*** charcol has joined #openstack-ironic | 00:02 | |
*** yangyapeng has joined #openstack-ironic | 00:04 | |
*** rpioso|eat has joined #openstack-ironic | 00:06 | |
*** rpioso|eat is now known as rpioso | 00:07 | |
*** yangyapeng has quit IRC | 00:09 | |
*** adisky__ has joined #openstack-ironic | 00:25 | |
*** rama_y has quit IRC | 00:26 | |
*** crushil has quit IRC | 00:29 | |
*** crushil has joined #openstack-ironic | 00:29 | |
*** rivers has quit IRC | 00:30 | |
*** rivers has joined #openstack-ironic | 00:30 | |
*** crushil has quit IRC | 00:30 | |
*** crushil has joined #openstack-ironic | 00:31 | |
*** chas has joined #openstack-ironic | 00:49 | |
*** chas has quit IRC | 00:53 | |
*** yangyapeng has joined #openstack-ironic | 01:11 | |
*** ijw has quit IRC | 01:11 | |
*** ijw has joined #openstack-ironic | 01:16 | |
*** mat128 has joined #openstack-ironic | 01:27 | |
*** baoli has joined #openstack-ironic | 01:30 | |
*** gcb has joined #openstack-ironic | 01:34 | |
*** zhangfei has joined #openstack-ironic | 01:39 | |
*** rama_y has joined #openstack-ironic | 01:54 | |
*** TuanLA has joined #openstack-ironic | 01:54 | |
*** phuongnh has joined #openstack-ironic | 02:06 | |
*** baoli has quit IRC | 02:24 | |
*** zshi_ has joined #openstack-ironic | 02:37 | |
*** mat128 has quit IRC | 02:39 | |
*** baoli has joined #openstack-ironic | 02:45 | |
openstackgerrit | Wang KaiFeng proposed openstack/ironic master: Put tests in correct location for ironic/api/controllers/v1/ https://review.openstack.org/500658 | 02:49 |
---|---|---|
*** chas has joined #openstack-ironic | 02:49 | |
TheJulia | gah! I just noticed that like 20 minutes ago | 02:51 |
*** jlabarre has quit IRC | 02:52 | |
*** chas has quit IRC | 02:54 | |
*** rama_y has quit IRC | 02:56 | |
*** baoli has quit IRC | 03:07 | |
*** ijw has quit IRC | 03:22 | |
*** rama_y has joined #openstack-ironic | 03:28 | |
*** namnh has joined #openstack-ironic | 03:37 | |
*** links has joined #openstack-ironic | 03:51 | |
*** zshi_ has quit IRC | 03:52 | |
*** hshiina has joined #openstack-ironic | 04:02 | |
*** ijw has joined #openstack-ironic | 04:03 | |
*** ijw has quit IRC | 04:08 | |
*** gcb has quit IRC | 04:13 | |
*** udesale has joined #openstack-ironic | 04:18 | |
*** rama_y has quit IRC | 04:22 | |
*** gcb has joined #openstack-ironic | 04:26 | |
*** marios has joined #openstack-ironic | 04:44 | |
*** chas has joined #openstack-ironic | 04:50 | |
*** pcaruana has joined #openstack-ironic | 04:52 | |
*** MattMan_1 has quit IRC | 04:52 | |
*** MattMan_1 has joined #openstack-ironic | 04:53 | |
*** baoli has joined #openstack-ironic | 04:54 | |
*** chas has quit IRC | 04:55 | |
*** baoli has quit IRC | 04:59 | |
*** jtomasek has joined #openstack-ironic | 05:02 | |
*** srart has joined #openstack-ironic | 05:32 | |
*** mjura has joined #openstack-ironic | 05:35 | |
*** hoonetorg has quit IRC | 06:05 | |
*** jaganathan has joined #openstack-ironic | 06:15 | |
*** chas has joined #openstack-ironic | 06:22 | |
*** hoonetorg has joined #openstack-ironic | 06:22 | |
*** litao__ has joined #openstack-ironic | 06:26 | |
Tengu | hello Ironic people :) | 06:26 |
Tengu | small question: would it be possible to have a disk_label=gpt and boot_mode=bios alltogether? | 06:27 |
*** srart has quit IRC | 06:31 | |
*** hjensas has joined #openstack-ironic | 06:33 | |
*** charcol has quit IRC | 06:35 | |
*** ccamacho has joined #openstack-ironic | 06:39 | |
*** srart has joined #openstack-ironic | 06:46 | |
*** trungnv has joined #openstack-ironic | 06:46 | |
*** rcernin has joined #openstack-ironic | 06:46 | |
*** stendulker has joined #openstack-ironic | 06:49 | |
openstackgerrit | Hironori Shiina proposed openstack/ironic master: Fix unit test for new fields in invaid API version https://review.openstack.org/468723 | 06:52 |
Tengu | hmm. I suspect the doc might be misleading for GPT activation… | 06:54 |
Tengu | pretty sure I didn't fully understand the thing. :] | 06:54 |
*** karimb has joined #openstack-ironic | 07:01 | |
*** tesseract has joined #openstack-ironic | 07:13 | |
*** gcb has quit IRC | 07:22 | |
*** gcb has joined #openstack-ironic | 07:22 | |
*** racedo has joined #openstack-ironic | 07:30 | |
*** marios has quit IRC | 07:34 | |
*** mgoddard has joined #openstack-ironic | 07:36 | |
*** ralonsoh has joined #openstack-ironic | 07:41 | |
Tengu | \o/ I haz GPT label on my compute nodes | 07:45 |
*** etingof has quit IRC | 07:54 | |
*** xuanjia has joined #openstack-ironic | 07:56 | |
*** udesale has quit IRC | 08:03 | |
*** wlfightup has joined #openstack-ironic | 08:04 | |
*** udesale has joined #openstack-ironic | 08:05 | |
openstackgerrit | Shivanand Tendulker proposed openstack/ironic-python-agent master: Fix to return 'root_uuid' as part of command status https://review.openstack.org/500719 | 08:07 |
openstackgerrit | Shivanand Tendulker proposed openstack/ironic master: Fix ``agent`` deploy interface to call ``boot.prepare_instance`` https://review.openstack.org/499050 | 08:10 |
*** marios has joined #openstack-ironic | 08:14 | |
*** marios_ has joined #openstack-ironic | 08:16 | |
*** priteau has joined #openstack-ironic | 08:24 | |
*** ChubYann has quit IRC | 08:26 | |
*** udesale has quit IRC | 08:27 | |
*** udesale has joined #openstack-ironic | 08:27 | |
*** marios_ has quit IRC | 08:28 | |
*** marios has quit IRC | 08:28 | |
*** marios has joined #openstack-ironic | 08:29 | |
openstackgerrit | Shivanand Tendulker proposed openstack/ironic master: Fix ``agent`` deploy interface to call ``boot.prepare_instance`` https://review.openstack.org/499050 | 08:29 |
*** rbartal has joined #openstack-ironic | 08:30 | |
*** lucas-afk is now known as lucasagomes | 08:32 | |
*** yangyapeng has quit IRC | 08:34 | |
*** yangyapeng has joined #openstack-ironic | 08:35 | |
*** ChubYann has joined #openstack-ironic | 08:40 | |
*** derekh has joined #openstack-ironic | 08:42 | |
*** dtantsur|afk is now known as dtantsur | 08:46 | |
dtantsur | Tengu: morning! it should be possible, I guess you have it now :) | 08:47 |
*** hshiina has quit IRC | 08:49 | |
Tengu | dtantsur: all is working fine now :) | 08:50 |
dtantsur | \o/ | 08:50 |
dtantsur | if you encountered any problems with our docs, please let me know | 08:51 |
dtantsur | GPT+BIOS is not something we test too often | 08:51 |
Tengu | dtantsur: guess it's my brain is a bit tired of all that tripleo thing ;) | 08:51 |
openstackgerrit | zhangbailin proposed openstack/ironic master: Update the url for ironic document https://review.openstack.org/500734 | 08:52 |
*** yangyapeng has quit IRC | 08:52 | |
Tengu | dtantsur: now I have another issue, but nothing related to ironic or even nova: how to share secrets, like x509 keypair between three controllers. Especially interesting as I want to use Let's Encrypt cert for Horizon/API endpoints. | 08:52 |
*** yangyapeng has joined #openstack-ironic | 08:53 | |
dtantsur | Tengu: there was some secret service in OpenStack, Barbican I think | 08:53 |
Tengu | I think I have a clean solution, have to investigate a bit more. | 08:53 |
Tengu | yep | 08:53 |
Tengu | but that one isn't set up in a secure way with tripleO | 08:53 |
Tengu | and, more over, there's a chicken-and-egg issue | 08:53 |
Tengu | ~> have to ensure barbican is up n'running in order to store SSL certificate, but if API are configured for SSL, Barbican will need SSL, but… | 08:54 |
dtantsur | ooops | 08:54 |
Tengu | yep. | 08:54 |
*** e0ne has joined #openstack-ironic | 08:54 | |
Tengu | thus, I think I can just add a service on the undercloud server, something like Custodia, and make a wrapper arround custodia and certbot. | 08:54 |
Tengu | something that will be installed on all controllers, and that will: check where's the VIP, if "master" check cert validity, if EOL request a new one and update custodia // if not "master" (no VIP), fetch the keypair from custodia | 08:55 |
Tengu | pretty sure this is the cleanest way to do that. And, probably, the safest. | 08:56 |
Tengu | more over, custodia can store data in freeIPA (Barbican can also do that, but it's not implemented in tripleO configuration) | 08:56 |
*** e0ne has quit IRC | 09:00 | |
*** amotoki__away is now known as amotoki | 09:07 | |
* dtantsur haven't heard about custodia | 09:10 | |
dtantsur | but overall sounds good, yeah | 09:10 |
Tengu | dtantsur: stumbled on it just today. reading doc seems to confirm it's actually a good idea. | 09:11 |
*** milan has joined #openstack-ironic | 09:11 | |
milan | morning Ironic! :) | 09:11 |
milan | #pixiesay -mflexing occupy coffee! | 09:11 |
PixieBoots | ᕙʕ⇀ᴥ⇀ʔᕗ: occupy coffee! | 09:11 |
dtantsur | yo milan! apparently, they've approved our visas Oo or at least they asked to come and pick them :) | 09:12 |
milan | dtantsur, awesome! :) | 09:13 |
milan | #pixiesay -mlikeaboss dtantsur is the new Berlin coffee agent :P | 09:13 |
PixieBoots | ʕ▀̿ᴥ▀̿ʔ: dtantsur is the new Berlin coffee agent :P | 09:13 |
dtantsur | hehe, wait, let us actually move there first :) | 09:14 |
dtantsur | I'm still a bit scared of the future process | 09:14 |
*** jaianshu has joined #openstack-ironic | 09:17 | |
milan | dtantsur, the Spice must flow, that's the only think that matters :D | 09:20 |
*** sambetts|afk is now known as sambetts | 09:20 | |
sambetts | Morning all | 09:21 |
*** e0ne has joined #openstack-ironic | 09:21 | |
milan | morning sambetts! Having an +2 A+ day? ;) | 09:21 |
sambetts | milan: heh, hope so ;) | 09:22 |
milan | :D sambetts great! https://review.openstack.org/#/c/487909/ | 09:23 |
patchbot | patch 487909 - ironic-inspector - Allow periodics to terminate inspector | 09:23 |
* milan fingers crossed :P | 09:23 | |
dtantsur | lol | 09:27 |
dtantsur | morning sambetts | 09:27 |
sambetts | o/ dtantsur | 09:29 |
openstackgerrit | zhangbailin proposed openstack/ironic master: Update the url for ironic document https://review.openstack.org/500734 | 09:29 |
openstackgerrit | Andrey Shestakov proposed openstack/ironic master: Change pxe dhcp options name to codes. https://review.openstack.org/499553 | 09:42 |
*** priteau has quit IRC | 09:51 | |
openstackgerrit | Merged openstack/ironic master: Put tests in correct location for ironic/api/controllers/v1/ https://review.openstack.org/500658 | 10:00 |
*** ankkumar has joined #openstack-ironic | 10:00 | |
dtantsur | milan: time to celebrate? ;) (re your patch +W) | 10:01 |
milan | dtantsur, OMG! :D | 10:01 |
milan | aarefiev_afk, sambetts, thanks for the reviews! :D | 10:02 |
dtantsur | milan: have you had lunch yet? | 10:02 |
sambetts | :) | 10:02 |
dtantsur | milan: we're planning East Village around soon | 10:02 |
milan | dtantsur, just a bout to go, but I'm in the office, so I guess I'll pass | 10:02 |
milan | (2h roundtrip) | 10:02 |
dtantsur | milan: it's only 15 mins by tram, c'mon! | 10:03 |
milan | hehe but there's a value added to the office --- I'm going to drop by the coffee shop to buy some beans! :D | 10:03 |
dtantsur | milan: nice! btw I'm going to the office afterwards | 10:04 |
milan | awesome! :) I'm very keen on sharing today! :D | 10:04 |
openstackgerrit | Merged openstack/ironic master: Troubleshooting docs: explain disabled compute services https://review.openstack.org/500542 | 10:05 |
dtantsur | okay, I'm heading towards food, back in an hour or so | 10:06 |
*** dtantsur is now known as dtantsur|lunch | 10:06 | |
*** ankkumar has quit IRC | 10:12 | |
vdrok | late morning all | 10:15 |
openstackgerrit | Merged openstack/ironic master: tox.ini: Add 'py36' to the default envlist https://review.openstack.org/498534 | 10:16 |
*** udesale__ has joined #openstack-ironic | 10:37 | |
*** udesale has quit IRC | 10:39 | |
*** udesale__ has quit IRC | 10:40 | |
*** udesale has joined #openstack-ironic | 10:40 | |
openstackgerrit | Shivanand Tendulker proposed openstack/ironic-python-agent master: Fix to return 'root_uuid' as part of command status https://review.openstack.org/500719 | 10:41 |
*** zshi_ has joined #openstack-ironic | 10:42 | |
*** yangyapeng has quit IRC | 10:43 | |
*** jkilpatr has joined #openstack-ironic | 10:59 | |
*** rpioso has quit IRC | 11:00 | |
*** jkilpatr_ has joined #openstack-ironic | 11:01 | |
*** jaganathan has quit IRC | 11:02 | |
*** jkilpatr has quit IRC | 11:05 | |
*** lucasagomes is now known as lucas-hungry | 11:05 | |
*** stendulker has quit IRC | 11:10 | |
*** jaianshu has quit IRC | 11:16 | |
*** zhangfei has quit IRC | 11:23 | |
*** Nisha_Agarwal has joined #openstack-ironic | 11:24 | |
*** Nisha_ has joined #openstack-ironic | 11:26 | |
*** priteau has joined #openstack-ironic | 11:26 | |
*** Nisha_Agarwal has quit IRC | 11:28 | |
*** fried_rice is now known as efried | 11:29 | |
*** fellypefca has joined #openstack-ironic | 11:36 | |
*** rcernin has quit IRC | 11:36 | |
*** rcernin has joined #openstack-ironic | 11:36 | |
fellypefca | Good morning, Ironic :) | 11:37 |
*** dprince has joined #openstack-ironic | 11:39 | |
*** Nisha_ has quit IRC | 11:41 | |
vdrok | good morning fellypefca | 11:47 |
*** rook|pto is now known as rook | 11:48 | |
fellypefca | vdrok :) | 11:49 |
openstackgerrit | Merged openstack/ironic-inspector master: Allow periodics to terminate inspector https://review.openstack.org/487909 | 11:49 |
openstackgerrit | Vladyslav Drok proposed openstack/ironic master: Revise deploy process documentation https://review.openstack.org/458867 | 11:50 |
*** dtantsur|lunch is now known as dtantsur | 11:57 | |
*** jaypipes has joined #openstack-ironic | 12:00 | |
*** wong_ has quit IRC | 12:04 | |
*** wong has joined #openstack-ironic | 12:04 | |
*** bfournie has quit IRC | 12:04 | |
*** bfournie has joined #openstack-ironic | 12:04 | |
*** mat128 has joined #openstack-ironic | 12:06 | |
*** bfournie has quit IRC | 12:09 | |
*** jlabarre has joined #openstack-ironic | 12:09 | |
*** TuanLA has quit IRC | 12:09 | |
*** litao__ has quit IRC | 12:13 | |
*** srobert has joined #openstack-ironic | 12:14 | |
*** srobert has quit IRC | 12:14 | |
*** srobert has joined #openstack-ironic | 12:15 | |
*** lucas-hungry is now known as lucasagomes | 12:18 | |
*** sauloaislan has joined #openstack-ironic | 12:20 | |
sauloaislan | Morning! | 12:20 |
fellypefca | Good morning sauloaislan | 12:23 |
*** udesale has quit IRC | 12:25 | |
*** yangyapeng has joined #openstack-ironic | 12:28 | |
*** mat128 has quit IRC | 12:33 | |
*** mat128 has joined #openstack-ironic | 12:34 | |
*** bfournie has joined #openstack-ironic | 12:49 | |
*** dyasny has joined #openstack-ironic | 12:50 | |
*** srobert_ has joined #openstack-ironic | 12:52 | |
*** srobert has quit IRC | 12:53 | |
*** baoli has joined #openstack-ironic | 12:54 | |
*** baoli has quit IRC | 12:54 | |
*** rloo has joined #openstack-ironic | 12:55 | |
*** zhurong has joined #openstack-ironic | 12:56 | |
*** persia has joined #openstack-ironic | 12:58 | |
*** baoli has joined #openstack-ironic | 13:00 | |
*** rbudden has joined #openstack-ironic | 13:00 | |
*** sasha21 has joined #openstack-ironic | 13:01 | |
*** Goneri has joined #openstack-ironic | 13:08 | |
*** zhurong has quit IRC | 13:10 | |
sauloaislan | dtantsur: Hi did you hade time to take a look at this backport? https://review.openstack.org/#/c/489598/ | 13:13 |
patchbot | patch 489598 - ironic (stable/ocata) - Removes agent mixin from oneview drivers | 13:13 |
dtantsur | sauloaislan: have your CI passed on it? | 13:13 |
sauloaislan | dtantsur: Yep | 13:13 |
*** srobert_ has quit IRC | 13:14 | |
*** srobert has joined #openstack-ironic | 13:15 | |
dtantsur | then I'll check it | 13:15 |
sauloaislan | dtantsur: Ok | 13:16 |
*** milan has quit IRC | 13:17 | |
dtantsur | sauloaislan: could you remind me the reasons we're backporting this? The release note only mentions a feature, the commit message only mentions *what* is done, but not *why*. | 13:17 |
*** phuongnh has quit IRC | 13:18 | |
sauloaislan | dtantsur We need this backport for Newton, because it can fix our agent job for stable/newton pacthes. And we need backport this patch first for Ocata, so we can backport this for newton.. | 13:22 |
dtantsur | rloo: morning! I've proposed a few document backports, mind checking them? https://review.openstack.org/#/q/status:open+project:openstack/ironic+branch:stable/pike | 13:22 |
rloo | hi dtantsur! looking... | 13:23 |
dtantsur | sauloaislan: why is it fixing a job in newton? is it also fixing it for ocata? | 13:23 |
sauloaislan | Ocata is ok, but my team ask you if we need to backport this patch for Ocata before newton, and you answer that we can't jump versions | 13:26 |
sauloaislan | So we are firts backport for Ocata to reach Newton | 13:26 |
*** xuanjia has quit IRC | 13:27 | |
sambetts | sauloaislan: if ocata is working, then why can't you backport from ocata to newton whatever made ocata work? | 13:27 |
sauloaislan | Ocata always works | 13:28 |
sauloaislan | What is not working is our agent job on CI for newton. And this patch can solve this issue. | 13:28 |
*** mrch has joined #openstack-ironic | 13:28 | |
dtantsur | sauloaislan: newton is EOL in IIRC 1 month. so the critical question is: is *the driver* broken, or is it only a problem with the CI? is the issue affecting end users, in other words? | 13:28 |
*** dsneddon has joined #openstack-ironic | 13:29 | |
dtantsur | yep, Newton EOL is 2017-10-11 | 13:29 |
dtantsur | if this issue does not affect end users, I think we can skip fixing it | 13:29 |
dtantsur | TheJulia, rloo, sambetts, wdyt ^^? | 13:29 |
sambetts | also if your agent job in ocata works why aren't you just backporting whatever merged in ocata to make it work ... | 13:30 |
sambetts | you shouldn't need something new if ocata is already working | 13:30 |
rloo | dtantsur: if it doesn't affect users, then it shouldn't go in | 13:31 |
rloo | dtantsur: do we insist on 3rd party CI working for stable/newton? | 13:31 |
sauloaislan | Not the driver, only the CI. But since it's end of life, ok. But we also have a bug with multitenant in the driver for Ocata and this backport can fix that too | 13:31 |
dtantsur | rloo: I don't. the branch will be deleted in October. | 13:32 |
rloo | dtantsur: ok, decided then. it shouldn't go in newton | 13:32 |
sauloaislan | Ok for us about newton.. | 13:32 |
dtantsur | sauloaislan: could you please confirm if this backport fixes the issue? if so, I'd like an LP bug and an update to the commit message. | 13:32 |
rloo | sauloaislan: you have a funny way of asking us to do things, i feel like you are throwing reasons one after the other after we knock a reason down... | 13:32 |
sauloaislan | rloo sorry, it's because our priority was fix our newton job. But we also have this bug on Ocata. | 13:34 |
sauloaislan | We will update the commit message with the bug.. sorry. | 13:34 |
rloo | dtantsur: question about release note for https://review.openstack.org/#/c/500545/ | 13:35 |
patchbot | patch 500545 - ironic (stable/pike) - Boot from volume fails with 'iscsi' deploy interface | 13:35 |
rloo | dtantsur: https://docs.openstack.org/releasenotes/ironic/unreleased.html | 13:36 |
rloo | dtantsur: it isn't clear to me whether something fails *after* the fix... | 13:36 |
rloo | dtantsur: "Boot from a volume fails when it is requested for a node ..." | 13:36 |
rloo | dtantsur: did it fail before, and the patch fixes it so it doesn't fail any more? | 13:36 |
rloo | dtantsur: oh, there are two release notes that have the same problem (for me) | 13:37 |
*** crushil has quit IRC | 13:37 | |
rloo | dtantsur: asking cuz you are backporting one of them. should the release note be fixed on master and also backported as separate backport patch? (am guessing yes while writing this) | 13:37 |
*** aparnav has joined #openstack-ironic | 13:38 | |
TheJulia | good morning | 13:38 |
dtantsur | rloo: yes, I think this is the way to do it: fix on master and backport | 13:38 |
rloo | dtantsur: thx for confirming! | 13:38 |
dtantsur | changing backported patches is an exceptional case | 13:38 |
dtantsur | morning TheJulia | 13:38 |
rloo | Good Monday Morning TheJulia, sauloaislan and everyone else! | 13:39 |
TheJulia | dtantsur: I'm good with leaving newton be unless the driver is broken in its present state. | 13:40 |
fellypefca | Good morning TheJulia, rloo, dtantsur :) | 13:40 |
rloo | morning fellypefca! | 13:40 |
TheJulia | so many people waking up | 13:41 |
dtantsur | :) | 13:41 |
dtantsur | TheJulia: it seems like the driver is fine, the CI is not. I wonder why, by the way.. | 13:41 |
TheJulia | dtantsur: my off the wall guess would be constraints | 13:42 |
TheJulia | well, upper constraints | 13:42 |
dtantsur | yeah, likely | 13:42 |
*** baoli_ has joined #openstack-ironic | 13:43 | |
*** srobert has quit IRC | 13:44 | |
*** trown is now known as trown|brb | 13:45 | |
*** baoli has quit IRC | 13:46 | |
*** crushil has joined #openstack-ironic | 13:47 | |
fellypefca | dtantsur:, rloo, TheJulia Sorry for the misunderstandings. We will formalize and open the bug that our driver presented in Ocata and update the commit message to make clear what we want with this backport | 13:48 |
fellypefca | dtantsur * | 13:48 |
rloo | thx fellypefca | 13:49 |
dtantsur | fellypefca: yes please! | 13:50 |
*** milan has joined #openstack-ironic | 13:51 | |
*** zshi_ has quit IRC | 13:51 | |
*** cdearborn has joined #openstack-ironic | 13:55 | |
*** jmorgan1 has joined #openstack-ironic | 13:57 | |
*** ijw has joined #openstack-ironic | 13:57 | |
*** trown|brb is now known as trown | 13:57 | |
*** etingof has joined #openstack-ironic | 14:00 | |
*** dtrainor has joined #openstack-ironic | 14:00 | |
vdrok | good morning rloo and TheJulia | 14:07 |
rloo | hi vdrok! | 14:08 |
vdrok | dtantsur: btw, regarding that multinode job failure on the whiteboard, the fix is at https://review.openstack.org/#/c/499545/2/nova/scheduler/utils.py but I'm pretty sure it won't be ever merged to nova. so wondering do we need to skip the test that's failing? or, as it happens not so often, just leave it and recheck when needed? | 14:09 |
patchbot | patch 499545 - nova - Allow reschedules for ironic computes if one force... | 14:09 |
*** srobert has joined #openstack-ironic | 14:10 | |
vdrok | what that test does is iterating through all compute hosts doing nova boot --availiability-zone nova:<host> | 14:11 |
openstackgerrit | Merged openstack/ironic master: Introduce keystoneauth adapters for clients https://review.openstack.org/469627 | 14:12 |
dtantsur | vdrok: why does it fail for us? because of cleaning? | 14:13 |
vdrok | dtantsur: because there is a gap between removal of allocation from placement (when node started cleaning and instance is deleted from nova), which basically frees the resources in placement, and resource tracker understanding that we need to set all resources to 0 | 14:14 |
vdrok | as RT is periodic | 14:14 |
vdrok | so if that failure happens ideally we should reschedule | 14:14 |
vdrok | or not let that happen :) | 14:15 |
openstackgerrit | Merged openstack/ironic master: Fix unit test for new fields in invaid API version https://review.openstack.org/468723 | 14:15 |
openstackgerrit | Merged openstack/ironic stable/pike: Fix indentation in few of the documentation pages https://review.openstack.org/500557 | 14:15 |
openstackgerrit | Merged openstack/ironic stable/pike: Update Nova configuration documentation https://review.openstack.org/500562 | 14:15 |
openstackgerrit | Merged openstack/ironic stable/pike: Minor improvements to the resource classes documentation https://review.openstack.org/500563 | 14:15 |
openstackgerrit | Merged openstack/python-ironicclient master: Updated from global requirements https://review.openstack.org/500023 | 14:16 |
milan | dtantsur, sambetts, guys, would be nice to merge https://review.openstack.org/#/c/471831 soon-ish to be able to get some time to correct any issues with it, wdyt? | 14:16 |
patchbot | patch 471831 - ironic-inspector - Refactoring the firewall | 14:16 |
dtantsur | vdrok: I like the "not let that happen" bit :) | 14:16 |
dtantsur | vdrok: we should ask nova folks to have virt drivers tell their availability, then we won't have this problem | 14:16 |
dtantsur | vdrok: like is_host_available(host) call, which will return False while node is not AVAILABLE | 14:16 |
* dtantsur has added this topic to the nova's PTG etherpad | 14:17 | |
vdrok | dtantsur: well, we'd need to have removal of allocation and that 'unavailability' update happening in a single placement request | 14:17 |
vdrok | there is such a call, just it happens only periodically :( | 14:18 |
dtantsur | vdrok: yeah, or we need a new notion of temporary unavailability in placement | 14:20 |
*** beekneemech is now known as bnemec | 14:25 | |
*** namnh has quit IRC | 14:26 | |
*** namnh has joined #openstack-ironic | 14:29 | |
openstackgerrit | Merged openstack/python-ironicclient stable/ocata: OSC 'node list' recognizes all provision states https://review.openstack.org/498413 | 14:37 |
*** links has quit IRC | 14:37 | |
*** mrch has quit IRC | 14:40 | |
*** karimb has quit IRC | 14:41 | |
*** udesale has joined #openstack-ironic | 14:44 | |
openstackgerrit | Ruby Loo proposed openstack/ironic master: [reno] Clarify fix for missing boot.prepare_instance https://review.openstack.org/500855 | 14:45 |
*** jtomasek has quit IRC | 14:45 | |
*** jtomasek has joined #openstack-ironic | 14:46 | |
*** jlabarre has quit IRC | 14:46 | |
*** udesale has quit IRC | 14:47 | |
*** marios has quit IRC | 14:47 | |
*** jheroux has joined #openstack-ironic | 14:47 | |
*** udesale has joined #openstack-ironic | 14:48 | |
*** marios has joined #openstack-ironic | 14:48 | |
*** rpioso has joined #openstack-ironic | 14:48 | |
rpioso | Good morning! | 14:49 |
*** udesale has quit IRC | 14:49 | |
fellypefca | Good morning rpioso | 14:50 |
rpioso | fellypefca: :) | 14:51 |
sauloaislan | Good morning rpioso | 14:51 |
rpioso | Hey sauloaislan! | 14:53 |
rpioso | I'm wondering if anyone has successfully used UEFI boot mode with the ironic.drivers.agent.AgentDeploy interface. | 14:54 |
dtantsur | I usually use iscsi deploy.. | 14:55 |
dtantsur | HPE folks might have tried it | 14:55 |
rpioso | dtantsur: ty | 14:56 |
vdrok | mornin rpioso | 14:57 |
rpioso | vdrok: Good morning! | 14:57 |
ricardoas | rpioso: not with oneview iirc | 14:57 |
rpioso | ricardoas: Thank you. I was hoping for a "Yes", though :( | 14:58 |
*** jlabarre has joined #openstack-ironic | 14:58 | |
dtantsur | rpioso: what difference do you expect between two deploy methods wrt UEFI? | 14:59 |
openstackgerrit | Sam Betts proposed openstack/ironic master: Switch API ref to use versionadded syntax throughout https://review.openstack.org/490008 | 14:59 |
rpioso | dtantsur: Hopefully none. | 14:59 |
TheJulia | rpioso: did you see the links I pasted into irc before you went to dinner last night? | 15:00 |
*** rbartal has quit IRC | 15:01 | |
rpioso | dtantsur: I'm scoping out the work to add UEFI boot mode support to drac. | 15:02 |
openstackgerrit | Ruby Loo proposed openstack/ironic master: [reno] Clarify fix for BFV & image_source https://review.openstack.org/500868 | 15:02 |
*** mjura has quit IRC | 15:03 | |
rpioso | TheJulia: I didn't at the time. Looking now ... | 15:03 |
TheJulia | rpioso: and I do remember, explicitly helping a internal team that had to use whole disk images with uefi because their network cards would not initialize in bios mode. | 15:03 |
TheJulia | well, internal team at hp | 15:04 |
rpioso | TheJulia: Now I see them. They were hidden after an IRC client disconnect message :( | 15:04 |
TheJulia | doh, sorry, I didn't think to tag you when I pasted them | 15:04 |
*** ccamacho has quit IRC | 15:06 | |
openstackgerrit | Sam Betts proposed openstack/ironic master: Add detail=True query string to API list endpoints https://review.openstack.org/431497 | 15:06 |
rpioso | TheJulia: I'm mistaken. The urls are the general ironic urls displayed on connection to our channel. | 15:07 |
rpioso | TheJulia: I didn't catch them. I was probably disconnected. | 15:07 |
TheJulia | ahh, it was before you showed as disconnected, but apparently I was having some network connectivity issues as well. You have the links or do you need them again? | 15:08 |
rpioso | TheJulia: Your recollection is encouraging. | 15:08 |
NobodyCam | Good Morning Ironic'ers :) | 15:08 |
fellypefca | NobodyCam o/ | 15:08 |
NobodyCam | :_ morning fellypefca | 15:08 |
rpioso | TheJulia: Please resend them if it's convenient. I could also search eavesdrop if you prefer. | 15:08 |
NobodyCam | s/_/)/ | 15:09 |
vdrok | good morning NobodyCam ! | 15:09 |
rpioso | Good morning, NobodyCam | 15:09 |
NobodyCam | hey hey morning vdrok and rpioso :) | 15:09 |
ricardoas | morning NobodyCam! | 15:09 |
NobodyCam | Nice short week this week :) | 15:09 |
NobodyCam | hey there morning ricardoas :) | 15:09 |
TheJulia | rpioso: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/ilo/boot.py#L331 and prepare_instance in that same file | 15:10 |
openstackgerrit | Yaguo Zhou proposed openstack/ironic master: Replace DbMigrationError with DBMigrationError https://review.openstack.org/500877 | 15:10 |
NobodyCam | morning TheJulia :) | 15:10 |
TheJulia | rpioso: the biggest issue, and I NobodyCam may remember this as well, was figuring out that the network cards would not initialize for pxe boot unles the hardware was in uefi boot mode | 15:11 |
TheJulia | err, I think NobodyCam | 15:11 |
openstackgerrit | Ruby Loo proposed openstack/ironic master: [install docs] ironic -> openstack baremetal CLI https://review.openstack.org/498970 | 15:12 |
dtantsur | morning NobodyCam | 15:12 |
NobodyCam | Morning dtantsur :) woo hoo it's not monday! :p | 15:12 |
dtantsur | \o/ | 15:12 |
dtantsur | even better news: | 15:12 |
rpioso | NobodyCam: lol | 15:12 |
dtantsur | TOMORROW IS NOT MONDAY EITHER!! | 15:12 |
NobodyCam | TheJulia: there was also a issue with the iLo not paying attention to set next boot device! but I came in, in the middle of the conversation | 15:13 |
rpioso | TheJulia: Thank you. I'll pull that link up shortly. I have a downstream meeting now. | 15:14 |
NobodyCam | hehehehee | 15:14 |
TheJulia | rpioso: The second issue was creation of a bootable image that was able to actually boot in uefi mode, because there were very specific hardware constraints that existed. Reading between the lines, it was very specialized gear. | 15:14 |
TheJulia | NobodyCam: oh, yeah, that was when they were trying to use ipmi in bios mode if memory serves | 15:14 |
rpioso | NobodyCam: I wrote, "I'm wondering if anyone has successfully used UEFI boot mode with the ironic.drivers.agent.AgentDeploy interface." | 15:14 |
NobodyCam | yep! | 15:14 |
NobodyCam | ah! | 15:15 |
* rpioso is in a downstream meeting | 15:15 | |
NobodyCam | :) enjoy :) | 15:15 |
*** e0ne has quit IRC | 15:32 | |
*** b4rti has quit IRC | 15:33 | |
*** rama_y has joined #openstack-ironic | 15:34 | |
*** dtrainor_ has joined #openstack-ironic | 15:34 | |
openstackgerrit | Dmitry Tantsur proposed openstack/python-ironicclient stable/pike: Skip warning when changing target_raid_config https://review.openstack.org/500893 | 15:35 |
*** baoli_ has quit IRC | 15:37 | |
fellypefca | sambetts could you take a look at the answer that ricardoas give to you? If you agreed I can make the changright now. :) https://review.openstack.org/#/c/468129/ | 15:38 |
patchbot | patch 468129 - ironic - Introduce hpOneView and ilorest to OneView | 15:38 |
*** dtrainor has quit IRC | 15:38 | |
*** baoli has joined #openstack-ironic | 15:38 | |
fellypefca | changes* | 15:39 |
openstackgerrit | Ilya Etingof proposed openstack/ironic-specs master: IPA to report node activity https://review.openstack.org/492702 | 15:41 |
openstackgerrit | Merged openstack/ironic stable/pike: Boot from volume fails with 'iscsi' deploy interface https://review.openstack.org/500545 | 15:42 |
*** lucasagomes is now known as lucas-afk | 15:46 | |
*** amarao has quit IRC | 15:48 | |
*** aparnav has quit IRC | 15:48 | |
rpioso | TheJulia: I've looked at that code in the ilo boot interface. | 15:49 |
openstackgerrit | Merged openstack/ironic master: [docs] Update Releasing Ironic Projects https://review.openstack.org/497536 | 15:54 |
dtantsur | folks, I need a backup for the PTG "Rethinking the vendor passthru API" topic please | 15:54 |
dtantsur | I wanted to talk about vendor passthru after the driver composition | 15:54 |
dtantsur | pas-ha, rloo ^^^ | 15:55 |
dtantsur | it may conflict with the nova discussion | 15:55 |
dtantsur | which is roughly scheduled to happen 3pm - 4pm on Wed | 15:55 |
dtantsur | vdrok: ^^^ | 15:55 |
*** stendulker has joined #openstack-ironic | 15:55 | |
rloo | dtantsur: i am not sure i even know what that topic is about, sec, let me look at the pad... | 15:56 |
rivers | Hi, does the deployment image starts the ironic python agent as a process on the baremetal node , prior to a lookup ? | 15:56 |
sambetts | rivers: yes | 15:56 |
sambetts | rivers: the first thing the deployment image does post-boot is start the ironic-python-agent | 15:57 |
rloo | dtantsur: that nova baremetal scheduling discussion is important. we should all be there? | 15:57 |
vdrok | dtantsur: I think I'll be on the nova thing as well | 15:57 |
rloo | dtantsur: I am reminded of the the neutron-ironic sessions where it seemed like all the neutron folks were there :) | 15:57 |
vdrok | if that owner-aware spec is going to be there | 15:57 |
dtantsur | rloo: I'm not sure we need to bring really everyone.. | 15:58 |
rloo | dtantsur: oh, we don't. i just thought i'd be funny if we did :) | 15:59 |
*** karimb has joined #openstack-ironic | 15:59 | |
dtantsur | wellll :) | 15:59 |
dtantsur | I'd prefer 2-4 people coming to nova, everyone else discussing e.g. clean steps API | 15:59 |
dtantsur | we don't have too much time | 15:59 |
*** etingof has quit IRC | 16:00 | |
dtantsur | pas-ha: around? | 16:01 |
rloo | dtantsur: wrt the vendor passthru, yeah, wanted to know ^^ | 16:01 |
dtantsur | rloo: the details are at line 38 of https://etherpad.openstack.org/p/ironic-queens-ptg-contentious-topics | 16:01 |
* pas-ha is AFK on mobile for couple if days | 16:01 | |
dtantsur | I think pas-ha was the only complaining about our vendor passthu | 16:01 |
rloo | dtantsur: looking at that now. I didn't realize it was even an issue :-( | 16:02 |
pas-ha | yep, I was | 16:02 |
rloo | dtantsur: would be nice to have a spec up first, that mentions the problem, and possible solutions? | 16:02 |
dtantsur | pas-ha: mind leading this topic if I get stuck in the nova room? | 16:02 |
pas-ha | Ok | 16:02 |
dtantsur | I'll put all my thougts to the etherpad | 16:02 |
dtantsur | thanks pas-ha! | 16:02 |
rloo | thx pas-ha! | 16:02 |
pas-ha | Cool, thx | 16:02 |
pas-ha | If we ever get to Denver in the first place - what's current situation in Houston? We ought to fly in US thru it.. | 16:03 |
*** hamzy has quit IRC | 16:04 | |
pas-ha | Any US folks have info on that? Sorry, wasn't following this topic lately | 16:04 |
rpioso | TheJulia: I understand that ISCSIDeploy and AgentDeploy call the two (2) methods in the boot interface that you pointed out, prepare_instance and prepare_node_for_deploy (indirectly by way of IloPXEBoot.prepare_ramdisk). | 16:05 |
dtantsur | folks, I have to call it a day, as I have to wake up super-early tomorrow | 16:06 |
TheJulia | rpioso: hopefully that helps explain how | 16:06 |
TheJulia | dtantsur: have a good evening | 16:06 |
dtantsur | meaning, I'll be out tomorrow morning | 16:06 |
dtantsur | I'm not sure when I'll come back, please bear with me ;) | 16:06 |
dtantsur | please keep reviewing the specs and adding your notes to PTG etherpads | 16:07 |
TheJulia | pas-ha: as I understand it, I think passenger service was supposed to resume this past Sunday, but your airline should have most current information | 16:07 |
rpioso | TheJulia: My question is about the differences between how ISCSIDeploy and AgentDeploy invoke those boot interface methods. | 16:07 |
*** dtantsur is now known as dtantsur|afk | 16:07 | |
pas-ha | TheJulia: thanks, we'll check it out | 16:08 |
*** Nisha_Agarwal has joined #openstack-ironic | 16:09 | |
*** marios has quit IRC | 16:09 | |
TheJulia | rpioso: not really sure I get exactly what your seeking to understand at this point, since each has slightly specific mechanics, although that can likely be unwound at this point | 16:10 |
vdrok | pas-ha: I guess here is the info https://www.facebook.com/flyiah/posts/973498529457002 | 16:11 |
vdrok | so international flights should be OK by the weekend | 16:11 |
rpioso | TheJulia: Both of their deploy() methods interact with the PXEBoot inteface in the same way. | 16:12 |
stendulker | TheJulia: Hi | 16:13 |
rpioso | TheJulia: That's not the case with their prepare() methods. | 16:13 |
TheJulia | vdrok: that was posted last week, so might already be good, but yeah, specific airline | 16:13 |
vdrok | ah, right | 16:13 |
TheJulia | rpioso: prepare_ramdisk vs prepare_instance? | 16:13 |
stendulker | TheJulia: A question on BFV. | 16:13 |
TheJulia | o/ stendulker | 16:13 |
TheJulia | stendulker: will do my best, but also having like 6 discussions right now :) | 16:14 |
stendulker | TheJulia: BFV works with cirros image, but not with Fedora or Ubuntu | 16:14 |
stendulker | TheJulia: Any thing additionally needs to be done in the images before cerating its cinder volume? | 16:14 |
TheJulia | hmmmm | 16:15 |
stendulker | TheJulia: It does work if we hardcode the iscsi initaitor info into the volume | 16:15 |
rpioso | TheJulia: I'm referring to ISCSIDeploy.prepare() vs AgentDeploy.prepare() wrt to their calls of boot.prepare_instance() and boot.prepare_ramdisk(). | 16:15 |
TheJulia | stendulker: very likely bootloader specific/ramdisk specific that gets constructed in the image. I had successful reports of ?fedora? I think, but I'm trying to remember who reported that. | 16:15 |
stendulker | But thts not expected, and cannot be done | 16:15 |
*** dtrainor_ has quit IRC | 16:16 | |
*** dtrainor_ has joined #openstack-ironic | 16:16 | |
TheJulia | stendulker: so booting the volume from a vm, then editing its contents, and then booting the same volume on baremetal? | 16:16 |
stendulker | yes these images have grub2 in it and somehow grub is faliling to mount the root patition. I doubt it at all attempts iscsi | 16:16 |
stendulker | no | 16:17 |
TheJulia | stendulker: got a kernel log from the console handy? | 16:17 |
stendulker | have whole disk images that are being used for hdd | 16:17 |
stendulker | not much of info over there.. | 16:19 |
TheJulia | rpioso: so, if I'm understanding it correctly, some of that is an artifact of default netboot instead of default boot to local disk. The iscsi deploy driver was originally always netboot based, where as agent was always local boot | 16:19 |
stendulker | do nota have access to BM right now | 16:19 |
stendulker | will capture the data | 16:19 |
stendulker | but there isn't anything special needs to done in image, right? | 16:19 |
TheJulia | stendulker: okay, we definitely need more info, likely it might need to be something dib related for ipxe based boots | 16:19 |
TheJulia | possibly, we're missing mechanics information that we need to understand in order to properly understand what is occurring or might be missing from the puzzle in the images. | 16:20 |
stendulker | rpioso: There is a bug in AgentDeploy | 16:20 |
TheJulia | stendulker: a bug? | 16:20 |
stendulker | rpioso: AgentDeploy always assumes disk boot with whole disk image | 16:21 |
stendulker | TheJulia: Yes | 16:21 |
openstackgerrit | Sam Betts proposed openstack/ironic-python-agent master: [LLDP] Skip NICs that say they are ready but are unreadable. https://review.openstack.org/434385 | 16:21 |
TheJulia | stendulker: by original design though, so i personally wouldn't call it a bug :) | 16:21 |
stendulker | TheJulia: It does not call prepare_instance() for whole disk image | 16:21 |
*** dtrainor_ has quit IRC | 16:21 | |
TheJulia | stendulker: which it should not | 16:21 |
stendulker | TheJulia, rpioso : https://review.openstack.org/#/c/499050/ | 16:21 |
patchbot | patch 499050 - ironic - Fix ``agent`` deploy interface to call ``boot.prep... | 16:21 |
*** dtrainor_ has joined #openstack-ironic | 16:22 | |
TheJulia | at least, should not based on original use model/behavior | 16:22 |
stendulker | TheJulia: It should call, otherwise drivers cannot do post install settings | 16:22 |
stendulker | TheJulia: Ex. If one wants to change the boot mode or any pther settings , its not possible | 16:22 |
stendulker | s/pther/other | 16:23 |
stendulker | Also agentDeploy fails with PXEBoot interface even for IPMI drivers | 16:23 |
stendulker | We have IPMI drivers based on iscsi and not agent so this defect was hidden | 16:24 |
rpioso | stendulker: Thank you for informing us of that bug :) | 16:24 |
rpioso | TheJulia, stendulker: It appears different from my concern, though. | 16:25 |
vdrok | stendulker: hrm, we do have jobs with agent_ipmitool driver, which is pxeboot+agentdeploy | 16:26 |
vdrok | and we boot the whole disk image there | 16:26 |
rpioso | TheJulia, stendulker: ISCSIDeploy.prepare() calls boot.prepare_instance() if the provision state is active or adopting. But ISCSIDeploy.prepare() calls boot.prepare_instance() only if the provision state is active. | 16:26 |
stendulker | vdrok: there is adifefrence | 16:26 |
vdrok | but yeah I see the problem around the post-deploy config you've described in the bug | 16:26 |
stendulker | iSCSI+PXE+WholeDisk= Disk boot | 16:27 |
*** etingof has joined #openstack-ironic | 16:27 | |
stendulker | sorry | 16:27 |
openstackgerrit | Sam Betts proposed openstack/ironic master: Switch API ref to use versionadded syntax throughout https://review.openstack.org/490008 | 16:27 |
stendulker | iSCSI+PXE+WholeDisk= PXE boot | 16:27 |
stendulker | AGENt+PXE+WholeDisk= DISK boot | 16:27 |
stendulker | The boot devices used differ | 16:28 |
vdrok | (1st we don't test I think) | 16:28 |
vdrok | yup | 16:28 |
rpioso | TheJulia, stendulker: Why are their provision state conditions on calling boot.prepare_instance() different? | 16:28 |
stendulker | when no boot_option is specified. i.e boot_option of netboot would be used | 16:28 |
stendulker | rpioso: boot.prepare_instance() is called once image is provisioned and before instance is powere on | 16:30 |
stendulker | rpioso: Can you point the line nos in the code where you see the difference | 16:32 |
rpioso | stendulker: ISCSIDeploy.prepare() - https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/iscsi_deploy.py#L519 | 16:35 |
rpioso | stendulker: AgentDeploy.prepare() - https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent.py#L460 | 16:36 |
TheJulia | which rreminds me, I need to work on the patch to change the default boot option | 16:36 |
stendulker | TheJulia: What is the patch about boot option? | 16:37 |
rpioso | stendulker: ISCSIDeploy.prepare() - line nos 519-520 | 16:37 |
TheJulia | rpioso: adopt state conditional is to create the configuration of an instance for network booting if configured | 16:37 |
rpioso | stendulker: AgentDeploy.prepare() - line nos 460-462 | 16:38 |
TheJulia | rpioso: same does not exist with agent deployment | 16:38 |
rpioso | TheJulia: Thank you so much for explaining that difference :) | 16:38 |
stendulker | yes, both has it | 16:38 |
TheJulia | stendulker: https://review.openstack.org/#/c/492114/ | 16:38 |
patchbot | patch 492114 - ironic - DNM: change default_boot_option to local | 16:38 |
*** sauloaislan has quit IRC | 16:39 | |
TheJulia | noted throughout that we intended to change it and that it would change in the future, we just never changed it | 16:39 |
stendulker | TheJulia: ok, thanks | 16:39 |
*** ijw has quit IRC | 16:40 | |
rpioso | TheJulia, stendulker: Thanks for your assistance. | 16:41 |
* rpioso is hungry | 16:41 | |
stendulker | rpioso: np, welcome | 16:41 |
*** rpioso is now known as rpioso|eat | 16:42 | |
* TheJulia thinks it is time for lunch | 16:43 | |
*** mgoddard has quit IRC | 16:45 | |
rloo | hey, who understands dbs and transactions? | 16:47 |
*** dtrainor_ has quit IRC | 16:48 | |
TheJulia | i do | 16:49 |
TheJulia | why? | 16:49 |
rloo | TheJulia: https://review.openstack.org/#/c/497666/1/ironic/db/sqlalchemy/api.py | 16:51 |
patchbot | patch 497666 - ironic - Add online data migrations for conductor version | 16:51 |
rloo | TheJulia: I don't understand the _session_for_write() stuff. And why it is an issue in this particular method (backfill_version_column()) and not in other methods that use that same _session_for_write() | 16:52 |
openstackgerrit | Anup Navare proposed openstack/ironic-inspector master: Improves documentation for discovery https://review.openstack.org/477640 | 16:56 |
*** namnh has quit IRC | 16:56 | |
*** derekh has quit IRC | 17:00 | |
*** rpioso|eat has quit IRC | 17:02 | |
*** ralonsoh has quit IRC | 17:02 | |
*** dtrainor_ has joined #openstack-ironic | 17:03 | |
*** stendulker has quit IRC | 17:03 | |
*** dtrainor_ has quit IRC | 17:07 | |
*** etingof has quit IRC | 17:08 | |
*** dtrainor_ has joined #openstack-ironic | 17:08 | |
*** harlowja has joined #openstack-ironic | 17:08 | |
*** etingof has joined #openstack-ironic | 17:08 | |
TheJulia | because, in theory, a conductor could update the record representing it's self mid-transaction if it is saving the entire row, which I bet slqalchemy does. This is likely less of an issue with ndodes since we don't update entire rows often | 17:09 |
*** rpioso|eat has joined #openstack-ironic | 17:10 | |
TheJulia | rloo: there is also a layer to this that is added by sqlalchemy | 17:11 |
rloo | TheJulia: db stuff is foreign to me. sigh. | 17:13 |
*** penick has joined #openstack-ironic | 17:16 | |
rloo | TheJulia: so is this incorrect too: https://github.com/openstack/ironic/blob/master/ironic/db/sqlalchemy/api.py#L786-L790 | 17:17 |
rloo | TheJulia: it gets the nodes that are reserved for that conductor; then it clears them. What if a node got reserved by a diff conductor in the meantime? | 17:18 |
TheJulia | then the write fails, and a deadlock condition occurs | 17:19 |
TheJulia | that at least has the decorator to retry | 17:19 |
*** dtrainor_ has quit IRC | 17:20 | |
TheJulia | so, a migration that could be control-c'ed shoud be okay to hit a deadlock as long as it can be restarted. a transaction would just rollback a change if a deadlock occurs | 17:21 |
*** hamzy has joined #openstack-ironic | 17:21 | |
* sambetts --> home | 17:22 | |
sambetts | Night all! | 17:22 |
sambetts | cya tomorrow | 17:22 |
TheJulia | goodnight sam | 17:22 |
*** sambetts is now known as sambetts|afk | 17:22 | |
rloo | TheJulia: oh, by 'reserved' I don't mean a DB lock. I mean the 'reservation' field for Node1 had been set to 'hostname', but got changed to 'fooname', then L790 updates node1.reservation to None? | 17:22 |
sambetts|afk | cya TheJulia, rloo | 17:23 |
rloo | have a good evening sambetts|afk! | 17:23 |
*** xuanjia has joined #openstack-ironic | 17:23 | |
rloo | TheJulia: I *think* that's a similar issue to what dtantsur|afk was referring to? | 17:23 |
TheJulia | so we're in database dependent behavior here | 17:24 |
*** tesseract has quit IRC | 17:25 | |
TheJulia | some when you pull a transaction give you a static view upon your selects as of when you start the transaction, some give you the latest data and just stack your changes for write out to the db once the transaction is exited. | 17:25 |
*** rcernin has quit IRC | 17:25 | |
*** Nisha_Agarwal has quit IRC | 17:25 | |
*** rama_y has quit IRC | 17:26 | |
TheJulia | rloo: in the case where another conductor grabbed the record, the transaction should fail, but oslo_db will retry because the database knows it was a deadlock | 17:27 |
*** ijw has joined #openstack-ironic | 17:27 | |
rloo | TheJulia: so to be safe, do what dtantsur|afk suggests? | 17:27 |
TheJulia | query yes, session, I'm not convinced. I guess we should for consistenccy and retry behavior which we get then | 17:28 |
TheJulia | as long as the method is decorated | 17:28 |
TheJulia | rloo: the difference, in my mind, is fast fail or retry and wait for the db to become happy again. | 17:29 |
rloo | TheJulia: i just want to make sure that method works ! | 17:30 |
TheJulia | :) | 17:31 |
rloo | thx TheJulia! | 17:31 |
TheJulia | Thinking about it, every time I saw a java connector raise deadlock exceptions to the surface, I was a sad panda, so better to have the mechanics in place. | 17:33 |
*** dsneddon has quit IRC | 17:36 | |
jlvillal | TheJulia: I'll point you at this stable patch I noticed: https://review.openstack.org/#/c/498415/ | 17:43 |
patchbot | patch 498415 - python-ironicclient (stable/newton) - OSC 'node list' recognizes all provision states | 17:43 |
jlvillal | TheJulia: You are the only one who can approve it :) | 17:43 |
*** dsneddon has joined #openstack-ironic | 17:44 | |
TheJulia | \o/ | 17:44 |
*** trown is now known as trown|lunch | 17:45 | |
openstackgerrit | Fellype Cavalcante proposed openstack/ironic master: Introduce hpOneView and ilorest to OneView https://review.openstack.org/468129 | 17:53 |
rloo | TheJulia: you good if we +A this? https://review.openstack.org/#/c/499819/ | 17:55 |
patchbot | patch 499819 - python-ironicclient (stable/pike) - Allow OS_BAREMETAL_API_VERSION=latest to work | 17:55 |
TheJulia | i am | 17:58 |
rloo | TheJulia: thx :) | 17:59 |
*** baoli has quit IRC | 18:00 | |
*** baoli has joined #openstack-ironic | 18:01 | |
*** srobert_ has joined #openstack-ironic | 18:02 | |
*** srobert_ has quit IRC | 18:03 | |
*** gmonteiro has joined #openstack-ironic | 18:03 | |
*** srobert_ has joined #openstack-ironic | 18:03 | |
*** exodusftw has quit IRC | 18:05 | |
*** srobert has quit IRC | 18:05 | |
*** karthiks has quit IRC | 18:06 | |
*** Nisha_Agarwal has joined #openstack-ironic | 18:06 | |
openstackgerrit | Anup Navare proposed openstack/ironic-ui master: Allow selecting interfaces while enrolling nodes https://review.openstack.org/473957 | 18:10 |
fellypefca | TheJulia Did you agree with this that we are discussing with sambetts|afk ? https://review.openstack.org/#/c/468129 | 18:11 |
patchbot | patch 468129 - ironic - Introduce hpOneView and ilorest to OneView | 18:11 |
*** srobert_ has quit IRC | 18:11 | |
*** srobert has joined #openstack-ironic | 18:12 | |
TheJulia | fellypefca: I do kind of agree. I could see sam's point though in that if I added a load balancer behind another path, but... unlikely. I've pulled up the API before and I think it calls for absolute paths off root too, but it has been a while | 18:16 |
fellypefca | TheJulia, yeah. It calls for absolute paths off root. We prefer to keep this without talk about the format of manager_url, and only in the last patch of this migration(that we will remove all python-oneviewclient), we will talk about this. Make sense? | 18:19 |
*** chas has quit IRC | 18:20 | |
fellypefca | TheJulia Did you think that we can have this patch merged by friday? We need this before the PTG(also my vocations :D), and I need to solve that until next week. | 18:21 |
TheJulia | I can't guarentee merged, but let me get a drink and I'll start reviewing them | 18:21 |
fellypefca | TheJulia sure! thanks :D | 18:22 |
*** karthiks has joined #openstack-ironic | 18:22 | |
openstackgerrit | Merged openstack/python-ironicclient stable/newton: OSC 'node list' recognizes all provision states https://review.openstack.org/498415 | 18:24 |
*** Nisha_Agarwal has quit IRC | 18:33 | |
*** crushil has quit IRC | 18:34 | |
openstackgerrit | Merged openstack/python-ironicclient stable/ocata: Updated from global requirements https://review.openstack.org/464861 | 18:35 |
rpioso|eat | I would appreciate a second +2 and workflow on https://review.openstack.org/#/c/498412/. It's a drac bug fix that dtantsur cherry picked to stable/ocata. | 18:35 |
patchbot | patch 498412 - ironic (stable/ocata) - Fix DRAC classic driver double manage/provide | 18:35 |
*** rpioso|eat is now known as rpioso | 18:36 | |
*** ijw has quit IRC | 18:41 | |
*** crushil has joined #openstack-ironic | 18:41 | |
openstackgerrit | Merged openstack/python-ironicclient stable/pike: Allow OS_BAREMETAL_API_VERSION=latest to work https://review.openstack.org/499819 | 18:44 |
TheJulia | rpioso: ack | 18:45 |
openstackgerrit | Ruby Loo proposed openstack/ironic master: Fix race condition in backfill_version_column() https://review.openstack.org/500955 | 18:45 |
*** ijw has joined #openstack-ironic | 18:46 | |
*** pcaruana has quit IRC | 18:49 | |
*** fellypefca is now known as fellypefca-afk | 18:50 | |
fellypefca-afk | Good night everyone :) | 18:50 |
TheJulia | fellypefca-afk: goodnight | 18:50 |
TheJulia | about to leave a review on the first patch, reno needs a little work | 18:51 |
TheJulia | which I feel bad about, but... | 18:51 |
fellypefca-afk | TheJulia Oh, I need to update the release note with this new way to talk about the path! | 18:53 |
rpioso | TheJulia: That would be awesome. Thank you in advance :) | 18:53 |
fellypefca-afk | TheJulia sounds better to not mention the hardcoded port o reno? It was a suggestion of yuriyz | 18:54 |
fellypefca-afk | on reno* | 18:54 |
TheJulia | yeah... unsure :( | 18:55 |
fellypefca-afk | TheJulia I think that since we will not mention the patch on our Help, we need to delete from the reno too, right? | 18:57 |
TheJulia | you mean port? | 18:57 |
fellypefca-afk | Both | 18:57 |
TheJulia | yeah | 18:57 |
fellypefca-afk | Ok. I will change now | 18:58 |
*** ijw has quit IRC | 18:58 | |
*** penick has quit IRC | 18:58 | |
*** penick has joined #openstack-ironic | 18:59 | |
openstackgerrit | Fellype Cavalcante proposed openstack/ironic master: Introduce hpOneView and ilorest to OneView https://review.openstack.org/468129 | 19:02 |
fellypefca-afk | TheJulia ^ I think the reno looks better in this format. | 19:03 |
*** yangyapeng has quit IRC | 19:04 | |
fellypefca-afk | Now I can go heh Good Night and thank you again TheJulia | 19:12 |
TheJulia | goodnight | 19:12 |
*** priteau has quit IRC | 19:15 | |
openstackgerrit | Merged openstack/ironic master: Switch API ref to use versionadded syntax throughout https://review.openstack.org/490008 | 19:17 |
rpioso | fellypefca-afk: Thank you for the +1 on that cherry picked bug fix. | 19:22 |
*** [1]rpioso has joined #openstack-ironic | 19:24 | |
*** jtomasek has quit IRC | 19:26 | |
* TheJulia needs a vacation | 19:27 | |
*** rpioso has quit IRC | 19:28 | |
*** priteau has joined #openstack-ironic | 19:37 | |
*** e0ne has joined #openstack-ironic | 19:41 | |
*** priteau has quit IRC | 19:41 | |
*** trown|lunch is now known as trown | 19:42 | |
*** ijw has joined #openstack-ironic | 19:45 | |
*** mat128 has quit IRC | 19:51 | |
*** mat128 has joined #openstack-ironic | 19:51 | |
*** mat128 has quit IRC | 19:52 | |
*** mat128 has joined #openstack-ironic | 19:53 | |
NobodyCam | TheJulia: you are in CO? now | 19:57 |
TheJulia | yeah | 19:57 |
NobodyCam | might I suggest https://www.google.com/maps/place/Chris+Park+Rd,+Durango,+CO+81301/@37.5236224,-107.8124101,17z/data=!3m1!4b1!4m5!3m4!1s0x873ebbcd1d0cbecd:0x152ad7e864f2485e!8m2!3d37.5236224!4d-107.8102214 for a vacation | 19:57 |
*** [1]rpioso has quit IRC | 19:58 | |
TheJulia | heh, might pass through there to/from ABQ at some point | 19:58 |
TheJulia | since 550 | 19:59 |
NobodyCam | lol that link didn't really have any info on it. :p https://www.recreation.gov/camping/chris-park/r/campgroundDetails.do?contractCode=NRSO&parkId=70056 | 19:59 |
*** priteau has joined #openstack-ironic | 19:59 | |
TheJulia | heh | 20:00 |
NobodyCam | :p | 20:00 |
NobodyCam | there literally nothing around that park. If I recall no telco service / no wifi / no connection to outside world | 20:01 |
*** yangyapeng has joined #openstack-ironic | 20:01 | |
TheJulia | I really enjoyed my last vacation to Maui. I jumped in the sea a few times, sat on the condo's patio and just drank Kona Brewing's IPA | 20:02 |
*** ijw has quit IRC | 20:05 | |
*** ijw has joined #openstack-ironic | 20:05 | |
openstackgerrit | Ruby Loo proposed openstack/ironic master: [doc] Non-word updates to releasing doc https://review.openstack.org/500967 | 20:06 |
TheJulia | rloo: moar red? | 20:06 |
* TheJulia thinks of just calling it a day | 20:06 | |
rloo | TheJulia: hey, someone asked for it... ;) | 20:07 |
rloo | TheJulia: call it a day! | 20:07 |
*** srobert has quit IRC | 20:15 | |
*** srobert has joined #openstack-ironic | 20:15 | |
*** efried is now known as efried_bbiab | 20:15 | |
TheJulia | yeah, going to go back to hacking on client library code in the morning | 20:20 |
*** rama_y has joined #openstack-ironic | 20:21 | |
*** mat128 has quit IRC | 20:22 | |
*** mat128 has joined #openstack-ironic | 20:23 | |
openstackgerrit | Ilya Etingof proposed openstack/ironic-python-agent master: Report /dev/disk/by-path on inspection https://review.openstack.org/498489 | 20:23 |
*** ijw has quit IRC | 20:24 | |
*** ijw has joined #openstack-ironic | 20:24 | |
*** jkilpatr_ has quit IRC | 20:25 | |
*** [1]rpioso has joined #openstack-ironic | 20:29 | |
*** e0ne has quit IRC | 20:30 | |
openstackgerrit | Ruby Loo proposed openstack/ironic master: [reno] Clarify fix for missing boot.prepare_instance https://review.openstack.org/500855 | 20:34 |
openstackgerrit | Ilya Etingof proposed openstack/ironic-lib master: Added `by_path` root device hint https://review.openstack.org/500524 | 20:34 |
*** rama_y has quit IRC | 20:39 | |
*** jkilpatr has joined #openstack-ironic | 20:41 | |
openstackgerrit | Merged openstack/ironic-specs master: Updates to the snmp-power-driver spec https://review.openstack.org/493978 | 20:41 |
*** rcernin has joined #openstack-ironic | 20:44 | |
openstackgerrit | Ilya Etingof proposed openstack/ironic-lib master: Added `by_path` root device hint https://review.openstack.org/500524 | 20:47 |
*** efried_bbiab is now known as efried | 20:52 | |
*** trown is now known as trown|outtypewww | 20:59 | |
*** mat128 has quit IRC | 21:01 | |
*** [1]rpioso has quit IRC | 21:01 | |
*** srobert has quit IRC | 21:01 | |
*** sasha2 has joined #openstack-ironic | 21:03 | |
*** jheroux has quit IRC | 21:04 | |
*** sasha2 has quit IRC | 21:04 | |
*** crushil has quit IRC | 21:05 | |
*** sasha2 has joined #openstack-ironic | 21:06 | |
*** penick has quit IRC | 21:07 | |
openstackgerrit | Merged openstack/python-ironicclient stable/pike: Skip warning when changing target_raid_config https://review.openstack.org/500893 | 21:10 |
*** harlowja has quit IRC | 21:13 | |
*** sasha21 has quit IRC | 21:15 | |
*** ijw has quit IRC | 21:17 | |
*** bfournie has quit IRC | 21:22 | |
openstackgerrit | Merged openstack/python-ironicclient master: tox.ini: Add 'py36' to the default envlist https://review.openstack.org/498610 | 21:27 |
*** sasha2 has quit IRC | 21:32 | |
*** sasha2 has joined #openstack-ironic | 21:32 | |
openstackgerrit | Merged openstack/ironic stable/ocata: Fix DRAC classic driver double manage/provide https://review.openstack.org/498412 | 21:37 |
openstackgerrit | Merged openstack/ironic master: Update documentation for ``ilo`` hardware type https://review.openstack.org/496480 | 21:39 |
*** crushil has joined #openstack-ironic | 21:43 | |
*** Goneri has quit IRC | 21:49 | |
*** rcernin has quit IRC | 21:53 | |
*** karimb has quit IRC | 21:56 | |
*** quakers has quit IRC | 22:03 | |
*** quakers has joined #openstack-ironic | 22:06 | |
*** quakers is now known as ddurst | 22:06 | |
*** [1]rpioso has joined #openstack-ironic | 22:07 | |
*** ddurst has quit IRC | 22:07 | |
*** ddurst has joined #openstack-ironic | 22:08 | |
*** ijw has joined #openstack-ironic | 22:13 | |
*** priteau has quit IRC | 22:15 | |
*** ijw has quit IRC | 22:21 | |
*** baoli has quit IRC | 22:30 | |
*** dprince has quit IRC | 22:42 | |
*** rpioso has joined #openstack-ironic | 22:45 | |
*** [1]rpioso has quit IRC | 22:45 | |
*** cdearborn has quit IRC | 23:05 | |
*** efried is now known as efried_zzz | 23:07 | |
*** rloo has quit IRC | 23:08 | |
*** bfournie has joined #openstack-ironic | 23:10 | |
*** bfournie has quit IRC | 23:12 | |
*** gmonteiro has quit IRC | 23:13 | |
openstackgerrit | Ilya Etingof proposed openstack/ironic-python-agent master: Report /dev/disk/by-path on inspection https://review.openstack.org/498489 | 23:33 |
* rpioso heads home | 23:36 | |
*** rpioso is now known as rpioso|afk | 23:36 | |
openstackgerrit | Ilya Etingof proposed openstack/ironic-python-agent master: Report /dev/disk/by-path on inspection https://review.openstack.org/498489 | 23:40 |
*** rpioso|afk has quit IRC | 23:43 | |
*** crushil has quit IRC | 23:46 | |
*** yuanying has quit IRC | 23:50 | |
*** yuanying has joined #openstack-ironic | 23:51 | |
*** crushil has joined #openstack-ironic | 23:52 | |
*** zshi_ has joined #openstack-ironic | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!