modular, so you can choose between a zookeeper backend or a dbapi (or sqlalchemy? not sure how best to model that) backend | 17:17 |
| lucasagomes | also, to make it trigger the api to refresh is hard cause if there's many APIs under a haproxy idk hw to trigger that one specifically | 17:18 |
| lucasagomes | so that's why I think that having all APIs to use a dist. in-mem cache mechanism would be a better implementation | 17:18 |
| russell_h | well | 17:18 |
| russell_h | so there are two cases you care about | 17:18 |
| russell_h | a conductor shuts down cleanly, or a conductor dies | 17:19 |
| devananda | russell_h: also, a conductor is only shut down temporarily | 17:19 |
| russell_h | if it dies, it won't be removed from the hashring until its database entry is past the TTL | 17:19 |
| devananda | is a third case | 17:19 |
| russell_h | ah, yes | 17:19 |
| russell_h | good point | 17:19 |
| russell_h | is there a mechanism for that today? | 17:19 |
| devananda | no | 17:20 |
| russell_h | its almost like you want to make the case where it removes itself from the database a special case | 17:20 |
| russell_h | for example, Cassandra has some tool you can run to remove a node from the ring | 17:21 |
| russell_h | otherwise it is presumed to be down, but eventually to return | 17:21 |
| devananda | exactly | 17:21 |
| devananda | this is akin to swift's tool that informs the cluster a rebalance is needed | 17:22 |
| *** matty_dubs|lunch is now known as matty_dubs | 17:24 | |
| lucasagomes | devananda, [off-topic] I know u and lifeless have spoken a good bit about using bittorrent or multicast for transfering the image to the nodes | 17:25 |
| *** martyntaylor has quit IRC | 17:25 | |
| lucasagomes | do you guys have any etherpad or wiki page or idk... with details of that? | 17:26 |
| devananda | lifeless: did ya'll sketch something out for the glance multicast stuff? | 17:26 |
| devananda | lucasagomes: not that I have on hand | 17:26 |
| lucasagomes | devananda, I see, ok, mind if I pick ur brain some time this week to talk about it? | 17:27 |
| *** ndipanov is now known as ndipanov_gone | 17:28 | |
| openstackgerrit | Jenkins proposed a change to openstack/ironic: Updated from global requirements https://review.openstack.org/79334 | 17:30 |
| devananda | lucasagomes: don't mind at all | 17:30 |
| lucasagomes | devananda, ack thanks, I will give it some thoughts and raise some questions | 17:30 |
| devananda | russell_h: we have a mapping impelemented in ironic.common.hash_ring. right now, it's sync'd on the API side by always polling the DB (abstracting that so it could use zookeeper would be great) | 17:33 |
| devananda | russell_h: where I think we might have an issue today is that the API will respond to changes in the ring faster than the conductors, and the tear_down() method may not be idempotent | 17:34 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Replace fixtures with mock in test_keystone.py https://review.openstack.org/79702 | 17:38 |
| *** rupsky has joined #openstack-ironic | 17:38 | |
| devananda | iow, if conductor A did a deploy, then is stopped, and a user requests the deploy be torn down, the API will route it to another conductor that doesn't have the deploy environment on disk yet | 17:38 |
| devananda | and I suspect that will raise exceptions | 17:38 |
| devananda | which it should not | 17:39 |
| russell_h | ah, gotcha | 17:39 |
| russell_h | its also going to crush the database | 17:39 |
| russell_h | when it lists out every conductor on every API request | 17:39 |
| russell_h | I started a fix for that the other day but never finished it | 17:40 |
| devananda | yea | 17:41 |
| devananda | I had thought the API cached it | 17:41 |
| *** rupsky has quit IRC | 17:41 | |
| devananda | but it doesn't | 17:41 |
| *** rupsky has joined #openstack-ironic | 17:41 | |
| devananda | the conductor is | 17:41 |
| russell_h | I can submit a fix for that today | 17:41 |
| devananda | http://git.openstack.org/cgit/openstack/ironic/tree/ironic/api/hooks.py#n97 | 17:43 |
| russell_h | what if we attached a hashring timestamp to every message | 17:43 |
| devananda | that's the obj we need to cache | 17:43 |
| *** tatyana has quit IRC | 17:43 | |
| russell_h | so if a conductor has an older hashring than the request, and the request appears to have been routed incorrectly (according to its view of the world), it would refresh | 17:44 |
| *** romcheg1 has quit IRC | 17:44 | |
| *** rsacharya has quit IRC | 17:44 | |
| russell_h | devananda: right, we also need the ConductorAPI to periodically refresh its hash rings | 17:44 |
| *** athomas has quit IRC | 17:45 | |
| russell_h | anyway, I'll fix the caching thing, we can address the expiration issues separately | 17:45 |
| lucasagomes | russell_h, devananda maybe open a bug for it and tag as rc ? | 17:48 |
| devananda | russell_h: i would prefer the refresh/resync be either continuous (eg, zookeeper), event-based (something wrouted wrong), or only-manual | 17:48 |
| *** rupsky has quit IRC | 17:49 | |
| devananda | periodic refresh alone will result in everything having different views for some time, when ever there's a change | 17:50 |
| *** max_lobur1 has quit IRC | 17:51 | |
| *** eghobo has quit IRC | 17:53 | |
| *** rupsky has joined #openstack-ironic | 17:53 | |
| devananda | russell_h: I think refresh when there's an unexpected message would be the easiest to implement on the conductor side | 17:53 |
| *** eghobo has joined #openstack-ironic | 17:55 | |
| *** derekh has quit IRC | 18:00 | |
| Shrews | adam_g: have you tried the latest devstack deploy? conductor mgr is giving me "Timeout reached when waiting callback for node" for a deploy. | 18:04 |
| devananda | Shrews: is the deploy starting, then timing out? or just not starting? | 18:05 |
| Shrews | devananda: starting, then timing out | 18:05 |
| devananda | Shrews: increase the timeout :) | 18:05 |
| NobodyCam | ok expanse report done and registered for atlanata too :) | 18:06 |
| Shrews | devananda: which setting is that? | 18:06 |
| devananda | Shrews: though, the default seems to be 30 minutes ... which is crazy if that's timing out | 18:06 |
| devananda | Shrews: deploy_callback_timeout | 18:06 |
| Shrews | devananda: yeah, it seemed crazy long already | 18:06 |
| devananda | when i ran through it last week, it was ~ 10 minutes | 18:07 |
| *** romcheg1 has joined #openstack-ironic | 18:07 | |
| devananda | given hpcs performance, i wasn't terribly surprised. 30 is too long though | 18:07 |
| Shrews | devananda: is there any reason to peg nova with NOVA_BRANCH=refs/changes/26/71026/15 anymore? | 18:08 |
| devananda | nope | 18:08 |
| devananda | should be nova and ironic, both from trunk now | 18:08 |
| Shrews | yeah, that's what I'm doing. just checking | 18:09 |
| devananda | Shrews: https://etherpad.openstack.org/p/IronicAndDevstackAgain seems quite out of date | 18:12 |
| devananda | s/quite/slightly/ | 18:12 |
| devananda | i'm giving it a run now | 18:12 |
| devananda | but if you have local notes and can toss them up, it'd be great | 18:12 |
| Shrews | devananda: change the patchset number from 20 to 21 | 18:13 |
| Shrews | otherwise, pretty up-to-date | 18:13 |
| Shrews | adam_g's 2nd bullet point is addressed by pulling down the latest devstack | 18:14 |
| devananda | ah | 18:14 |
| devananda | Shrews: i've updated the localrc. how's this look: http://paste.openstack.org/show/73169/ | 18:14 |
| Shrews | devananda: add HOST_IP | 18:16 |
| devananda | static value? | 18:16 |
| devananda | i thought that was interpreted | 18:16 |
| Shrews | that actually may only needed for us using HP cloud instances | 18:17 |
| Shrews | be* needed | 18:17 |
| Shrews | pretty much what i have, though | 18:18 |
| adam_g | Shrews, devstack just completed. will let you know | 18:26 |
| Shrews | i just bumped the timeout. going to try again | 18:27 |
| Shrews | biab | 18:29 |
| lucasagomes | I'm done for today | 18:30 |
| lucasagomes | have a good night everybody | 18:30 |
| NobodyCam | night lucasagomes | 18:30 |
| adam_g | nova boot gets me a running libvirt domain and an instance that is ACTIVE (though power state is Shutdown) | 18:30 |
| adam_g | Shrews, ^ | 18:30 |
| lucasagomes | adam_g, it takes a while to sync the power state from ironic to nova | 18:31 |
| adam_g | ah | 18:31 |
| lucasagomes | I think the nova default's periodic task is 180 sec | 18:31 |
| adam_g | looks like it.. running now | 18:33 |
| lucasagomes | adam_g, actually more 600 secs, see CONF.sync_power_state_interval in nova | 18:33 |
| lucasagomes | anyhoo, gotta run have a good night | 18:33 |
| Shrews | adam_g: you need to periodically check the provisioning state with 'ironic node-show' | 18:33 |
| Shrews | need a walk, back soon | 18:33 |
| devananda | adam_g: Shrews: i copied some of my old notes to that etherpad. not sure if they're still necessary or if you guys have other work arounds in place | 18:34 |
| devananda | g'night, lucasagomes | 18:35 |
| *** jbjohnso_ has quit IRC | 18:37 | |
| *** jbjohnso_ has joined #openstack-ironic | 18:42 | |
| adam_g | is tftp expected to be served directly from the ironic node (via the infrastructure network) or from within the neutron tenant fixed network? | 18:46 |
| devananda | adam_g: mgmt network. though i'm not sure that's well separated today | 18:52 |
| NobodyCam | walkies ...brb | 18:52 |
| *** digambar has quit IRC | 18:53 | |
| *** max_lobur has joined #openstack-ironic | 18:54 | |
| lifeless | devananda: lucasagomes https://etherpad.openstack.org/p/glance-multicast-image-transfer | 19:03 |
| *** harlowja has quit IRC | 19:04 | |
| *** harlowja has joined #openstack-ironic | 19:08 | |
| *** lazy_prince has quit IRC | 19:14 | |
| *** killer_prince has joined #openstack-ironic | 19:24 | |
| *** SpamapS has quit IRC | 19:26 | |
| *** SpamapS_ has joined #openstack-ironic | 19:29 | |
| *** SpamapS_ is now known as SpamapS | 19:30 | |
| *** ifarkas has quit IRC | 19:33 | |
| *** jgrimm has quit IRC | 19:34 | |
| devananda | adam_g: have you seen something overwrite /etc/resolv.conf with the host's IP ? | 19:36 |
| devananda | this is causing the git clone for diskiage-builder to fail | 19:37 |
| devananda | so it's happening before that | 19:37 |
| devananda | perhaps when neutron starts? | 19:37 |
| adam_g | devananda, something being devstack related or a general 'something'? | 19:37 |
| adam_g | resolvconf will do that | 19:37 |
| devananda | right, so i'm explicitly mangling resolv.conf before i start | 19:38 |
| NobodyCam | woo hoo : undercloud | CREATE_IN_PROGRESS | 19:38 |
| NobodyCam | brb | 19:38 |
| devananda | adam_g: and resolvconf is mangling it in a different way that actually breaks it, because 10.0.0.3 can't resolv github | 19:39 |
| adam_g | devananda, ive been doing the same without issue. | 19:40 |
| * devananda tries agai | 19:40 | |
| devananda | n | 19:40 |
| adam_g | devananda, you should be able to turn it off 'resolvconf --disable-updates' | 19:41 |
| adam_g | Shrews, what are you suing for FIXED_RANGE and NETWORK_GATEWAY on HPC? having trouble with instances reaching the tftp server, and my neutron L3 agents are all 'DOWN' | 19:50 |
| Shrews | adam_g: i was using FIXED_RANGE=172.16.0.0/24 and NETWORK_GATEWAY=172.16.0.1, but i suspect issues with the gateway i've chosen. trying something else, atm | 19:52 |
| NobodyCam | ahh new error InstanceDeployFailure: PXE deploy failed for instance :) | 19:56 |
| openstackgerrit | Adam Gandelman proposed a change to openstack/python-ironicclient: Avoid traceback with insufficient auth credentials https://review.openstack.org/79494 | 19:57 |
| *** jbjohnso_ has quit IRC | 19:58 | |
| *** romcheg1 has quit IRC | 20:02 | |
| *** jistr has quit IRC | 20:02 | |
| *** romcheg1 has joined #openstack-ironic | 20:02 | |
| Shrews | I don't understand enough about neutron, but I'm getting a floating IP assigned to the external network bridge (br-ex). I thought the NETWORK_GATEWAY ip should be assigned to it? | 20:11 |
| *** vkozhukalov has quit IRC | 20:13 | |
| *** _LXXIII has joined #openstack-ironic | 20:18 | |
| *** _LXXIII has quit IRC | 20:20 | |
| *** _LXXIII has joined #openstack-ironic | 20:21 | |
| *** _LXXIII has quit IRC | 20:23 | |
| *** _LXXIII has joined #openstack-ironic | 20:23 | |
| *** rupsky has quit IRC | 20:27 | |
| *** rupsky has joined #openstack-ironic | 20:30 | |
| devananda | urgh... lost connectivity to my devstack VM -again- | 20:33 |
| lifeless | Shrews: on a nova compute host? or Ironic test setup/ or ... ? | 20:33 |
| Shrews | lifeless: devstack setup | 20:34 |
| devananda | ironic-client should not be spitting out the whole token when in debug mode ... | 20:35 |
| devananda | but - yay, deploy started. waiting for callback | 20:37 |
| * Shrews very interested in devananda's results | 20:38 | |
| devananda | took 3 false starts to get the environment prepared properly | 20:38 |
| devananda | resolvconf broke once, so git clone dib failed | 20:39 |
| devananda | something borked the network and i had to "nova floating-ip-disassociate" to be able to reconnect it this last time | 20:39 |
| adam_g | devananda, do VMs successfully tftpboot? | 20:39 |
| devananda | oh. that happened again | 20:40 |
| devananda | yay | 20:40 |
| devananda | adam_g: waiting | 20:40 |
| *** mrda_away is now known as mrda | 20:41 | |
| devananda | NobodyCam: nova lists the vm state as ACTIVE while it's still waiting for callback | 20:41 |
| devananda | NobodyCam: i consider that a bug. the nova driver should keep the state as SPAWNING until it's actually available | 20:41 |
| NobodyCam | devananda: ya | 20:41 |
| Shrews | devananda++ | 20:42 |
| Shrews | that's confused me more than once | 20:42 |
| adam_g | also--is the special network configuration for baremetal in lib/neutron relevant to ironic? or is that special to nova-baremetal? it looks like the patch should be special casing for ironic as well, but its currently not | 20:42 |
| mrda | morning all | 20:42 |
| *** jgrimm has joined #openstack-ironic | 20:42 | |
| devananda | adam_g: good point. the lib/baremetal stuff is legacy cruft. it should be ripped out IMO. | 20:42 |
| devananda | adam_g: it depended on bm_poseur which asn't been maintained for ~1yr | 20:43 |
| adam_g | devananda, ahh--im just wondering if neutron is setting up its networks appropriately for the ironic case--i imagine the same stuff its doing in the presence of n-baremetal needs to happen for ironic as well? | 20:44 |
| *** ceoqia has joined #openstack-ironic | 20:44 | |
| NobodyCam | ahh the driver just calls set_provision_state and then moves on | 20:45 |
| NobodyCam | https://github.com/openstack/ironic/blob/master/ironic/nova/virt/ironic/driver.py#L425 | 20:45 |
| devananda | adam_g: dont make assumptions based on lib/baremetal | 20:45 |
| devananda | it may be totally wrong | 20:45 |
| adam_g | devananda, oh no, im not :) i'm looking at lib/neutron | 20:45 |
| devananda | oh | 20:46 |
| adam_g | it does a 'is_baremetal && do_some_specific_ovs_setup' | 20:46 |
| devananda | hmm | 20:46 |
| adam_g | wondering if it needs to do a is_ironic as well--testing that now | 20:47 |
| devananda | is that part of the patch ya'll are working on? if so, it's probably reasonable | 20:47 |
| *** ceoqia has quit IRC | 20:47 | |
| adam_g | something is screwy for me with gateways and routers in neutron once everything is up | 20:47 |
| adam_g | and the instances cannot reach outside to the tftp server | 20:47 |
| adam_g | im curious if your nodes are doing the same now that you've gotten as far as me :) | 20:48 |
| Shrews | adam_g: i think you and i are hunting the same dragon. | 20:48 |
| devananda | i'm barely able to maintain a stable network conncetion to the HOST :( | 20:48 |
| adam_g | Shrews, does 'sudo ip netns list' show you a namespace for dhcp + router? | 20:49 |
| devananda | adam_g: so far i haven't seen any POST come back from the VM | 20:50 |
| devananda | it hasnt even started to copy the iamge | 20:50 |
| NobodyCam | devananda: looks like driver is waiting for nodes to destroy but not waiting on deploy :-p | 20:50 |
| devananda | there should be a POST to the API to /vendor_passthru/, then disk is mounted and image copied | 20:50 |
| Shrews | adam_g: no, but i'm rerunning stack.sh atm | 20:50 |
| Shrews | funny thing is devananda and I watched this process work last week | 20:51 |
| openstackgerrit | A change was merged to openstack/ironic: Conductor logs better error if seamicroclient missing https://review.openstack.org/79520 | 20:51 |
| Shrews | adam_g: fyi, as a quick experiment, i changed the 'if is_baremetal' check in lib/neutron to add '|| is_service_enabled ironic', but it bails out with the "Neutron settings for baremetal not set" error. | 20:58 |
| Shrews | i think we need to figure out what that code is actually doing and decide if we need it | 20:58 |
| *** jdob has quit IRC | 20:58 | |
| *** linggao has quit IRC | 21:02 | |
| devananda | we also need better error handling around delete | 21:02 |
| devananda | and to allow deletes during wait-callback | 21:03 |
| devananda | i thought there ws a patch for that? | 21:03 |
| lifeless | devananda: did you see the glance etherpad I linked? | 21:04 |
| devananda | lifeless: didn't read it yet. E_MEETINGS | 21:05 |
| NobodyCam | ya I thought so to? wasn't it looked at / landed on a review jam | 21:05 |
| devananda | NobodyCam: and yet -- 2014-03-11 21:01:25.394 21781 WARNING wsme.api [-] Client-side error: Node c5394778-5d81-4b56-b7b2-1e2164eb92ec is already being provisioned. | 21:05 |
| devananda | when i try to delet it | 21:05 |
| devananda | NobodyCam: the bigger problem is that nova won't retry a delete | 21:08 |
| devananda | NobodyCam: once it gets an exception during the delete attempt, it marks the node state and backs off, and prevcents any more user attempts to delete it | 21:09 |
| NobodyCam | oh | 21:10 |
| devananda | NobodyCam: there's got to be a sync deleted states periodic task in nova. either it's not running, very long, or our driver isn't tied in yet | 21:10 |
| devananda | http://paste.openstack.org/show/73191/ | 21:12 |
| *** romcheg1 has quit IRC | 21:28 | |
| *** harlowja has quit IRC | 21:28 | |
| *** rwsu has quit IRC | 21:28 | |
| *** _LXXIII has quit IRC | 21:28 | |
| *** max_lobur has quit IRC | 21:28 | |
| *** rloo has quit IRC | 21:28 | |
| *** dshulyak has quit IRC | 21:28 | |
| *** dtantsur has quit IRC | 21:28 | |
| *** rainya has quit IRC | 21:28 | |
| *** jgrimm has quit IRC | 21:28 | |
| *** dwalleck_ has quit IRC | 21:28 | |
| *** antonym has quit IRC | 21:28 | |
| *** ndipanov_gone has quit IRC | 21:28 | |
| *** mrda has quit IRC | 21:28 | |
| *** JoshNang has quit IRC | 21:28 | |
| *** bigjools has quit IRC | 21:28 | |
| *** Madasi has quit IRC | 21:28 | |
| *** shortstop has quit IRC | 21:28 | |
| *** eghobo has quit IRC | 21:28 | |
| *** hemna has quit IRC | 21:28 | |
| *** lynxman has quit IRC | 21:28 | |
| *** sbadia has quit IRC | 21:28 | |
| *** dhellmann has quit IRC | 21:28 | |
| *** pradipta has quit IRC | 21:28 | |
| *** lsmola has quit IRC | 21:28 | |
| *** morgabra has quit IRC | 21:28 | |
| *** ewindisch has quit IRC | 21:28 | |
| *** Shrews has quit IRC | 21:28 | |
| *** blamar has quit IRC | 21:28 | |
| *** davidlenwell has quit IRC | 21:28 | |
| *** Slower_ has quit IRC | 21:28 | |
| *** mdurnosvistov has quit IRC | 21:28 | |
| *** toure has quit IRC | 21:28 | |
| *** aignatov has quit IRC | 21:28 | |
| *** romcheg has quit IRC | 21:28 | |
| *** Hefeweizen has quit IRC | 21:28 | |
| *** datajerk1 has quit IRC | 21:28 | |
| *** openstackgerrit has quit IRC | 21:28 | |
| *** early has quit IRC | 21:28 | |
| *** pleia2 has quit IRC | 21:28 | |
| *** comstud has quit IRC | 21:28 | |
| *** matty_dubs has quit IRC | 21:28 | |
| *** lifeless has quit IRC | 21:28 | |
| *** mordred has quit IRC | 21:28 | |
| *** Haomeng has quit IRC | 21:28 | |
| *** lucasagomes has quit IRC | 21:28 | |
| *** greghaynes has quit IRC | 21:28 | |
| *** JayF has quit IRC | 21:28 | |
| *** EmilienM has quit IRC | 21:28 | |
| *** jroll has quit IRC | 21:29 | |
| *** wendar has quit IRC | 21:29 | |
| *** russellb has quit IRC | 21:29 | |
| *** adam_g has quit IRC | 21:29 | |
| *** mikal has quit IRC | 21:29 | |
| *** Manishanker has quit IRC | 21:29 | |
| *** ekarlso has quit IRC | 21:29 | |
| *** pquerna has quit IRC | 21:29 | |
| *** anteaya has quit IRC | 21:29 | |
| *** NobodyCam has quit IRC | 21:29 | |
| *** devananda has quit IRC | 21:29 | |
| *** wendar has joined #openstack-ironic | 21:41 | |
| *** russellb has joined #openstack-ironic | 21:41 | |
| *** _LXXIII has joined #openstack-ironic | 21:41 | |
| *** jroll has joined #openstack-ironic | 21:41 | |
| *** rloo has joined #openstack-ironic | 21:41 | |
| *** max_lobur has joined #openstack-ironic | 21:41 | |
| *** dtantsur has joined #openstack-ironic | 21:41 | |
| *** adam_g has joined #openstack-ironic | 21:41 | |
| *** dshulyak has joined #openstack-ironic | 21:41 | |
| *** greghaynes has joined #openstack-ironic | 21:41 | |
| *** lucasagomes has joined #openstack-ironic | 21:41 | |
| *** tatyana has joined #openstack-ironic | 21:41 | |
| *** EmilienM has joined #openstack-ironic | 21:41 | |
| *** derekh has joined #openstack-ironic | 21:41 | |
| *** mikal has joined #openstack-ironic | 21:41 | |
| *** JayF has joined #openstack-ironic | 21:41 | |
| *** mdurnosvistov_ has joined #openstack-ironic | 21:41 | |
| *** jgrimm has joined #openstack-ironic | 21:41 | |
| *** eghobo has joined #openstack-ironic | 21:41 | |
| *** dwalleck_ has joined #openstack-ironic | 21:41 | |
| *** hemna has joined #openstack-ironic | 21:41 | |
| *** blamar has joined #openstack-ironic | 21:41 | |
| *** romcheg has joined #openstack-ironic | 21:41 | |
| *** Manishanker has joined #openstack-ironic | 21:41 | |
| *** antonym has joined #openstack-ironic | 21:41 | |
| *** Hefeweizen has joined #openstack-ironic | 21:41 | |
| *** datajerk1 has joined #openstack-ironic | 21:41 | |
| *** ndipanov_gone has joined #openstack-ironic | 21:41 | |
| *** davidlenwell has joined #openstack-ironic | 21:41 | |
| *** mrda has joined #openstack-ironic | 21:41 | |
| *** JoshNang has joined #openstack-ironic | 21:41 | |
| *** Haomeng has joined #openstack-ironic | 21:41 | |
| *** Slower_ has joined #openstack-ironic | 21:41 | |
| *** openstackgerrit has joined #openstack-ironic | 21:41 | |
| *** toure has joined #openstack-ironic | 21:41 | |
| *** mdurnosvistov has joined #openstack-ironic | 21:41 | |
| *** shortstop has joined #openstack-ironic | 21:41 | |
| *** Madasi has joined #openstack-ironic | 21:41 | |
| *** bigjools has joined #openstack-ironic | 21:41 | |
| *** lynxman has joined #openstack-ironic | 21:41 | |
| *** sbadia has joined #openstack-ironic | 21:41 | |
| *** dhellmann has joined #openstack-ironic | 21:41 | |
| *** pradipta has joined #openstack-ironic | 21:41 | |
| *** lsmola has joined #openstack-ironic | 21:41 | |
| *** early has joined #openstack-ironic | 21:41 | |
| *** morgabra has joined #openstack-ironic | 21:41 | |
| *** ewindisch has joined #openstack-ironic | 21:41 | |
| *** Shrews has joined #openstack-ironic | 21:41 | |
| *** aignatov has joined #openstack-ironic | 21:41 | |
| *** ekarlso has joined #openstack-ironic | 21:41 | |
| *** pleia2 has joined #openstack-ironic | 21:41 | |
| *** comstud has joined #openstack-ironic | 21:41 | |
| *** pquerna has joined #openstack-ironic | 21:41 | |
| *** anteaya has joined #openstack-ironic | 21:41 | |
| *** matty_dubs has joined #openstack-ironic | 21:41 | |
| *** lifeless has joined #openstack-ironic | 21:41 | |
| *** mordred has joined #openstack-ironic | 21:41 | |
| *** NobodyCam has joined #openstack-ironic | 21:41 | |
| *** devananda has joined #openstack-ironic | 21:41 | |
| *** dickson.freenode.net sets mode: +o devananda | 21:41 | |
| *** rainya has joined #openstack-ironic | 21:41 | |
| devananda | adam_g: how have you not lost network access to the HPCS instance you're testing in? | 21:43 |
| devananda | adam_g: or are you not using it any more? | 21:43 |
| * devananda takes a break from meetings and gets some fresh air | 21:44 | |
| NobodyCam | :) | 21:44 |
| *** harlowja has joined #openstack-ironic | 21:45 | |
| adam_g | devananda, networking has been stable since last week | 21:46 |
| *** rwsu has joined #openstack-ironic | 21:46 | |
| adam_g | devananda, at least in east | 21:46 |
| devananda | adam_g: urgh. my instance is in the same az as the one i gave you, but keeps losing net | 21:47 |
| adam_g | region-b.geo-1 is what ive been using | 21:49 |
| devananda | adam_g: did you get access to spin up your own? | 21:49 |
| adam_g | devananda, yup. | 21:49 |
| devananda | adam_g: cool | 21:50 |
| devananda | i'll delete the one i made last week then | 21:50 |
| *** dwalleck_ has quit IRC | 22:00 | |
| *** dwalleck has joined #openstack-ironic | 22:02 | |
| *** max_lobur1 has joined #openstack-ironic | 22:04 | |
| *** max_lobur has quit IRC | 22:04 | |
| *** dwalleck_ has joined #openstack-ironic | 22:05 | |
| *** mdurnosvistov_ has quit IRC | 22:06 | |
| NobodyCam | just me or is anyone seeing oddness with https://github.com/openstack/nova | 22:06 |
| openstackgerrit | Ruby Loo proposed a change to openstack/ironic: Default value for database.connection config option https://review.openstack.org/79787 | 22:07 |
| NobodyCam | i'm getting page not found... other sites seem to be working, thou could just me TimeWarnner | 22:07 |
| rloo | NobodyCam: works for me | 22:08 |
| NobodyCam | rloo: thank you :) | 22:08 |
| *** dwalleck has quit IRC | 22:09 | |
| NobodyCam | hehe now its working for me :_p | 22:09 |
| NobodyCam | rloo: has the magic today :) | 22:10 |
| rloo | NobodyCam: just need to be patient ;) | 22:10 |
| *** _LXXIII has quit IRC | 22:11 | |
| *** rupsky_ has joined #openstack-ironic | 22:13 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Generic MixinVendorInterface using static mapping https://review.openstack.org/78384 | 22:13 |
| *** rupsky has quit IRC | 22:14 | |
| devananda | GheRivero: i'm landing ^ now. should https://review.openstack.org/#/c/71332/ be reopened on top of that? | 22:17 |
| *** matty_dubs is now known as matty_dubs|gone | 22:19 | |
| NobodyCam | brb | 22:23 |
| *** max_lobur1 has quit IRC | 22:25 | |
| *** tatyana has quit IRC | 22:27 | |
| *** dwalleck_ has quit IRC | 22:28 | |
| adam_g | Shrews, matching what the gate uses in localrc got me a functional neutron setup | 22:41 |
| adam_g | NETWORK_GATEWAY=10.1.0.1 | 22:41 |
| adam_g | FIXED_RANGE=10.1.0.0/24 | 22:41 |
| adam_g | FIXED_NETWORK_SIZE=25 | 22:41 |
| adam_g | paste fail.. FIXED_NETWORK_SIZE=256 | 22:42 |
| *** openstackstatus has joined #openstack-ironic | 22:44 | |
| NobodyCam | odd Warning: given size (11264) exceeds max allowable size (0) | 22:50 |
| NobodyCam | ok my deploys are failing at batemetal deploy helper with ^^ error from sfdisk | 23:01 |
| NobodyCam | https://review.openstack.org/#/c/79344 ahh | 23:12 |
| *** lucasagomes has quit IRC | 23:16 | |
| openstackgerrit | A change was merged to openstack/ironic: Generic MixinVendorInterface using static mapping https://review.openstack.org/78384 | 23:29 |
| openstackgerrit | Zhongyue Luo proposed a change to openstack/ironic: Remove unused method from timeutils https://review.openstack.org/78848 | 23:33 |
| *** rupsky_ has quit IRC | 23:41 | |
| adam_g | thoughts on an optional DIB element to backdoor deploy ramdisk with an ssh server for easier debugging? | 23:44 |
| NobodyCam | I'd love to at least get some logging from the deploy ramdisk | 23:45 |
| NobodyCam | :-p | 23:45 |
| NobodyCam | there use to break points in the ramdisk where it would drop to a prompt | 23:46 |
| NobodyCam | but I haven't looked at it in a long while | 23:46 |
| jroll | fwiw we ship an ssh server in our ramdisk image, it's super helpful | 23:46 |
| jroll | we also plan to ship logs back from our deploy agent. not sure how yet. | 23:47 |
| NobodyCam | :) | 23:47 |
| JoshNang | jroll: maybe ship them back with the heartbeat? | 23:48 |
| russell_h | I'd say we just ship them in a separate API call | 23:48 |
| russell_h | then we can make it pluggable what to do with them or something | 23:48 |
| russell_h | I haven't seen much in the way of logging infrastructure in openstack, but maybe I missed something | 23:49 |
| jroll | yeah, I was thinking just set up a log handler to something outside of the code, and configure that separately | 23:49 |
| jroll | e.g. logstash or whatever | 23:50 |
| NobodyCam | ssh root@$SEED_IP | 23:50 |
| NobodyCam | doh ww | 23:50 |
| *** dividehex has joined #openstack-ironic | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!