| *** baoli has quit IRC | 00:00 | |
| *** wajdi_ has joined #openstack-ironic | 00:01 | |
| *** wajdi has quit IRC | 00:04 | |
| *** wajdi_ has quit IRC | 00:05 | |
| jroll | rloo: release notes look good to me rendered: http://docs-draft.openstack.org/70/362870/6/check/gate-python-ironicclient-releasenotes/8cdfef8//releasenotes/build/html/current-series.html | 00:13 |
|---|---|---|
| rloo | jroll: ugh. deprecation notes, it has 'openstack' not 'OpenStack'. | 00:19 |
| *** Sukhdev has joined #openstack-ironic | 00:20 | |
| *** baoli has joined #openstack-ironic | 00:21 | |
| jroll | blah | 00:23 |
| jroll | I'll fix real quick | 00:23 |
| rloo | jroll: i +A'd your patch. you can add a new patch to fix that one or not, up to you. | 00:24 |
| rloo | jroll: and if you do fix it, i guess it should be s/Deprecated/Deprecates/ | 00:24 |
| jroll | yep, did that | 00:25 |
| *** watanabe_isao has joined #openstack-ironic | 00:25 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/python-ironicclient: Correct a couple small grammar things in release notes https://review.openstack.org/363996 | 00:26 |
| jroll | rloo: ^ there's the update, I think it's small enough for a single +2 unless NobodyCam is still around | 00:26 |
| *** thrash is now known as thrash|g0ne | 00:26 | |
| * jroll wants to get these out | 00:27 | |
| rloo | jroll: +A'd | 00:28 |
| jroll | thanks | 00:28 |
| * jroll will release in the morning | 00:28 | |
| openstackgerrit | Merged openstack/python-ironicclient: Clean up release notes for 1.7 https://review.openstack.org/362870 | 00:29 |
| *** baoli has quit IRC | 00:30 | |
| rloo | why'd that take so long to render, yet so fast to merge! | 00:30 |
| jroll | check queue is always longer than gate | 00:31 |
| jroll | and I think it was backed up, seemed someone had pushed a ton of commits right before | 00:31 |
| openstackgerrit | Ruby Loo proposed openstack/python-ironicclient: Adds --chassis arg to 'openstack baremetal node set' https://review.openstack.org/363999 | 00:35 |
| *** baoli has joined #openstack-ironic | 00:35 | |
| rloo | jroll: well, hopefully your last change is fast too | 00:36 |
| jroll | rloo: doesn't matter, I'm done for the night :P | 00:36 |
| jroll | just needs to be faster than my sleep | 00:36 |
| rloo | jroll: true. have a good sleep :) | 00:36 |
| jroll | thanks, enjoy your night :) | 00:37 |
| rloo | jroll: :) | 00:37 |
| *** sabeen has quit IRC | 00:37 | |
| *** rloo has quit IRC | 00:39 | |
| openstackgerrit | Merged openstack/python-ironicclient: Correct a couple small grammar things in release notes https://review.openstack.org/363996 | 00:39 |
| jroll | well snap, maybe I'm not done | 00:41 |
| TheJulia | :( | 00:45 |
| jroll | TheJulia: no, I'm excited! | 00:48 |
| * jroll marked another thing off his todo list | 00:48 | |
| *** mtanino has quit IRC | 00:49 | |
| jroll | tomorrow will be driver deprecation day | 00:51 |
| jroll | gnight, everyone | 00:51 |
| TheJulia | goodnight! | 00:51 |
| *** vishwanathj has quit IRC | 00:53 | |
| *** wlfightup has quit IRC | 00:53 | |
| *** Goneri has joined #openstack-ironic | 00:54 | |
| *** Sukhdev has quit IRC | 00:55 | |
| openstackgerrit | Naohiro Tamura proposed openstack/python-ironicclient: Add a new command "ironic node-inject-nmi" https://review.openstack.org/247905 | 01:05 |
| openstackgerrit | Naohiro Tamura proposed openstack/python-ironicclient: Add a new OSC command for Inject NMI https://review.openstack.org/357661 | 01:07 |
| *** adnavare has quit IRC | 01:10 | |
| *** aNupoisc has quit IRC | 01:10 | |
| *** priteau has joined #openstack-ironic | 01:11 | |
| *** Sukhdev has joined #openstack-ironic | 01:12 | |
| *** Sukhdev has quit IRC | 01:13 | |
| *** hoangcx has joined #openstack-ironic | 01:14 | |
| *** priteau has quit IRC | 01:16 | |
| *** chlong has joined #openstack-ironic | 01:18 | |
| *** pmannidi has quit IRC | 01:20 | |
| *** pmannidi has joined #openstack-ironic | 01:21 | |
| *** baoli_ has joined #openstack-ironic | 01:28 | |
| *** baoli has quit IRC | 01:28 | |
| *** baoli has joined #openstack-ironic | 01:30 | |
| *** hoangcx has quit IRC | 01:31 | |
| *** baoli_ has quit IRC | 01:33 | |
| *** ijw has quit IRC | 02:01 | |
| *** sdake has quit IRC | 02:06 | |
| *** hoangcx has joined #openstack-ironic | 02:07 | |
| *** phuongnh has joined #openstack-ironic | 02:11 | |
| *** sabeen has joined #openstack-ironic | 02:28 | |
| *** syed_ has quit IRC | 02:32 | |
| *** Goneri has quit IRC | 02:35 | |
| *** Sukhdev has joined #openstack-ironic | 02:48 | |
| *** amotoki has joined #openstack-ironic | 02:51 | |
| *** hoangcx has quit IRC | 02:58 | |
| openstackgerrit | Ruby Loo proposed openstack/python-ironicclient: Move duplicated info to new v1/utils.py https://review.openstack.org/364038 | 03:00 |
| *** wajdi has joined #openstack-ironic | 03:02 | |
| *** amotoki has quit IRC | 03:08 | |
| *** amotoki has joined #openstack-ironic | 03:21 | |
| *** amotoki has quit IRC | 03:30 | |
| *** sabeen has quit IRC | 03:31 | |
| *** amotoki has joined #openstack-ironic | 03:35 | |
| *** vikrant has joined #openstack-ironic | 03:36 | |
| *** sabeen has joined #openstack-ironic | 03:42 | |
| *** hj-hpe has quit IRC | 03:45 | |
| *** amotoki has quit IRC | 04:01 | |
| *** Sukhdev has quit IRC | 04:09 | |
| *** vikrant has quit IRC | 04:09 | |
| *** vikrant has joined #openstack-ironic | 04:10 | |
| *** MattMan has quit IRC | 04:12 | |
| *** MattMan has joined #openstack-ironic | 04:12 | |
| *** fragatina has quit IRC | 04:14 | |
| *** amotoki has joined #openstack-ironic | 04:16 | |
| *** baoli has quit IRC | 04:24 | |
| *** links has joined #openstack-ironic | 04:32 | |
| *** fragatina has joined #openstack-ironic | 04:34 | |
| *** krtaylor has quit IRC | 04:34 | |
| *** rajinir has quit IRC | 04:35 | |
| *** BadCub has quit IRC | 04:37 | |
| *** amit213 has quit IRC | 04:37 | |
| *** Ng has quit IRC | 04:38 | |
| *** NobodyCam has quit IRC | 04:38 | |
| *** betherly has quit IRC | 04:38 | |
| *** vdrok has quit IRC | 04:38 | |
| *** amit213 has joined #openstack-ironic | 04:40 | |
| *** vdrok has joined #openstack-ironic | 04:40 | |
| *** betherly has joined #openstack-ironic | 04:40 | |
| *** NobodyCam has joined #openstack-ironic | 04:41 | |
| *** BadCub has joined #openstack-ironic | 04:42 | |
| *** Ng has joined #openstack-ironic | 04:42 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/363863 | 04:56 |
| *** moshele has joined #openstack-ironic | 04:57 | |
| *** pcaruana has quit IRC | 04:57 | |
| *** cppforlife_ has quit IRC | 05:03 | |
| *** sergek has quit IRC | 05:03 | |
| *** agordeev has quit IRC | 05:03 | |
| *** kun_huang has quit IRC | 05:04 | |
| *** rama_y has joined #openstack-ironic | 05:05 | |
| *** sergek has joined #openstack-ironic | 05:05 | |
| *** agordeev has joined #openstack-ironic | 05:05 | |
| *** cppforlife_ has joined #openstack-ironic | 05:05 | |
| *** harlowja_at_home has joined #openstack-ironic | 05:06 | |
| *** rama_y has quit IRC | 05:06 | |
| *** kun_huang has joined #openstack-ironic | 05:07 | |
| *** mgkwill has quit IRC | 05:07 | |
| *** rama_y has joined #openstack-ironic | 05:07 | |
| *** mgkwill has joined #openstack-ironic | 05:08 | |
| *** pas-ha has quit IRC | 05:10 | |
| *** pas-ha has joined #openstack-ironic | 05:10 | |
| *** fyxim has quit IRC | 05:13 | |
| *** fyxim has joined #openstack-ironic | 05:14 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Update driver requirement for iRMC https://review.openstack.org/364071 | 05:14 |
| *** yibo has joined #openstack-ironic | 05:18 | |
| *** fragatin_ has joined #openstack-ironic | 05:22 | |
| *** sdake has joined #openstack-ironic | 05:23 | |
| *** fragatina has quit IRC | 05:25 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Update driver requirement for iRMC https://review.openstack.org/364077 | 05:30 |
| *** mjura has joined #openstack-ironic | 05:42 | |
| *** ChubYann has quit IRC | 05:42 | |
| *** syed_ has joined #openstack-ironic | 05:50 | |
| *** harlowja_at_home has quit IRC | 05:51 | |
| *** mjura has quit IRC | 05:51 | |
| *** mjura has joined #openstack-ironic | 05:51 | |
| *** hoangcx has joined #openstack-ironic | 05:56 | |
| *** sabeen has quit IRC | 05:57 | |
| *** Sukhdev has joined #openstack-ironic | 06:01 | |
| *** rcernin has quit IRC | 06:08 | |
| *** ljjjustin has quit IRC | 06:21 | |
| *** ljjjustin has joined #openstack-ironic | 06:21 | |
| *** vnogin has quit IRC | 06:22 | |
| *** vnogin has joined #openstack-ironic | 06:22 | |
| *** pcaruana has joined #openstack-ironic | 06:30 | |
| *** rcernin has joined #openstack-ironic | 06:41 | |
| *** Sukhdev has quit IRC | 06:42 | |
| *** rcernin has quit IRC | 06:45 | |
| *** rcernin has joined #openstack-ironic | 06:45 | |
| *** hoangcx has quit IRC | 06:52 | |
| *** hoangcx_ has joined #openstack-ironic | 06:52 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Fix updating port.portgroup_uuid for node https://review.openstack.org/364095 | 06:53 |
| *** tesseract- has joined #openstack-ironic | 07:03 | |
| *** rama_y has quit IRC | 07:06 | |
| *** rama_y has joined #openstack-ironic | 07:06 | |
| pas-ha | morning Ironicers :) | 07:06 |
| *** rama_y has quit IRC | 07:08 | |
| *** rama_y has joined #openstack-ironic | 07:08 | |
| openstackgerrit | zhufl proposed openstack/ironic: Remove unused LOG and CONF https://review.openstack.org/364109 | 07:16 |
| *** jlvillal has quit IRC | 07:22 | |
| *** rama_y has quit IRC | 07:23 | |
| *** jlvillal has joined #openstack-ironic | 07:23 | |
| *** rama_y has joined #openstack-ironic | 07:24 | |
| *** jpich has joined #openstack-ironic | 07:25 | |
| *** praneshp has quit IRC | 07:26 | |
| *** ifarkas_afk is now known as ifarkas | 07:28 | |
| *** rama_y has quit IRC | 07:29 | |
| *** phuongnh has quit IRC | 07:30 | |
| *** phuongnh has joined #openstack-ironic | 07:31 | |
| *** hoangcx_ has quit IRC | 07:35 | |
| *** hoangcx has joined #openstack-ironic | 07:40 | |
| *** mgoddard has joined #openstack-ironic | 07:43 | |
| *** sdake has quit IRC | 07:43 | |
| *** sdake has joined #openstack-ironic | 07:44 | |
| *** phuongnh has quit IRC | 07:50 | |
| openstackgerrit | Merged openstack/ironic-lib: Sync tools/tox_install.sh https://review.openstack.org/363086 | 07:57 |
| *** dtantsur|afk is now known as dtantsur | 07:58 | |
| dtantsur | Morning Ironic, morning pas-ha | 07:58 |
| *** ohamada has joined #openstack-ironic | 07:59 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** ljjjustin has quit IRC | 08:00 | |
| *** zzzeek has joined #openstack-ironic | 08:00 | |
| *** syed_ has quit IRC | 08:02 | |
| *** jaosorior has joined #openstack-ironic | 08:04 | |
| *** athomas has joined #openstack-ironic | 08:04 | |
| *** hoangcx has quit IRC | 08:05 | |
| *** milan has joined #openstack-ironic | 08:10 | |
| *** parlos has quit IRC | 08:10 | |
| *** ohamada has quit IRC | 08:10 | |
| *** ohamada has joined #openstack-ironic | 08:11 | |
| openstackgerrit | Merged openstack/ironic: Replace DOS line endings with Unix https://review.openstack.org/363900 | 08:22 |
| dtantsur | jroll, do you plan on an ironic release this week? | 08:24 |
| dtantsur | also clients were just released \o/ | 08:24 |
| *** narasimha_SV has joined #openstack-ironic | 08:27 | |
| narasimha_SV | http://paste.openstack.org/show/565646/ | 08:27 |
| narasimha_SV | anyone facing this error | 08:27 |
| narasimha_SV | I have deployed ironic with the master code | 08:27 |
| narasimha_SV | I am getting above error in nova-compute configured with ironic | 08:27 |
| narasimha_SV | the ironic section mentioned are the values from nova.conf | 08:28 |
| *** Romanenko_K has joined #openstack-ironic | 08:33 | |
| *** priteau has joined #openstack-ironic | 08:37 | |
| dtantsur | narasimha_SV, on my environment, the admin_url has /v2.0 prefix, maybe that can help | 08:38 |
| dtantsur | s/prefix/postfix/ | 08:38 |
| *** lucas-dinner is now known as lucasagomes | 08:38 | |
| lucasagomes | good morning all | 08:38 |
| dtantsur | in your case http://192.168.56.101:35357/v2.0 | 08:39 |
| dtantsur | morning lucasagomes | 08:39 |
| openstackgerrit | Merged openstack/ironic: Add a newline at the end of release note files https://review.openstack.org/363919 | 08:39 |
| ifarkas | morning all | 08:39 |
| dtantsur | morning ifarkas as well :) | 08:39 |
| *** e0ne has joined #openstack-ironic | 08:44 | |
| *** hwoarang_ has joined #openstack-ironic | 08:54 | |
| *** derekh has joined #openstack-ironic | 08:54 | |
| *** jhesketh_ has joined #openstack-ironic | 08:54 | |
| *** HenryG_ has joined #openstack-ironic | 08:54 | |
| *** amotoki_ has joined #openstack-ironic | 08:55 | |
| openstackgerrit | Szymon Borkowski proposed openstack/ironic-inspector: Add PCI devices plugin to inspector https://review.openstack.org/352295 | 08:56 |
| *** jlvillal_ has joined #openstack-ironic | 08:57 | |
| *** jlvillal has quit IRC | 08:58 | |
| *** amotoki has quit IRC | 08:58 | |
| *** vikrant has quit IRC | 08:58 | |
| *** Fdaisuke has quit IRC | 08:58 | |
| *** tangchen has quit IRC | 08:58 | |
| *** ricardoas has quit IRC | 08:58 | |
| *** mrtenio has quit IRC | 08:58 | |
| *** jhesketh has quit IRC | 08:58 | |
| *** mat128 has quit IRC | 08:58 | |
| *** bapalm has quit IRC | 08:58 | |
| *** HenryG has quit IRC | 08:58 | |
| *** anteaya has quit IRC | 08:58 | |
| *** sambetts|afk has quit IRC | 08:58 | |
| *** stevebaker has quit IRC | 08:58 | |
| *** hwoarang has quit IRC | 08:58 | |
| *** Guest40699 has quit IRC | 08:58 | |
| *** bapalm has joined #openstack-ironic | 08:59 | |
| *** mat128 has joined #openstack-ironic | 08:59 | |
| *** tangchen has joined #openstack-ironic | 09:00 | |
| *** Fdaisuke has joined #openstack-ironic | 09:00 | |
| *** mrtenio has joined #openstack-ironic | 09:00 | |
| *** jlvillal_ is now known as jlvillal | 09:00 | |
| *** jlvillal is now known as Guest77956 | 09:01 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Rely on portgroup standalone_ports_supported https://review.openstack.org/360647 | 09:02 |
| *** vikrant has joined #openstack-ironic | 09:03 | |
| *** ricardoas has joined #openstack-ironic | 09:03 | |
| *** anteaya has joined #openstack-ironic | 09:03 | |
| *** sambetts|afk has joined #openstack-ironic | 09:03 | |
| *** stevebaker has joined #openstack-ironic | 09:03 | |
| *** Guest40699 has joined #openstack-ironic | 09:03 | |
| milan | morning dtantsur ifarkas lucasagomes! :) | 09:12 |
| milan | and morning Ironic! :) | 09:12 |
| dtantsur | morning milan | 09:12 |
| narasimha_SV | dtantsur: thanks it worked | 09:14 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/363863 | 09:15 |
| lucasagomes | milan, hello there, good morning | 09:16 |
| milan | o/ :) | 09:16 |
| *** rbartal has joined #openstack-ironic | 09:16 | |
| *** HenryG_ is now known as HenryG | 09:17 | |
| openstackgerrit | Merged openstack/ironic-python-agent: Fix IPA for stable/mitaka with noauth mode https://review.openstack.org/359991 | 09:19 |
| *** appu103_ has joined #openstack-ironic | 09:22 | |
| *** sambetts|afk is now known as sambetts | 09:23 | |
| sambetts | Morning all | 09:23 |
| openstackgerrit | Bruno Cornec proposed openstack/bifrost: Fix minor documentation issue https://review.openstack.org/364185 | 09:25 |
| lucasagomes | morning | 09:25 |
| openstackgerrit | Bruno Cornec proposed openstack/bifrost: Fix minor documentation issue https://review.openstack.org/364185 | 09:26 |
| lucasagomes | can I get few eyes on https://review.openstack.org/#/c/360658/ ? It's a simple patch and fixes this error here: http://logs.openstack.org/15/342915/7/check/gate-tempest-dsvm-ironic-ipa-wholedisk-agent_ssh-tinyipa/0516c52/logs/screen-ir-cond.txt.gz#_2016-09-01_08_20_27_561 | 09:26 |
| *** bcornec1 has joined #openstack-ironic | 09:26 | |
| dtantsur | morning sambetts | 09:28 |
| *** HenryG has quit IRC | 09:29 | |
| *** HenryG has joined #openstack-ironic | 09:29 | |
| *** hwoarang_ is now known as hwoarang | 09:31 | |
| openstackgerrit | Bruno Cornec proposed openstack/bifrost: Fix minor documentation issue https://review.openstack.org/364189 | 09:33 |
| sambetts | o/ lucasagomes dtantsur | 09:33 |
| sambetts | lucasagomes: Nice +W i've seen at least 2 or 3 people in here asking whats causing that error when they stack recently | 09:36 |
| lucasagomes | sambetts, yeah it's really misleading when you look at the logs and see that error | 09:37 |
| lucasagomes | sambetts, dtantsur thanks for approving | 09:37 |
| lucasagomes | dtantsur, re: vim, I just find it easier to have it to remove trialing spaces for me, automation! | 09:38 |
| lucasagomes | maybe we could configure bashate to fail on trailing spaces as pep8 does | 09:38 |
| openstackgerrit | Merged openstack/ironic: Fix incorrect order of params of assertEqual() https://review.openstack.org/359580 | 09:42 |
| *** daemontool has joined #openstack-ironic | 09:45 | |
| lucasagomes | has anyone seem this before http://logs.openstack.org/15/342915/7/check/gate-tempest-dsvm-ironic-ipa-wholedisk-agent_ssh-tinyipa/0516c52/logs/screen-ir-cond.txt.gz#_2016-09-01_08_20_16_733 ? | 09:48 |
| lucasagomes | apparently there's a race on heartbeat() ? I can see that the node was suppose to be on "deploywait" state few lines above, and then it's failing cause it can move from "deploying" to "resume" | 09:49 |
| lucasagomes | sounds like the node object that tried to change the provision state was outdated | 09:49 |
| vsaienk0 | lucasagomes: I've seen this on multinode job, but didn't have enough time to debug, might be pas-ha may help, he also looked at this issue | 09:52 |
| lucasagomes | vsaienk0, thanks, first time I've seem it | 09:53 |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add portgroup support to osc plugin https://review.openstack.org/362130 | 10:00 |
| vdrok | morning ironic, pas-ha dtantsur lucasagomes milan sambetts vsaienk0 ! | 10:00 |
| *** sborkows has joined #openstack-ironic | 10:01 | |
| dtantsur | morning vdrok, vsaienk0 | 10:01 |
| sborkows | Hello | 10:01 |
| *** sdake has quit IRC | 10:01 | |
| milan | morning pas-ha sborkows vdrok vsaienk0! :) | 10:01 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/363863 | 10:04 |
| sambetts | lucasagomes: it seems like prepare_image and therefore continue deploy ends up getting called twice in this log for the same node, so I think there might be a locking issue somewhere | 10:09 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/ironic: [WIP] Testing latest u-c https://review.openstack.org/318440 | 10:10 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/ironic: [WIP] Testing latest u-c https://review.openstack.org/318440 | 10:10 |
| lucasagomes | sambetts, yeah | 10:10 |
| lucasagomes | morning vdrok | 10:10 |
| dtantsur | found an unpleasant behavior in our OSC plugin: https://bugs.launchpad.net/python-ironicclient/+bug/1619229 | 10:12 |
| openstack | Launchpad bug 1619229 in python-ironicclient "OSC plugin fails when ironic service catalog entry contains /v1 postfix" [High,New] | 10:12 |
| dtantsur | jroll, ^^^ | 10:12 |
| sambetts | lucasagomes: errrr.. this isn't good ... I see an exclusive lock claimed on the node, and then a second exclusive lock claimed before the first is released | 10:12 |
| pas-ha | lucasagomes: also it seems we need to configure log collection properly on gates (not to /var/lib/ironic) | 10:13 |
| lucasagomes | pas-ha, yeah there's a patch on gate right now fixing that | 10:14 |
| pas-ha | ok, cool | 10:14 |
| lucasagomes | pas-ha, https://review.openstack.org/#/c/360658/ | 10:14 |
| lucasagomes | sambetts, I think that what is happening is, we are not updating the object upon getting a exclusive lock | 10:15 |
| lucasagomes | sambetts, 2 threads gets a shared lock and therefor holds an instance of the node object | 10:16 |
| lucasagomes | once they upgrade the lock, the object is not refreshed | 10:16 |
| sambetts | lucasagomes: so both claim an exclusive lock? | 10:16 |
| lucasagomes | so the last thread to upgrade the lock still have some old information in the instance | 10:16 |
| lucasagomes | sambetts, I'm unsure on that yet... lemme dig more | 10:17 |
| lucasagomes | in any case I will open a bug on it | 10:17 |
| lucasagomes | (if there's none already) | 10:17 |
| sambetts | lucasagomes: I can see the first heartbeat coming in upgrading the lock to an exclusive lock, then the second heartbeat comes in a claims a shared and then also successfully upgrades it lock too | 10:17 |
| sambetts | lucasagomes: http://logs.openstack.org/15/342915/7/check/gate-tempest-dsvm-ironic-ipa-wholedisk-agent_ssh-tinyipa/0516c52/logs/screen-ir-cond.txt.gz#_2016-09-01_08_19_45_274 | 10:18 |
| sambetts | lucasagomes: from this point down | 10:18 |
| sambetts | lucasagomes: I can see 2 successful upgrades before a exclusive release | 10:18 |
| sambetts | unless our logging is wrong | 10:18 |
| *** rbartal has quit IRC | 10:19 | |
| lucasagomes | yeah there's something fishy there | 10:19 |
| *** thrash|g0ne is now known as thrash | 10:19 | |
| *** jhesketh_ is now known as jhesketh | 10:20 | |
| sambetts | lucasagomes: there is a heartbeat that comes in midway through the continue deploy call, that shouldn't have even been allowed and should have responded with a node locked error | 10:20 |
| lucasagomes | there's def two logs about upgrading the log before a release | 10:20 |
| sambetts | lucasagomes: I wonder if it has something to do with using the new API over the old vendor one | 10:22 |
| lucasagomes | sambetts, def something to investigate since it touches those bits | 10:23 |
| sambetts | lucasagomes: the old vendor one used to fail instantly if you tried to use it and the node was locked | 10:23 |
| lucasagomes | opened https://bugs.launchpad.net/ironic/+bug/1619232 | 10:23 |
| openstack | Launchpad bug 1619232 in Ironic "Heartbeat()'s race condition: InvalidState: Can not transition from state 'deploying' on event 'resume' (no defined transition)" [High,New] | 10:23 |
| openstackgerrit | Bruno Cornec proposed openstack/bifrost: Fix minor documentation issue https://review.openstack.org/364189 | 10:26 |
| jroll | dtantsur: yeah, I was thinking about an ironic release, maybe after merging power state notifications? we need to get IPA out the door as well | 10:26 |
| jroll | dtantsur: I see your osc bug :( we should be able to backport that no problem | 10:26 |
| lucasagomes | sambetts, so the thing is that _lock() has a rety mechanism | 10:26 |
| dtantsur | jroll, cool, I'm thinking about inspector release too, as soon as we finish with the PCI capabilities patch | 10:26 |
| lucasagomes | sambetts, so I still think it may be a case of two shared locks trying to upgrade | 10:26 |
| dtantsur | jroll, osc bug.. yeah. I still don't quite understand where it originates tbh | 10:27 |
| lucasagomes | and once the second one succeed it's outdated | 10:27 |
| jroll | lucasagomes: we've been seeing that deploy/resume thing too, just started seeing it this week, thought it was a downstream hack bug | 10:27 |
| jroll | wow | 10:27 |
| lucasagomes | jroll, yeah investigating it now | 10:27 |
| dtantsur | this week, hmm.. what did we merge this week? | 10:27 |
| jroll | dtantsur: well, we saw it in 6.0.0+hacks this week | 10:27 |
| jroll | so something there | 10:28 |
| dtantsur | aha | 10:28 |
| jroll | and I think with old IPA endpoint? | 10:28 |
| jroll | in case that matters | 10:28 |
| jroll | almost certainly old IPA endpoint | 10:28 |
| dtantsur | with old? hmm, interesting. I suspect we did something wrong around new IPA endpoints... | 10:28 |
| jroll | dtantsur: I think we changed something around locking with the old, actually | 10:29 |
| dtantsur | milan, this is why I don't like merging things late in a cycle ^^^ | 10:29 |
| sambetts | so I think the problem is that the the way locks were handled for the old vendor_passthrough method is a little differnt to other API calls | 10:29 |
| dtantsur | this is the 3rd (?) regression around that | 10:29 |
| jroll | okay, so heartbeat doesn't require a shared lock | 10:30 |
| jroll | it used to | 10:30 |
| jroll | or did it, hm | 10:30 |
| lucasagomes | jroll, I think it required an exclusive one, no? | 10:31 |
| jroll | oh, right, because vendor passthru requires one no? | 10:31 |
| lucasagomes | yeah, I think before it was always exclusive | 10:31 |
| lucasagomes | but now, I think it's up to the methods to upgrade the lock | 10:31 |
| sambetts | heartbeat used to require an exclisive lock | 10:31 |
| jroll | I'm looking at this bit of the change: https://github.com/openstack/ironic/commit/a9216bb07f4ccd4dc202fd1f5c14b36a08ac6046#diff-3e6e90ffe62c55e7f724a6bda068aebeR462 | 10:31 |
| lucasagomes | but I wonder why would IPA be using the old endpoint ? | 10:31 |
| jroll | yeah, in CI we should be using the new | 10:32 |
| jroll | but I believe we're using the old downstream | 10:32 |
| lucasagomes | I see | 10:32 |
| dtantsur | anyway, we do upgrade the lock: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent_base_vendor.py#L451 | 10:32 |
| lucasagomes | lemme look at the logs see if I can figure out whether it's old or new | 10:32 |
| *** syed_ has joined #openstack-ironic | 10:33 | |
| sambetts | dtantsur: yeah we can see that happening, but twice, from different requests | 10:33 |
| sambetts | dtantsur: before the first releases it | 10:33 |
| dtantsur | mmm, a bug in upgrade_lock? | 10:33 |
| jroll | yeah, lucas' theory is two heartbeats at the same time, and the upgrade doesn't refresh the object | 10:33 |
| jroll | could be | 10:33 |
| jroll | this is on the new endpoint http://logs.openstack.org/15/342915/7/check/gate-tempest-dsvm-ironic-ipa-wholedisk-agent_ssh-tinyipa/0516c52/logs/screen-ir-api.txt.gz#_2016-09-01_08_19_45_276 | 10:34 |
| jroll | also, don't have two heartbeats nearby in api logs, so maybe racing against power sync loop? | 10:34 |
| dtantsur | do we also have two messages "Node %(node)s successfully reserved"? | 10:34 |
| lucasagomes | apparently it's the new one http://logs.openstack.org/15/342915/7/check/gate-tempest-dsvm-ironic-ipa-wholedisk-agent_ssh-tinyipa/0516c52/logs/screen-ir-api.txt.gz#_2016-09-01_08_22_05_350 | 10:35 |
| sambetts | dtantsur: http://logs.openstack.org/15/342915/7/check/gate-tempest-dsvm-ironic-ipa-wholedisk-agent_ssh-tinyipa/0516c52/logs/screen-ir-cond.txt.gz#_2016-09-01_08_19_45_274 | 10:35 |
| jroll | this actually doesn't seem like a race to me, it isn't happening during a heartbeat | 10:35 |
| lucasagomes | jroll, yeah heartbeats are pretty close to each other there | 10:36 |
| jroll | see the timestamps | 10:36 |
| jroll | not really? they're seconds apart | 10:36 |
| * lucasagomes looks | 10:36 | |
| lucasagomes | true, at least ~4s | 10:36 |
| dtantsur | hmm, it looks like upgrade_lock does not quite work... | 10:36 |
| jroll | oh wait, fun | 10:36 |
| sambetts | the second heartbeat is coming in during the continue_deploy, which should fail | 10:36 |
| sambetts | because of the lcoking | 10:37 |
| dtantsur | I don't see a "Node successfully reserved" message nearby | 10:37 |
| jroll | 8:20:06 starts fetching command status, 8:20:16 finishes fetching it | 10:37 |
| jroll | there's a heartbeat in the middle | 10:37 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Fix exception handling in NodesController._lookup https://review.openstack.org/364222 | 10:37 |
| jroll | it tries to upgrade the lock at 8:20:11 and then releases it (still shared) at 8:20:13 | 10:37 |
| vdrok | morning jroll | 10:37 |
| sambetts | jroll: if you look at 8:19:45 thats where we enter the continue_deploy function | 10:38 |
| jroll | hi vdrok | 10:38 |
| sambetts | jroll: and there is a heartbeat half way through it | 10:38 |
| sambetts | jroll: which then tries to continue_deploy itself | 10:38 |
| vdrok | sambetts: a small issue in subcontroller handling - https://review.openstack.org/364222 | 10:38 |
| openstackgerrit | Merged openstack/ironic: Fix updating port.portgroup_uuid for node https://review.openstack.org/364095 | 10:38 |
| jroll | sambetts: yeah, seems like it | 10:39 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Rely on portgroup standalone_ports_supported https://review.openstack.org/360647 | 10:39 |
| sambetts | vdrok: Nice catch, Thanks! | 10:39 |
| jroll | man, thread IDs or something would be really nice here | 10:41 |
| jroll | has someone filed a bug for this, btw? | 10:41 |
| dtantsur | meh, something messy is going on there.. yeah, ++ for thread IDs | 10:41 |
| dtantsur | https://bugs.launchpad.net/ironic/+bug/1619232 > | 10:42 |
| openstack | Launchpad bug 1619232 in Ironic "Heartbeat()'s race condition: InvalidState: Can not transition from state 'deploying' on event 'resume' (no defined transition)" [High,New] - Assigned to Lucas Alvares Gomes (lucasagomes) | 10:42 |
| jroll | thanks | 10:42 |
| jroll | I'm going to make an e-r thing for this, just for tracking | 10:42 |
| jroll | we should also figure out how to log request IDs in conductor | 10:42 |
| lucasagomes | ++ | 10:42 |
| *** parlos has joined #openstack-ironic | 10:44 | |
| jroll | we have them downstream, not sure what config is missing in CI | 10:44 |
| *** stendulker has joined #openstack-ironic | 10:45 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/363863 | 10:48 |
| jroll | oh, we do have request ids, what was I thinking | 10:48 |
| * jroll clearly needs more coffee | 10:48 | |
| *** dtantsur is now known as dtantsur|lunch | 10:49 | |
| sambetts | something weird is happening definatly because I see the actions being taken by continue_deploy before a Node reserve for log | 10:54 |
| openstackgerrit | Merged openstack/ironic: Remove unused LOG and CONF https://review.openstack.org/364109 | 10:55 |
| jroll | e-r check is landing | 10:56 |
| jroll | sambetts: yeah, it's odd, I need to read through it again | 10:56 |
| sambetts | jroll: its a shame request IDs aren't on all the logs, e.g. we're logging the req ID on the attempting to get shared lock log, but not on the Upgrading shared lock log | 10:59 |
| jroll | sambetts: yeah, I don't understand that | 11:00 |
| sambetts | jroll: and in fact looking at the req ids for any consistency I see the heart beat is one request ID and the logs for the continue deploy stuff is a different req ID :/ | 11:01 |
| jroll | sambetts: well, as we suspect, right? | 11:01 |
| sambetts | jroll: oh wow... soo in fact the using the req IDs http://logs.openstack.org/15/342915/7/check/gate-tempest-dsvm-ironic-ipa-wholedisk-agent_ssh-tinyipa/0516c52/logs/screen-ir-cond.txt.gz#_2016-09-01_08_19_35_275 | 11:02 |
| sambetts | this is the heart beat that is starting the continue_deploy, and in fact there are 2 heartbeats that come in after that point | 11:02 |
| openstackgerrit | Merged openstack/ironic: DevStack: Configure retrieving logs from the deploy ramdisk https://review.openstack.org/360658 | 11:03 |
| jroll | sambetts: right, the one that actually does work takes 20s | 11:04 |
| milan | dtantsur|lunch, I see | 11:04 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/363863 | 11:04 |
| sambetts | yeah, so I guess the question is what is different between using upgrade_lock and just passing shared=False into the task manager | 11:05 |
| jroll | I don't see this entering the deploying state | 11:06 |
| jroll | so I like the theory of the object isn't refreshed | 11:06 |
| lucasagomes | jroll, yeah I don't see that either, so that;s why I think the object is just outdated | 11:07 |
| lucasagomes | but again, reserve_node should update it >.< | 11:07 |
| sambetts | http://logs.openstack.org/15/342915/7/check/gate-tempest-dsvm-ironic-ipa-wholedisk-agent_ssh-tinyipa/0516c52/logs/screen-ir-cond.txt.gz#_2016-09-01_08_19_45_416 | 11:07 |
| sambetts | ? | 11:07 |
| lucasagomes | sambetts, it goes back to deploy-wait few lines below | 11:08 |
| sambetts | lucasagomes: thats expected | 11:08 |
| sambetts | lucasagomes: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent.py#L228 | 11:08 |
| lucasagomes | sambetts, right and then after that it fails saying the node was in "deploying" state | 11:08 |
| *** sturivnyi has joined #openstack-ironic | 11:09 | |
| lucasagomes | but there was no transaction from deploy wait to deploying up to that point | 11:09 |
| jroll | interesting that it blows up in reboot_to_instance, and we need to be in DEPLOYWAIT to get in there | 11:09 |
| sambetts | lucasagomes: but thats because the heartbeat started midway through the node doing the contine_deploy code, so at the time it was right, but that should have been blocked by the exclusive locked | 11:09 |
| sambetts | the second heartbeat* | 11:10 |
| *** joprovost has joined #openstack-ironic | 11:11 | |
| sturivnyi | Hi All=) Would be grateful if you could review new patches. Tests for Ironic CLI. | 11:11 |
| sturivnyi | https://review.openstack.org/#/c/251458/ https://review.openstack.org/#/c/259354/ https://review.openstack.org/#/c/262055/ https://review.openstack.org/#/c/291802/ https://review.openstack.org/#/c/342060/ https://review.openstack.org/#/c/343796/ https://review.openstack.org/#/c/344174/ https://review.openstack.org/#/c/344272/ https://review.openstack.org/#/c/344715/ https://review.openstack.org/#/c/360549/ https://review.openstack.org/#/c/3 | 11:11 |
| sturivnyi | 60636/ | 11:11 |
| jroll | sturivnyi: gerrit's "topic" feature may be helpful to you | 11:11 |
| jroll | lucasagomes: sambetts: I wonder if we're doing a process_event without an exclusive lock somewhere | 11:12 |
| sambetts | the second heartbeat comes in when the node is in deploying, oh I guess the _lock retrying thing sets self.node so should have updated that tasks node | 11:12 |
| lucasagomes | jroll, so, one thing that would be nice to see (and we will when we get more data) is what drivers does it affects, because agent driver flips the state to deploywait when the image is being copied | 11:13 |
| lucasagomes | where iscsi doesn't | 11:13 |
| lucasagomes | it may happen at that point | 11:13 |
| jroll | lucasagomes: yeah, wouldn't be surprised if it's only agent | 11:13 |
| * jroll does logstash magic | 11:15 | |
| openstackgerrit | Sergii Turivnyi proposed openstack/python-ironicclient: Negative tests for testing actions with node-list command https://review.openstack.org/360549 | 11:15 |
| sborkows | dtantsur|lunch: I fixed my patch about pci_devices processing hook in Ironic -> now it's pointing to capabilities, as you suggest. You can check it out on review. | 11:15 |
| sambetts | lucasagomes: this is a weird line to see http://logs.openstack.org/15/342915/7/check/gate-tempest-dsvm-ironic-ipa-wholedisk-agent_ssh-tinyipa/0516c52/logs/screen-ir-cond.txt.gz#_2016-09-01_08_19_47_303 | 11:16 |
| lucasagomes | sambetts, yeah apparently when the node was still in deploying state | 11:17 |
| sambetts | lucasagomes: how can we be releasing a shared_lock while in an exclusive one, I wonder who is calling release_resources here | 11:17 |
| *** joprovost1 has joined #openstack-ironic | 11:18 | |
| sambetts | lucasagomes: I'm guessing it was the second heartbeat which failed to upgrade | 11:19 |
| sambetts | lucasagomes: and then the third heartbeat no longer sees an exclusive lock so claims it for itself | 11:20 |
| *** joprovost has quit IRC | 11:20 | |
| *** joprovost1 is now known as joprovost | 11:20 | |
| sambetts | lucasagomes: although it shouldn't release the exclusive if the lock type wasn't shared :/ damn... | 11:21 |
| TheJulia | good morning everyone | 11:21 |
| sambetts | o/ TheJulia | 11:22 |
| dtantsur|lunch | morning TheJulia | 11:22 |
| dtantsur|lunch | sborkows, will check, lemme finish my lunch :) | 11:22 |
| lucasagomes | TheJulia, morning | 11:23 |
| lucasagomes | sambetts, yeah... it's kinda odd | 11:23 |
| lucasagomes | s/kinda/totally | 11:24 |
| lucasagomes | I wonder why that lock was released at that point | 11:24 |
| sambetts | lucasagomes: I know why (I think) I think that log comes from the task manager of the second heartbeat, i.e. the second heartbeat comes in the task manager is setup with a shared lock, then tries to upgrade it, the upgrade fails, then it exits in the task_manager context so releases the shared lock it claims | 11:25 |
| sambetts | lucasagomes: here https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L2204 | 11:26 |
| lucasagomes | sambetts, interesting, I can see task_manager doesn't check for shared when relasing the resources | 11:27 |
| lucasagomes | that might be the case | 11:27 |
| lucasagomes | lemme test it | 11:27 |
| lucasagomes | oh no it does urgh... wait... | 11:28 |
| sambetts | yeah... it does check it : / | 11:28 |
| sambetts | lucasagomes: unless there is a race between upgrading and release_resources *mind splodes* | 11:29 |
| lucasagomes | hah yeah | 11:29 |
| lucasagomes | I think I need some energy, I will grab some lunch and continue looking at it after | 11:30 |
| sambetts | same | 11:30 |
| *** lucasagomes is now known as lucas-hungry | 11:33 | |
| mat128 | morning everyone | 11:35 |
| TheJulia | good morning mat128 | 11:35 |
| *** joprovost has quit IRC | 11:36 | |
| *** dprince has joined #openstack-ironic | 11:36 | |
| *** joprovost has joined #openstack-ironic | 11:37 | |
| *** moshele has quit IRC | 11:38 | |
| *** baoli has joined #openstack-ironic | 11:41 | |
| *** baoli_ has joined #openstack-ironic | 11:42 | |
| *** baoli has quit IRC | 11:46 | |
| jroll | sambetts: I agree with it failing to upgrade and just releasing it | 11:49 |
| jroll | man this is a rough one | 11:49 |
| sambetts | jroll: which is the right behavior as far as I can tell, but looking at the release_resources logic releasing a shared lock shouldn't release a exclusive, but equally we shouldn't be able to claim a shared lock if we already have an exclusive right? | 11:50 |
| sambetts | multiple shared yes, exclusive + shared no right? | 11:51 |
| jroll | sambetts: it reads to me that you should be able to get a shared lock if there's already an exclusive one | 11:51 |
| jroll | which seems a bit wrong | 11:51 |
| sambetts | yeah thats not right IMO | 11:52 |
| jroll | but idk if it hurts anything | 11:52 |
| sambetts | jroll: only that important information can become out of date if the thing with the exclusive lock change them and the thing with the shared lock doesn't notice | 11:53 |
| jroll | mmm, yeah | 11:54 |
| jroll | sambetts: we do get a new node object if we try to upgrade, though | 11:54 |
| sambetts | yeah, which is why I'm finding it really hard to work out whats going wrong here :/ | 11:55 |
| jroll | indeed | 11:55 |
| mat128 | jroll: is that your deploying/deploy_wait issue from the other day? | 11:56 |
| jroll | mat128: yep | 11:57 |
| lucas-hungry | yeah, at the moment we do allow a shared + exclusive I'm not sure if it's wrong either... it would be hard to prevent in case we have a shared + shared and one of them upgrade the lock | 11:57 |
| sambetts | yup :'( | 11:57 |
| *** dtantsur|lunch is now known as dtantsur | 12:04 | |
| dtantsur | morning mat128 | 12:04 |
| sambetts | so stepping back a bit, it seems like: the first heartbeat is handled correctly, the second heartbeat fails to upgrade (which is the correct behavior), the third heartbeat requests the lock upgrade during the first heartbeats fsm_event to wait call-back, then takes the lock when the first heartbeat releases it | 12:05 |
| mat128 | o/ dtantsur | 12:05 |
| sambetts | the forth heartbeat does the same as the second | 12:05 |
| *** jaosorior has quit IRC | 12:06 | |
| sambetts | and so does the next | 12:06 |
| *** jaosorior has joined #openstack-ironic | 12:06 | |
| openstackgerrit | Imre Farkas proposed openstack/ironic: DRAC OOB inspection https://review.openstack.org/342915 | 12:19 |
| *** joprovost has quit IRC | 12:20 | |
| *** lucas-hungry is now known as lucasagomes | 12:21 | |
| jroll | sambetts: this has to be something not refreshing the node object properly | 12:23 |
| jroll | I'm not sure I understand *how* | 12:23 |
| jroll | but it has to be | 12:23 |
| sambetts | jroll: yeah thats what I'm looking into now | 12:24 |
| *** sturivnyi has quit IRC | 12:24 | |
| sambetts | jroll: I can't work out why we've not seen it before now | 12:24 |
| jroll | I'm surprised an update query will return a db object | 12:24 |
| *** wajdi has quit IRC | 12:24 | |
| jroll | sambetts: previously vendor passthru would immediately request an exclusive lock, rather than a shared + upgrade | 12:25 |
| sambetts | jroll: that changed not to long ago, but before dtantsur's new API | 12:25 |
| sambetts | jroll: because we made it possible to have shared look vendor passthrough | 12:26 |
| jroll | sambetts: that would explain why I see it on the old endpoint too :) | 12:26 |
| jroll | old heartbeat endpoint, to be clear | 12:26 |
| sambetts | https://github.com/openstack/ironic/blob/master/ironic/conductor/manager.py#L257 | 12:26 |
| sambetts | jroll: ah right so you see it on that one too | 12:26 |
| * lucasagomes looks what synchronize_session=False might do | 12:27 | |
| * jroll playing around in a python session | 12:27 | |
| lucasagomes | we do update the node with that flag when reserving it | 12:27 |
| sambetts | jroll: interesting, so it must be an issue with the upgrade_lock thing, but what I don't understand is that upgrade_lock calls the same code path as a task_manager created with shared=False | 12:27 |
| jroll | sambetts: except you wouldn't have a node object to start with | 12:28 |
| jroll | so if it doesn't refresh the object correctly, that's the difference | 12:28 |
| sambetts | https://github.com/openstack/ironic/blob/master/ironic/conductor/task_manager.py#L209 | 12:29 |
| jroll | ah | 12:30 |
| sambetts | both call _lock() which should set self.node | 12:30 |
| jroll | so maybe there's a race in both, just much larger window for the upgrade case | 12:30 |
| jroll | I suddenly want this for ironic https://docs.djangoproject.com/en/1.10/ref/django-admin/#shell | 12:31 |
| jroll | gives you a python shell with config loaded and such | 12:31 |
| sambetts | heh | 12:31 |
| openstackgerrit | Milan Kováčik proposed openstack/ironic-inspector: Bypassing NodeInfo.acquire_lock() to try OOL https://review.openstack.org/360311 | 12:34 |
| * milan out, bye Ironic! | 12:36 | |
| sambetts | milan: cya dude! | 12:36 |
| milan | sambetts, cya! :D | 12:36 |
| sambetts | jroll: I guess there is always going to be a problem with code that does, node = task.node, followed by a task.upgrade_lock() because then node is no longer == task.node because object references | 12:37 |
| *** milan has quit IRC | 12:37 | |
| sambetts | jroll: trying to see if that happens anywhere in this | 12:37 |
| jroll | sambetts: I don't think it does happen here, at least not in heartbeat() | 12:37 |
| *** stendulker has quit IRC | 12:47 | |
| *** links has quit IRC | 12:51 | |
| sambetts | jroll: in the task_manager what type of object is self.context? | 12:51 |
| jroll | sambetts: ironic.common.context.RequestContext | 12:51 |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic: Fix Ubuntu 16.04 pxelinux path handling https://review.openstack.org/326024 | 12:51 |
| sambetts | jroll: I'm a little confused why we're using the context here https://github.com/openstack/ironic/blob/master/ironic/objects/node.py#L271 | 12:52 |
| *** syed_ has quit IRC | 12:52 | |
| sambetts | https://github.com/openstack/ironic/blob/master/ironic/objects/base.py#L78 | 12:52 |
| sambetts | I'd expect a Node() object to be passed into _from_db_object right? | 12:53 |
| jroll | whoa wat | 12:53 |
| sambetts | jroll: its in most of the functions in that class, I don't get what wizardry this is | 12:54 |
| jroll | sec, let me finish what I'm doing | 12:54 |
| sambetts | sure | 12:54 |
| *** Goneri has joined #openstack-ironic | 12:54 | |
| jroll | context is super magic though | 12:54 |
| jroll | we can't be that wrong, right? | 12:54 |
| sambetts | i'd hope not, I was just looking that the reservation code to try and work out if it was possible the object being returns wasn't being updated right and saw that and though wait what :/ | 12:56 |
| *** causten has joined #openstack-ironic | 12:58 | |
| lucasagomes | sambetts, probably just converting db object to rpc objects? | 12:58 |
| lucasagomes | not sure | 12:59 |
| jroll | muahahahaha | 13:00 |
| jroll | repro'd it | 13:00 |
| sambetts | Nice! | 13:00 |
| jroll | and I feel like the fsm is the suspect | 13:00 |
| *** parlos has quit IRC | 13:01 | |
| sambetts | oh? is it nor processing the event correctly even though the logging seems to say it is? | 13:01 |
| jroll | we initialize the fsm at TaskManager init() time, with a start state | 13:01 |
| jroll | but we don't refresh that if we upgrade the lock | 13:01 |
| jroll | so the fsm still thinks we're in the old state | 13:02 |
| sambetts | so self.node is right but the fsm isn't *face palm* | 13:02 |
| jroll | yep | 13:02 |
| sambetts | so it is simply a bug in upgrade_lock essentially then \o/ | 13:03 |
| jroll | I think I can write a unit test for this | 13:03 |
| jroll | and the fix seems easy, just re-init the fsm | 13:03 |
| *** appu103_ has quit IRC | 13:03 | |
| jroll | the test is the hard part :) | 13:03 |
| lucasagomes | jroll, hell hah | 13:04 |
| sambetts | the UT should be pretty straight forward: create 2 task_managers, use 1 to trigger a fsm_event, and then upgrade_lock on the other, and check the task.fsm right? | 13:05 |
| lucasagomes | perhaps just initialize the fsm when process_event is called ? | 13:06 |
| *** vikrant has quit IRC | 13:06 | |
| lucasagomes | that's the only place it's used | 13:06 |
| *** rloo has joined #openstack-ironic | 13:07 | |
| jroll | sambetts: oh, I think you're right, was thinking it needed to change state while blocking on upgrade_lock() | 13:07 |
| * lucasagomes checks if it's not being used outside that calss | 13:07 | |
| lucasagomes | yeah we do not use it outside the TaskManager class (only in tests) | 13:07 |
| *** cdearborn has joined #openstack-ironic | 13:08 | |
| jroll | okay, that makes sense | 13:08 |
| lucasagomes | so just check if self.fsm is None and if so initialize it the first time process_event() is called | 13:08 |
| jroll | though | 13:08 |
| jroll | then we'll be initializing it twice in some requests | 13:08 |
| jroll | e.g. for resume then wait | 13:08 |
| jroll | so I'm almost thinking do it in _lock() | 13:09 |
| jroll | with a bonus of not being able to advance the fsm without an exclusive lock :) | 13:09 |
| lucasagomes | right, I was thinking of checking if "self.fsm is None" | 13:09 |
| sambetts | we need to do it anywhere we replace self.node with a fresh version from the DB | 13:09 |
| lucasagomes | but yeah _lock() makes sense too | 13:09 |
| jroll | sambetts: ah yeah, true | 13:09 |
| sambetts | jroll: which right now luckly is only in the __init__ and in _lock | 13:10 |
| jroll | yep | 13:10 |
| *** cdearborn has quit IRC | 13:10 | |
| openstackgerrit | Merged openstack/ironic: Updated from global requirements https://review.openstack.org/363863 | 13:10 |
| jroll | so yeah, just re-init in _lock() I think | 13:10 |
| jroll | well, idk, I'll figure something cleaner out | 13:11 |
| *** pionar has joined #openstack-ironic | 13:11 | |
| jroll | don't want to initialize it twice if I can help it | 13:11 |
| sambetts | jroll: I wonder if its worth having a node property on the task manager with getters and setters, and then in the setter just to the fsm init, then we don't have to remember it in the future | 13:12 |
| jroll | oh, I like that | 13:12 |
| *** sabeen has joined #openstack-ironic | 13:12 | |
| jroll | yeah that seems best | 13:13 |
| *** cdearborn has joined #openstack-ironic | 13:13 | |
| jroll | sambetts: btw, the way I figured this out was trying to reproduce it with just setting something in node.extra, and couldn't | 13:13 |
| jroll | so I knew it was related to process_event() somehow | 13:13 |
| *** amotoki_ has quit IRC | 13:14 | |
| sambetts | jroll: so it wasn't just tied to heartbeat then | 13:15 |
| jroll | right | 13:15 |
| jroll | was just bouncing a node through manage->provide->done | 13:15 |
| *** [1]cdearborn has joined #openstack-ironic | 13:15 | |
| *** [1]cdearborn has quit IRC | 13:18 | |
| sambetts | where in the code was that triggering the upgrade_lock? or did you have heartbeats going at the same time? | 13:18 |
| jroll | I was triggering it myself | 13:18 |
| jroll | so I had two python shells | 13:18 |
| jroll | one ran a function to get exclusive lock, sleep, process_event | 13:18 |
| jroll | the other grabbed a shared lock, upgraded it, ran process event | 13:19 |
| jroll | run one than the other in quick succession | 13:19 |
| sambetts | ++ got it, I thought you where running it against a live Ironic, and I was trying to work out how it was triggering it :P | 13:19 |
| jroll | heh, no | 13:19 |
| jroll | too many moving pieces there :P | 13:19 |
| sambetts | totally :-P | 13:20 |
| jroll | I also made a todo to write ironic_shell.py | 13:20 |
| openstackgerrit | Georgy Dyuldin proposed openstack/ironic: Add Serial-over-LAN enable/disable test https://review.openstack.org/364303 | 13:20 |
| jroll | which will set up config, import objects, give you a context to use, etc | 13:20 |
| *** cdearborn has quit IRC | 13:21 | |
| sambetts | that would be cool, basically like PBRing a breakpoint in an empty UT | 13:21 |
| jroll | yeah, basically | 13:21 |
| jroll | it was super useful working in django | 13:21 |
| sambetts | same for RoR | 13:21 |
| jroll | especially for modifying the db for test, objects interface is way nicer than our api (sadly) | 13:22 |
| sambetts | well hopfully we can fix that with v2 :-P | 13:22 |
| jroll | yeah | 13:22 |
| jroll | I guess the main part is I know python better than curl :P | 13:23 |
| jroll | and bpython/etc makes it nicer because tab complete and such | 13:23 |
| *** cdearborn has joined #openstack-ironic | 13:25 | |
| sambetts | haha, yeah, I really enjoyed working with Rails because of its strict resourceful routing stuff which made designing resourcful APIs e.g. <object type>/<uuid>/<object type/<uuid> super obvious | 13:25 |
| jroll | +1 | 13:26 |
| jroll | need to move to flask or something else more explicit | 13:26 |
| lucasagomes | yeah | 13:26 |
| lucasagomes | anyway great catch jroll, my hair was turning grey already :D | 13:27 |
| jroll | lucasagomes: thanks, these are the fun ones to me | 13:27 |
| jroll | (probably why I have gray hairs coming in) | 13:27 |
| jroll | :P | 13:27 |
| *** wajdi has joined #openstack-ironic | 13:27 | |
| lucasagomes | heh yeah | 13:27 |
| wajdi | good morning ironicers | 13:27 |
| lucasagomes | wajdi, morning | 13:27 |
| wajdi | hey lucasagomes! | 13:28 |
| Romanenko_K | Hi! Good morning ) | 13:28 |
| jroll | lucasagomes: I've got a concert tonight to take the frustration out :D | 13:28 |
| wajdi | which concert? | 13:28 |
| lucasagomes | jroll, nice, who's playing? | 13:29 |
| jroll | prophets against rage - it's a mix of rage against the machine, cypress hill, and maybe one other group? | 13:29 |
| wajdi | so lucky | 13:29 |
| sambetts | oooh nice! | 13:29 |
| wajdi | I've been really wanting to see them | 13:29 |
| jroll | er prophets *of* rage | 13:29 |
| jroll | oh, public enemy is the other one | 13:29 |
| jroll | "Prophets of Rage is Tom Morello, Tim Commerford, Brad Wilk, Chuck D, B-Real, and DJ Lord" http://prophetsofrage.com/ | 13:29 |
| jroll | gonna be awesome \o/ | 13:29 |
| lucasagomes | nice, tom morello is fecking awesome | 13:30 |
| * lucasagomes will check their songs | 13:30 | |
| jroll | yeah :D | 13:30 |
| wajdi | I missed it in Montreal =/ | 13:30 |
| *** joprovost has joined #openstack-ironic | 13:30 | |
| jroll | :( | 13:30 |
| wajdi | I've been really bad with my event tracking lately. Missed a bunch of other things I was planning on going to see. We have a fantastic movie festival in Montreal called Fantasia, I also missed. | 13:31 |
| jroll | wajdi: get in the car and drive to detroit :P | 13:31 |
| *** amotoki has joined #openstack-ironic | 13:31 | |
| wajdi | jroll: you know that's not impossible! :) | 13:31 |
| jroll | I know :D | 13:31 |
| lucasagomes | I'm waiting for black sabbath here in january, probably my last chance to see those guys playing | 13:31 |
| * lucasagomes bought the tickets already | 13:32 | |
| wajdi | sweet | 13:32 |
| *** causten has quit IRC | 13:32 | |
| openstackgerrit | Georgy Dyuldin proposed openstack/ironic: Add Serial-over-LAN console validation test https://review.openstack.org/364313 | 13:34 |
| jroll | heh, that'll be awesome | 13:34 |
| *** causten has joined #openstack-ironic | 13:37 | |
| *** causten has quit IRC | 13:38 | |
| openstackgerrit | Georgy Dyuldin proposed openstack/ironic: Add Serial-over-LAN console validation test https://review.openstack.org/364313 | 13:39 |
| *** rajinir has joined #openstack-ironic | 13:43 | |
| *** amotoki has quit IRC | 13:43 | |
| *** causten has joined #openstack-ironic | 13:44 | |
| *** causten has quit IRC | 13:46 | |
| jroll | I think this exposes a weirdness in task manager tests, too :P | 13:46 |
| jroll | and found some code to delete \o/ | 13:46 |
| lucasagomes | yay, there's an NOTE left by devananda about some juno compat code | 13:47 |
| lucasagomes | probably can be deleted too | 13:47 |
| lucasagomes | (L221) | 13:47 |
| jroll | yeah, that one | 13:47 |
| jroll | so when I moved this fsm init around, our test nodes (with provision_state None) all started bombing out because the fsm init was above that compat code | 13:48 |
| openstackgerrit | Galyna Zholtkevych proposed openstack/ironic: Optionally power off the node when needed https://review.openstack.org/360330 | 13:49 |
| *** openstackgerrit has quit IRC | 13:49 | |
| sambetts | jroll: move the compat code into the setter too? | 13:49 |
| *** openstackgerrit has joined #openstack-ironic | 13:49 | |
| jroll | sambetts: gonna kill it with fire | 13:50 |
| sambetts | oh awesome \o/ | 13:50 |
| jroll | I updated the tests to use provision_state=AVAILABLE for the test node objects | 13:50 |
| jroll | writing commit message now, sec | 13:50 |
| *** amotoki has joined #openstack-ironic | 13:50 | |
| *** mrtenio has quit IRC | 13:51 | |
| Guest77956 | /me read the backscroll and impressed with the team effort to figure out the reason for the bug. Nice job everyone! | 13:54 |
| * Guest77956 read the backscroll and impressed with the team effort to figure out the reason for the bug. Nice job everyone! | 13:54 | |
| * Guest77956 also needs to login as himself. | 13:54 | |
| lucasagomes | Guest30374, hey john :-) | 13:54 |
| *** Guest77956 is now known as jlvillal | 13:55 | |
| jlvillal | lucasagomes, how did you know??? | 13:55 |
| jlvillal | :) | 13:55 |
| Romanenko_K | We have bug in ironicclient with node-validate when passing argument with tilde: https://bugs.launchpad.net/python-ironicclient/+bug/1619291 | 13:55 |
| openstack | Launchpad bug 1619291 in python-ironicclient "Node-validate with tilde in argument causes server error" [Undecided,New] | 13:55 |
| lucasagomes | whois :-) | 13:55 |
| jlvillal | lucasagomes, Ah! | 13:56 |
| jlvillal | lucasagomes, I thought you recognized my style of writing :) | 13:56 |
| sambetts | hey jlvillal | 13:56 |
| jroll | Romanenko_K: hi, can you provide server logs there? | 13:56 |
| jlvillal | Good morning sambetts lucasagomes jroll TheJulia rloo and everyone else | 13:56 |
| lucasagomes | morning | 13:56 |
| vdrok | morning TheJulia wajdi and jlvillal | 13:56 |
| wajdi | morning vdrok | 13:57 |
| jlvillal | Hey vdrok | 13:57 |
| TheJulia | Good morning! | 13:57 |
| jlvillal | and wajdi :) | 13:57 |
| vdrok | Romanenko_K: jroll here is the fix already - https://review.openstack.org/#/c/364222/ | 13:57 |
| dtantsur | morning jlvillal | 13:57 |
| wajdi | hey jlvillal! | 13:57 |
| sambetts | vdrok is faster than me :-P | 13:57 |
| jlvillal | dtantsur, :) | 13:57 |
| rloo | hi and morning all, jlvillal (alias Guest77956), lucasagomes, vdrok, wajdi, TheJulia, Romanenko_K, jroll | 13:57 |
| lucasagomes | rloo, hi there | 13:57 |
| vdrok | morning rloo :) | 13:57 |
| jroll | vdrok: magic! | 13:57 |
| jroll | hey rloo :) | 13:57 |
| wajdi | o/ rloo | 13:57 |
| sambetts | hey rloo | 13:57 |
| Romanenko_K | Hello, Ruby! | 13:58 |
| Romanenko_K | Thanks, Vlad. | 13:58 |
| lucasagomes | vdrok, that was quick | 13:58 |
| * jroll +W | 13:59 | |
| vdrok | heh, thanks :) | 13:59 |
| *** amotoki has quit IRC | 13:59 | |
| *** mgoddard_ has joined #openstack-ironic | 14:00 | |
| *** abashmak has joined #openstack-ironic | 14:00 | |
| *** anush__ has joined #openstack-ironic | 14:00 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Rely on portgroup standalone_ports_supported https://review.openstack.org/360647 | 14:00 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Integrate portgroups with ports to support LAG https://review.openstack.org/332177 | 14:00 |
| jroll | sambetts: guess I need to remove that compat code in this patch as well: - Failed: 1033 | 14:01 |
| *** abashmak has left #openstack-ironic | 14:01 | |
| *** joprovost1 has joined #openstack-ironic | 14:01 | |
| sambetts | jroll: :-P yay | 14:01 |
| *** jheroux has joined #openstack-ironic | 14:02 | |
| *** mrtenio has joined #openstack-ironic | 14:02 | |
| *** joprovost has quit IRC | 14:02 | |
| *** joprovost1 is now known as joprovost | 14:02 | |
| *** mgoddard has quit IRC | 14:03 | |
| *** michauds has joined #openstack-ironic | 14:03 | |
| *** jaosorior has quit IRC | 14:04 | |
| vdrok | jroll: we can set this one https://bugs.launchpad.net/ironic/+bug/1618754 as ref-approved right? | 14:04 |
| openstack | Launchpad bug 1618754 in Ironic "[RFE] Add portgroups support to Ironic" [Undecided,In progress] - Assigned to Vasyl Saienko (vsaienko) | 14:04 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Fix invalid chassis/nodes/chassis_uuid API url https://review.openstack.org/316149 | 14:04 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Fix all possible node subresources routing collisions https://review.openstack.org/315766 | 14:04 |
| jroll | vdrok: yeah | 14:04 |
| vdrok | rfe-approved I mean | 14:04 |
| jroll | heh, I know :) | 14:05 |
| *** vishwanathj has joined #openstack-ironic | 14:06 | |
| *** joprovost1 has joined #openstack-ironic | 14:12 | |
| *** joprovost has quit IRC | 14:13 | |
| *** joprovost1 is now known as joprovost | 14:13 | |
| robcresswell | sambetts: Should ironic-ui be publishing its patches here btw? Or is there some separate channel? | 14:13 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: Refresh fsm in task when a shared lock is upgraded https://review.openstack.org/364341 | 14:13 |
| jroll | sambetts: lucasagomes: ^^ | 14:13 |
| jroll | I almost wonder if I should split and delete the crutch first, but /shrug | 14:14 |
| robcresswell | sambetts: As in, the openstackgerrit bot messages :) | 14:14 |
| jroll | robcresswell: it should, I'm not sure it's configured to do so | 14:14 |
| robcresswell | jroll: Ah, thanks | 14:14 |
| *** amotoki has joined #openstack-ironic | 14:14 | |
| jroll | robcresswell: I believe that's a project-config patch, if you want to set it up | 14:14 |
| robcresswell | it is indeed. gerritbot time. | 14:15 |
| *** bcornec1 has quit IRC | 14:20 | |
| *** fxpester has joined #openstack-ironic | 14:22 | |
| lucasagomes | jroll, nice! I like the setter | 14:22 |
| * lucasagomes +2;d | 14:22 | |
| robcresswell | betherly: https://review.openstack.org/#/c/364339/ | 14:22 |
| robcresswell | betherly: Adding translation :) | 14:23 |
| betherly | robcresswell: thanks! | 14:23 |
| jroll | lucasagomes: yeah, that was sambetts idea | 14:23 |
| *** mtanino has joined #openstack-ironic | 14:24 | |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add basic tests for OSC plugin baremetal node commands https://review.openstack.org/301687 | 14:24 |
| *** mkam has joined #openstack-ironic | 14:24 | |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add basic tests for OSC plugin baremetal node commands https://review.openstack.org/301687 | 14:24 |
| *** praneshp has joined #openstack-ironic | 14:24 | |
| *** watanabe_isao has quit IRC | 14:25 | |
| jcook | jroll thanks for fixing that | 14:25 |
| openstackgerrit | Ruby Loo proposed openstack/python-ironicclient: Adds --chassis arg to 'openstack baremetal node set' https://review.openstack.org/363999 | 14:25 |
| jroll | jcook: totes | 14:26 |
| *** syed_ has joined #openstack-ironic | 14:26 | |
| openstackgerrit | Pavlo Shchelokovskyy proposed openstack/ironic: Allow using TempURLs for deploy images https://review.openstack.org/344160 | 14:31 |
| dtantsur | what would folks say about a configuration option to change default boot_option (local, netboot)? | 14:32 |
| dtantsur | I can imagine a few use cases for always using local, for example | 14:32 |
| jroll | dtantsur: can we just get rid of netboot? :) | 14:32 |
| TheJulia | I would second getting rid of netboot | 14:33 |
| * jroll thinks of one use case for always using local - sanity | 14:33 | |
| dtantsur | jroll, this is step 2 :) as soon as we have an option, we can one day flip its value, the set a timer for removal | 14:33 |
| jroll | also, with multitenant networking, netboot never works | 14:33 |
| *** rama_y has joined #openstack-ironic | 14:33 | |
| sambetts | jroll: I have some ideas about that ;) | 14:33 |
| dtantsur | yep, this is one of use cases | 14:33 |
| TheJulia | Well, we could make an intermediate tool like sambetts and I were discussing in Austin to enable multitenant netbooting, but yeah | 14:33 |
| jroll | sambetts: pls no | 14:33 |
| *** mkam has quit IRC | 14:34 | |
| dtantsur | jroll, wanna have an RFE for that or just me quickly hack something? | 14:34 |
| jroll | dtantsur: I prefer we have a full plan to get rid of netboot, honestly | 14:34 |
| jroll | or at least a discussion | 14:34 |
| sambetts | the only advantage of netboot is that you can update the kernal on the fly by uploading it to glance right? | 14:34 |
| sambetts | tbh getting rid of netboot makes us more similar to VMs | 14:34 |
| sambetts | I think | 14:35 |
| TheJulia | I don't think so, you can update it on disk though, but that is also a risk. | 14:35 |
| jroll | sambetts: not so sure about that, I feel like there's more to do (update pxe config etc) | 14:35 |
| jroll | which isn't accessible by the user | 14:35 |
| lucasagomes | sambetts, hmm well, there's also the case that, with netboot you make sure that node machines are going to boot before your control plane | 14:35 |
| lucasagomes | jroll, ^ | 14:35 |
| jroll | dtantsur: I'd prefer to wait for ocata for any of this, I think, it's pretty late | 14:35 |
| dtantsur | well, an option would be a quick and small change, but whatever... | 14:35 |
| jroll | lucasagomes: can you rephrase? I don't get you | 14:35 |
| lucasagomes | jroll, say you have a power outrage and all machines are turned off | 14:35 |
| dtantsur | I'm writing some docs, and I'm looking forward to explain our capabilities syntax again... | 14:35 |
| TheJulia | I think it would be interesting to hear from some of the operators on netbooting, because I know some shops have put specialized code for varying things where they are using netbooting | 14:36 |
| lucasagomes | so you started turning then back on | 14:36 |
| dtantsur | one use case for netboot is easy changing kernel options btw | 14:36 |
| jroll | lucasagomes: yeah, I understand the implications here, I don't understand the folks that don't want their machines up if their control plane is down | 14:36 |
| TheJulia | lucasagomes: wouldn't that be better as shedding/recovery management? | 14:36 |
| jroll | dtantsur: how is it easier in netboot? | 14:36 |
| jroll | for a user, anyway | 14:36 |
| sambetts | dtantsur: tenants can do a grub-update right? | 14:36 |
| lucasagomes | I don't use that either, but that was something in old ages of tripleO | 14:36 |
| lucasagomes | maybe lifeless has a better use-case for that | 14:37 |
| sambetts | dtantsur: so it actually removes control from users | 14:37 |
| dtantsur | yeah, fair. I don't care about netboot much | 14:37 |
| *** jheroux has quit IRC | 14:37 | |
| dtantsur | I care about a long and fail-prone procedure to enable local boot rather | 14:37 |
| * sambetts is good with killing netboot btw, just playing devils advacate ;) | 14:37 | |
| dtantsur | in tripleo we have a special pre-deploy validation ensuing local boot is done right | 14:37 |
| robcresswell | jroll: https://review.openstack.org/#/c/364347 | 14:37 |
| jroll | dtantsur: anyway, it's a bit late for new features, even if it's just a config, I'd prefer to wait for ocata, I think | 14:37 |
| TheJulia | so, if we kill netbooting, we kill some of the boot from volume scenarios | 14:38 |
| robcresswell | I think thats right, will need some ironic core +1s I imagine. | 14:38 |
| *** hj-hpe has joined #openstack-ironic | 14:38 | |
| jroll | robcresswell: I hope ptl +1 is enough :D | 14:38 |
| * TheJulia thinks a very deep discussion, with whisky, may be required | 14:38 | |
| jroll | TheJulia: +2 | 14:38 |
| dtantsur | jroll, well, fine. I'll prepare an RFE for the option and for a plan to switch it to "local" by default eventually. I won't lay out the plan for removing netboot completely though. | 14:38 |
| jroll | dtantsur: sounds good | 14:39 |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add tests for --wrap option for driver-properties output https://review.openstack.org/297077 | 14:39 |
| sambetts | dtantsur: cool :) | 14:39 |
| jroll | dtantsur: I'd also like to explore a plan to switch to multitenant network by default and get rid of 'flat' (because flat is just neutron with provisioning network == tenant network) | 14:39 |
| jroll | but that's a separate thing | 14:39 |
| sambetts | which is technically acheiveable in the multitenant case too if you wanted | 14:40 |
| dtantsur | oh yeah :) let folks adopt multitenancy first :D | 14:40 |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Functional tests for portgroups in ironicclient https://review.openstack.org/297727 | 14:40 |
| TheJulia | dtantsur: ++ | 14:40 |
| jroll | sambetts: exactly | 14:41 |
| *** jheroux has joined #openstack-ironic | 14:42 | |
| lucasagomes | dtantsur, sambetts and btw, I kinda disagree that it's easier to change the kernel/initrd for netboot | 14:42 |
| *** causten has joined #openstack-ironic | 14:42 | |
| TheJulia | I think that would have to be navigated carefully because there are surely people out there that will have nodes on flat networks with switchgear that might not have an ml2 driver... | 14:42 |
| lucasagomes | I find it harder, with local boot you just need to apt-get/dnf/pacman kernel | 14:42 |
| lucasagomes | for netboot the tenant requires the operator to change it | 14:43 |
| lucasagomes | and we don't even have a proper mechanism to that | 14:43 |
| lucasagomes | which sucks :-) | 14:43 |
| dtantsur | jroll and others: https://bugs.launchpad.net/ironic/+bug/1619339 | 14:43 |
| openstack | Launchpad bug 1619339 in Ironic "[RFE] Switching to local boot by default" [Wishlist,Confirmed] | 14:43 |
| jroll | TheJulia: yeah, good point | 14:43 |
| TheJulia | lucasagomes: Agree with you there | 14:43 |
| robcresswell | jroll: Ha, sorry, I'd not realised who I was talking to. Thanks for the +1. | 14:43 |
| jroll | robcresswell: no worries :) | 14:44 |
| *** parlos has joined #openstack-ironic | 14:45 | |
| sambetts | lucasagomes: totally agree | 14:45 |
| *** mkam has joined #openstack-ironic | 14:46 | |
| *** narasimha_SV has quit IRC | 14:48 | |
| *** joprovost has quit IRC | 14:49 | |
| *** jcoufal has joined #openstack-ironic | 14:50 | |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add test for 'baremetal port create' command https://review.openstack.org/332880 | 14:50 |
| *** sborkows has quit IRC | 14:50 | |
| *** ChubYann has joined #openstack-ironic | 14:51 | |
| openstackgerrit | Kyrylo Romanenko proposed openstack/python-ironicclient: Add test for ironic port-list command https://review.openstack.org/329068 | 14:53 |
| *** pionar has quit IRC | 14:53 | |
| *** Sukhdev has joined #openstack-ironic | 14:54 | |
| *** pionar has joined #openstack-ironic | 14:55 | |
| *** pionar has quit IRC | 14:56 | |
| *** joprovost has joined #openstack-ironic | 14:57 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic-specs: Update openstackclient commands https://review.openstack.org/363921 | 14:57 |
| *** krtaylor has joined #openstack-ironic | 14:58 | |
| openstackgerrit | Galyna Zholtkevych proposed openstack/ironic: Optionally power off the node when needed https://review.openstack.org/360330 | 15:00 |
| *** mjura has quit IRC | 15:01 | |
| *** pcaruana has quit IRC | 15:03 | |
| *** causten has quit IRC | 15:04 | |
| *** hshiina has joined #openstack-ironic | 15:05 | |
| *** rcernin has quit IRC | 15:06 | |
| *** mgoddard_ has quit IRC | 15:08 | |
| *** mgoddard has joined #openstack-ironic | 15:09 | |
| openstackgerrit | Merged openstack/ironic: Fix exception handling in NodesController._lookup https://review.openstack.org/364222 | 15:13 |
| openstackgerrit | Merged openstack/ironic: Add standalone_ports_supported to portgroup - DB https://review.openstack.org/361053 | 15:15 |
| openstackgerrit | Merged openstack/ironic: Update test_update_portgroup_address_no_vif_id test https://review.openstack.org/363747 | 15:15 |
| dtantsur | calling it a day, see you tomorrow | 15:17 |
| JayF | o/ | 15:17 |
| TheJulia | goodnight dtantsur | 15:17 |
| *** dtantsur is now known as dtantsur|afk | 15:17 | |
| jlvillal | dtantsur|afk: ciao | 15:17 |
| *** rbrud has joined #openstack-ironic | 15:17 | |
| *** jcoufal_ has joined #openstack-ironic | 15:23 | |
| *** sdake has joined #openstack-ironic | 15:23 | |
| lucasagomes | dtantsur|afk, see ya | 15:23 |
| *** Sukhdev has quit IRC | 15:24 | |
| *** praneshp has quit IRC | 15:25 | |
| *** jcoufal has quit IRC | 15:25 | |
| *** causten has joined #openstack-ironic | 15:27 | |
| *** causten has quit IRC | 15:28 | |
| *** causten has joined #openstack-ironic | 15:28 | |
| *** ifarkas is now known as ifarkas_afk | 15:28 | |
| openstackgerrit | Vladyslav Drok proposed openstack/python-ironicclient: Negative tests for testing actions with node https://review.openstack.org/361583 | 15:30 |
| *** rcernin has joined #openstack-ironic | 15:31 | |
| devananda | in theory, I agree: local boot and multitenant networking should be the defaults. | 15:32 |
| devananda | we need some good messaging any time we change default behaviours though, since that is a breaking config change (if I didn't have a setting in my ironic.conf, and I upgrade, my environment will probably break) | 15:33 |
| lucasagomes | devananda, ++ | 15:34 |
| lucasagomes | change defaults are hard | 15:34 |
| lucasagomes | we still use elilo for uefi by default (which is not even mainteined anymore) | 15:36 |
| lucasagomes | we probably should change that too and makre grub or ipxe default for uefi | 15:36 |
| lucasagomes | make* | 15:36 |
| *** causten has joined #openstack-ironic | 15:37 | |
| devananda | +1 | 15:38 |
| *** jheroux has quit IRC | 15:42 | |
| *** causten has quit IRC | 15:43 | |
| openstackgerrit | Sam Betts proposed openstack/ironic: Add Network Interface Attach and Detach APIs https://review.openstack.org/327046 | 15:43 |
| *** causten has joined #openstack-ironic | 15:45 | |
| *** causten has quit IRC | 15:45 | |
| *** ricardoas has left #openstack-ironic | 15:48 | |
| *** nicodemos has quit IRC | 15:49 | |
| vdrok | is anyone able to do openstack endpoint delete on an devstack-created endpoint? I get 404 :( | 15:53 |
| vdrok | while being able to delete any endpoints I created myself | 15:54 |
| *** mkam has quit IRC | 15:55 | |
| *** causten_ has joined #openstack-ironic | 15:57 | |
| openstackgerrit | Sam Betts proposed openstack/python-ironicclient: Add python API for node interface attach/detach https://review.openstack.org/364420 | 16:05 |
| openstackgerrit | Mario Villaplana proposed openstack/ironic-specs: Update notifications spec https://review.openstack.org/364422 | 16:06 |
| *** aNupoisc has joined #openstack-ironic | 16:07 | |
| *** adnavare has joined #openstack-ironic | 16:07 | |
| mariojv | ^ there's the spec update i was discussing with rloo and JayF yesterday, updates to the base classes/framework are coming shortly | 16:07 |
| *** mkam has joined #openstack-ironic | 16:07 | |
| *** yogi has joined #openstack-ironic | 16:08 | |
| *** ijw has joined #openstack-ironic | 16:11 | |
| openstackgerrit | Oleksandr Berezovskyi proposed openstack/bifrost: Add possibility to set source branch for ironic-inspector https://review.openstack.org/363300 | 16:12 |
| *** adnavare has quit IRC | 16:20 | |
| *** aNupoisc has quit IRC | 16:20 | |
| *** causten_ has quit IRC | 16:20 | |
| lucasagomes | I'm calling it a day | 16:21 |
| *** e0ne has quit IRC | 16:21 | |
| lucasagomes | have a great evening all, talk to you all later | 16:21 |
| *** lucasagomes is now known as lucas-dinner | 16:21 | |
| *** hshiina has quit IRC | 16:23 | |
| NobodyCam | Good Morning Ironic'ers | 16:24 |
| NobodyCam | have a good night lucas-dinner | 16:24 |
| NobodyCam | morning vdrok :) | 16:24 |
| openstackgerrit | Ruby Loo proposed openstack/python-ironicclient: osc node power on & off commands https://review.openstack.org/364440 | 16:25 |
| vdrok | good morning NobodyCam ! | 16:25 |
| *** krtaylor has quit IRC | 16:25 | |
| *** causten has joined #openstack-ironic | 16:25 | |
| NobodyCam | :) | 16:25 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-lib: Add match_root_device_hints() to the utils.py module https://review.openstack.org/358000 | 16:28 |
| lucas-dinner | quick fix ^ | 16:28 |
| * lucas-dinner goes now! | 16:28 | |
| *** snizar has joined #openstack-ironic | 16:28 | |
| *** jpich has quit IRC | 16:29 | |
| * sambetts -> home | 16:35 | |
| sambetts | night everyone | 16:35 |
| *** sambetts is now known as sambetts|afk | 16:35 | |
| *** sabeen has quit IRC | 16:36 | |
| *** ohamada has quit IRC | 16:36 | |
| openstackgerrit | Oleksandr Berezovskyi proposed openstack/bifrost: Add possibility to set source branch for ironic-inspector https://review.openstack.org/363300 | 16:37 |
| NobodyCam | night sambetts|afk | 16:38 |
| *** ijw has quit IRC | 16:40 | |
| *** daemontool has quit IRC | 16:40 | |
| *** snizar has quit IRC | 16:41 | |
| *** trown is now known as trown|lunch | 16:45 | |
| *** aNupoisc has joined #openstack-ironic | 16:46 | |
| *** adnavare has joined #openstack-ironic | 16:46 | |
| *** amotoki has quit IRC | 16:51 | |
| *** causten has quit IRC | 16:53 | |
| jlvillal | vdrok: Someone told me that they increased presentations to allow three presenters | 16:54 |
| NobodyCam | morning jlvillal :) | 16:55 |
| jlvillal | Not sure if anyone knows if that is true or not. I had thought the limit was 2. | 16:55 |
| jlvillal | Hey NobodyCam | 16:55 |
| NobodyCam | :) | 16:55 |
| *** derekh has quit IRC | 16:56 | |
| openstackgerrit | Oleksandr Berezovskyi proposed openstack/bifrost: Add possibility to set source branch for ironic-inspector https://review.openstack.org/363300 | 16:59 |
| *** tesseract- has quit IRC | 17:01 | |
| *** causten has joined #openstack-ironic | 17:05 | |
| *** athomas has quit IRC | 17:05 | |
| openstackgerrit | Merged openstack/ironic: Refresh fsm in task when a shared lock is upgraded https://review.openstack.org/364341 | 17:08 |
| *** krtaylor has joined #openstack-ironic | 17:08 | |
| jroll | woooo | 17:10 |
| JayF | jroll: the nice thing is, that should backport cleanly to 6.0.0 | 17:10 |
| JayF | jroll: we should totally get that in our downstream branch | 17:10 |
| jroll | JayF: totes | 17:11 |
| jroll | maybe even mitaka? | 17:11 |
| jroll | seems like it's always been there, but something changed to make it hit more | 17:11 |
| JayF | I certainly think we should backport that to mitaka if it does so cleanly | 17:12 |
| JayF | jroll: were we running at-least-mitaka downstream before our last deploy? | 17:12 |
| jroll | JayF: dunno | 17:12 |
| jroll | probably not? | 17:12 |
| JayF | just thinking, it only pops up in the gate occassinoally, but we deployed and it started happening a lot | 17:12 |
| JayF | so just wondering if it was actually in mitaka or not | 17:12 |
| JayF | I think you're right, that we weren't even that far ahead | 17:13 |
| jroll | yeah, I think it's exacerbated in newton | 17:13 |
| jroll | with the "don't require exclusive lock for all vendor passthru" thing | 17:13 |
| JayF | well can we backport it to mitaka? given it drops support for NOSTATE? | 17:13 |
| jroll | oh, good point | 17:14 |
| jroll | I just tried and it conflicts anyway | 17:14 |
| jroll | probably not worth it | 17:14 |
| *** causten has quit IRC | 17:15 | |
| openstackgerrit | Ruby Loo proposed openstack/python-ironicclient: use utils.key_value_pairs_to_dict() https://review.openstack.org/364474 | 17:20 |
| *** causten has joined #openstack-ironic | 17:24 | |
| *** early has quit IRC | 17:26 | |
| * jlvillal reads the fsm shared lock patch and again thinks Python is cool with the @property feature :) | 17:29 | |
| mat128 | jroll, JayF: is there a way to reproduce this easily? | 17:29 |
| mat128 | I can try on liberty | 17:30 |
| JayF | mat128: we deployed 6.0.0 to our downstream cloud, and saw the # of times that error appeared become our #1 most common error | 17:30 |
| JayF | mat128: tbh we were happy when it hit in the gate as an upstream bug | 17:31 |
| JayF | mat128: it had to have been introduced sometime in the last 6 months or so, because that's about how large of a "bite" of the code we took | 17:31 |
| JayF | mat128: so IDK how to reproduce it directly, other than doing a lot of builds :) | 17:31 |
| *** ricardoas has joined #openstack-ironic | 17:32 | |
| jroll | mat128: it's incredibly hard to reproduce in live ironic | 17:32 |
| jroll | mat128: the unit test in that patch, you could run against your code | 17:32 |
| jroll | but I think code in newton made it easier to hit in real life | 17:32 |
| *** e0ne has joined #openstack-ironic | 17:32 | |
| *** ijw_ has joined #openstack-ironic | 17:35 | |
| *** causten has quit IRC | 17:38 | |
| mat128 | JayF, jroll: no occurrence at all for "Can not transition from state", unless the error message changed from liberty. | 17:38 |
| mat128 | backporting new unit test to stable/liberty to see if that triggers it | 17:40 |
| *** jtomasek is now known as jtomasek|afk | 17:45 | |
| openstackgerrit | Ruby Loo proposed openstack/python-ironicclient: osc node power on & off commands https://review.openstack.org/364440 | 17:46 |
| mat128 | JayF, jroll: The unit test fails on liberty: http://paste.openstack.org/show/565760/ | 17:49 |
| mat128 | I'd say it's backport worthy | 17:49 |
| JayF | mat128: it can't be backported in it's current form, as it removes support for NOSTATE as well | 17:50 |
| jroll | mat128: yeah, it's just not aggravated in liberty, because we don't upgrade the lock | 17:50 |
| jroll | I don't think it's worth it, honestly | 17:50 |
| mat128 | jroll: ah that's it | 17:50 |
| mat128 | the upgrade | 17:50 |
| mat128 | ok then, never mind | 17:50 |
| * jroll finds the change that made this happen more | 17:50 | |
| *** ijw_ has quit IRC | 17:51 | |
| *** leitan has joined #openstack-ironic | 17:52 | |
| leitan | Hi guys, i have a general doubt about ironic + neutron, can anyone lend me a hand on this ? | 17:52 |
| *** e0ne has quit IRC | 17:53 | |
| *** sabeen has joined #openstack-ironic | 17:53 | |
| *** jjohnson2 has quit IRC | 17:54 | |
| jroll | mat128: I guess it was the change to refactor heartbeat out | 17:54 |
| *** early has joined #openstack-ironic | 17:56 | |
| jroll | leitan: don't ask to ask, just ask :) | 17:56 |
| leitan | jroll, haha ok ! | 17:57 |
| *** harlowja_ has joined #openstack-ironic | 17:58 | |
| leitan | jroll, 1 - in mitaka is possible to work with "flat networks per tenant" to place the baremetal on a "private" l2 space on each tenant once deployed ? 2 - in an standard dvr neutron deployment where i have DHCP agents on many nodes, in order to provide DHCP to the baremetals, i need to have the bridge mappings for the baremetal network (ironic-provisioning), on every node that i have a dhcp-agent 3 - cleaning net id is the same l2 network i create | 17:59 |
| leitan | as ironic-provisioning on neutron ? | 17:59 |
| *** harlowja has quit IRC | 18:00 | |
| openstackgerrit | Merged openstack/ironic-lib: Add match_root_device_hints() to the utils.py module https://review.openstack.org/358000 | 18:00 |
| jroll | leitan: I was just stepping away, I'll try to answer quickly: 1) I think so, but the control plane needs to be able to access each of those L2 networks, 2) I think yes, 3) for mitaka and/or flat network case, yes, in multitenant mode in newton they can be different | 18:01 |
| *** e0ne has joined #openstack-ironic | 18:01 | |
| jroll | hopefully someone else is around that can answer followups/elaborate | 18:01 |
| * jroll bbiab | 18:01 | |
| leitan | jroll, thanks for the quick answer ! until someone elaborate a lil bit more it will be great to have some reference topologies to the solution, maybe my question is really simple but it can get it clear from the docs | 18:04 |
| openstackgerrit | Fabio Dassan proposed openstack/pyghmi: Implement M4 and Kent support https://review.openstack.org/361457 | 18:05 |
| *** trown|lunch is now known as trown | 18:06 | |
| leitan | jroll, one more, quick, the ironic conductor needs to have direct connection to the IPMI network (that can be isolated) is that right ? | 18:07 |
| openstackgerrit | Merged openstack/bifrost: Fix minor documentation issue https://review.openstack.org/364189 | 18:09 |
| *** dstepanenko_ has joined #openstack-ironic | 18:10 | |
| dstepanenko_ | hey | 18:11 |
| openstackgerrit | Oleksandr Berezovskyi proposed openstack/bifrost: Add possibility to set source branch for ironic-inspector https://review.openstack.org/363300 | 18:11 |
| *** hoonetorg has quit IRC | 18:11 | |
| *** ijw has joined #openstack-ironic | 18:16 | |
| *** hoonetorg has joined #openstack-ironic | 18:16 | |
| dstepanenko_ | I'm working on investigating ability to perform openstack rolling upgrade on multinode cluster containing nova, neutron, glance, cinder, keystone, swift, heat, ceilometer and ironic from mitaka to newton version. Is it possible to perform rolling upgrade for ironic? If no, which bugs/features blocks it? If I want to upgrade cluster with openstack components listed above, which particular order of components should I chose? | 18:16 |
| *** openstackgerrit has quit IRC | 18:18 | |
| *** openstackgerrit has joined #openstack-ironic | 18:19 | |
| *** linuxgeek has quit IRC | 18:22 | |
| *** sabeen has quit IRC | 18:24 | |
| natorious | hey JayF | 18:26 |
| JayF | natorious: sup | 18:26 |
| *** mtanino__ has joined #openstack-ironic | 18:26 | |
| natorious | re: https://review.openstack.org/#/c/363959/ | 18:26 |
| natorious | 1) Please update all existing metrics to use this decorator if it makes sense. | 18:27 |
| natorious | do you know where the existing metrics are? | 18:27 |
| JayF | I'd just grep the code for metrics_utils | 18:27 |
| natorious | only in ipa or elsewhere? | 18:27 |
| JayF | https://github.com/openstack/ironic-python-agent/blob/dfaf92b611d7c4abb98345fbcf21ef3ed4bb9d3b/ironic_python_agent/api/controllers/v1/status.py#L52 is a good example | 18:27 |
| JayF | in IPA | 18:27 |
| *** mtanino has quit IRC | 18:27 | |
| natorious | ok | 18:27 |
| JayF | because like the devdocs say, we have to use that obtuse context manager syntax | 18:28 |
| JayF | but your decorator should get around it, so it'll be cleaner | 18:28 |
| natorious | and do you have a link to the dev docs project so I can look at updating that? | 18:28 |
| JayF | it's in the same repo | 18:28 |
| natorious | k | 18:28 |
| JayF | doc/source/metrics.rst iirc, something like that | 18:28 |
| JayF | http://docs.openstack.org/developer/ironic-python-agent/metrics.html is the rendered version | 18:28 |
| natorious | my commit message, I was limiting to 50 characters. Is that not needed then? | 18:29 |
| JayF | just the first line limited to that | 18:29 |
| natorious | ah, ok | 18:29 |
| JayF | you can do two newlines then put a larger description, like it says in the link I posted w/the comment | 18:29 |
| *** jjohnson2 has joined #openstack-ironic | 18:29 | |
| *** linuxgeek has joined #openstack-ironic | 18:31 | |
| *** mtanino has joined #openstack-ironic | 18:31 | |
| *** harlowja_ has quit IRC | 18:31 | |
| *** harlowja has joined #openstack-ironic | 18:32 | |
| *** mtanino__ has quit IRC | 18:33 | |
| *** ijw has quit IRC | 18:35 | |
| openstackgerrit | Mario Villaplana proposed openstack/ironic-specs: Update notifications spec https://review.openstack.org/364422 | 18:43 |
| mariojv | JayF: updated to respond to your comment ^ | 18:43 |
| *** linuxgeek has quit IRC | 18:48 | |
| *** fragatin_ has quit IRC | 18:48 | |
| *** david-lyle has quit IRC | 18:49 | |
| *** linuxgeek has joined #openstack-ironic | 18:49 | |
| *** ijw has joined #openstack-ironic | 18:49 | |
| *** david-lyle has joined #openstack-ironic | 18:50 | |
| vdrok | jroll: pecan.expose return value is different in case of exception, we'll have to change the client too :( | 18:51 |
| *** aNupoisc has quit IRC | 18:51 | |
| vdrok | it returns only the following - {"error_message": "{\"code\": 400, \"description\": \"Expected a logical name or UUID but received 123~123.\", \"title\": \"Bad Request\"}"} | 18:51 |
| *** adnavare has quit IRC | 18:52 | |
| vdrok | without possibility of traceback as it seems | 18:52 |
| vdrok | I'll put up a patch in a couple of hours | 18:52 |
| vdrok | also jroll JayF have you seen this one? https://review.openstack.org/336281 | 18:54 |
| vdrok | seems to work | 18:54 |
| *** linuxgeek has quit IRC | 19:00 | |
| *** linuxgeek has joined #openstack-ironic | 19:01 | |
| *** linuxgeek has quit IRC | 19:02 | |
| jroll | vdrok: woo | 19:04 |
| jroll | vdrok: :( for pecan.expose thing thuogh | 19:05 |
| rloo | jroll: the hash-ring code in ironic & nova. are they the same/similar? wondering if we should try to move to oslo (or are you already doing that) | 19:06 |
| jroll | rloo: it's planned for ocata, I plan to do it unless someone else wants to | 19:06 |
| mariojv | are warnings expected when building ironic docs? i'm getting 15 docs not in toctree, unrelated to any changes i'm making | 19:06 |
| * jroll has a todo on his list already | 19:06 | |
| rloo | jroll: thx. off my todo (actually, my todo would have been 'open a bug about it') | 19:06 |
| jroll | mariojv: not really, but the sphinx/pbr option currently doesn't work so it does happen | 19:07 |
| JayF | mariojv: for bonus points, push a different patch fixing them | 19:07 |
| JayF | mariojv: or at least file a bug | 19:07 |
| jroll | ++ | 19:07 |
| mariojv | thanks | 19:07 |
| natorious | JayF: would you rather IPA metrics module go under ironic_python_agent.common or the root | 19:09 |
| JayF | natorious: it should probably go under common? But I never concern my self too much with location for things (I probably should) | 19:09 |
| *** fragatina has joined #openstack-ironic | 19:10 | |
| natorious | JayF: was there previously a metrics module in ironic_python_agent.common or am I just imagining things? | 19:11 |
| JayF | natorious: that was a patchset we pulled into downstream that never landed upstream | 19:11 |
| natorious | ok, its all making sense. Now I realize why dtantsu|afk had no idea what the merge req was about lol | 19:13 |
| natorious | *dtantsur | 19:13 |
| *** harlowja_ has joined #openstack-ironic | 19:13 | |
| *** adnavare has joined #openstack-ironic | 19:16 | |
| *** aNupoisc has joined #openstack-ironic | 19:17 | |
| *** harlowja has quit IRC | 19:17 | |
| *** mkam has quit IRC | 19:28 | |
| *** ijw has quit IRC | 19:31 | |
| *** ijw has joined #openstack-ironic | 19:32 | |
| *** mkam has joined #openstack-ironic | 19:34 | |
| *** sabeen has joined #openstack-ironic | 19:34 | |
| jlvillal | TheJulia: Cool to see the infra cloud is using Bifrost :) | 19:36 |
| jlvillal | http://lists.openstack.org/pipermail/openstack-dev/2016-September/102707.html | 19:37 |
| *** ijw has quit IRC | 19:37 | |
| rloo | mariojv: thx for the notifications update. | 19:39 |
| TheJulia | jlvillal: I'm glad it is finally up :) | 19:39 |
| jlvillal | TheJulia: kudos! | 19:39 |
| jroll | leitan: yes, you need direct connection between ironic-conductor and ipmi | 19:41 |
| TheJulia | jlvillal: the kudos go to the infra team, I just make things go. | 19:42 |
| mat128 | wow | 19:42 |
| jroll | that's awesome :) | 19:43 |
| jlvillal | TheJulia: I still think you deserve kudos :P | 19:43 |
| mariojv | np rloo | 19:43 |
| *** mkam has quit IRC | 19:43 | |
| jlvillal | For that Bifrost stuff, you know. | 19:43 |
| mariojv | i'll update today again | 19:43 |
| leitan | jroll, ok, so i can put the ironic-conductor on the same node as the nova-compute that is acting as compute_driver=ironic.IronicDriver , then in that machine i can have also the ml2 bridge mappings to connect to the provisioning network, and the dhcp agents for it | 19:43 |
| mariojv | code fixing is like half done | 19:43 |
| *** mkam has joined #openstack-ironic | 19:43 | |
| rloo | mariojv: thx. i just skimmed the CRUD spec; it doesn't conform to the new event types. | 19:43 |
| rloo | mariojv: but a good exercise to go through, to see what the new event types might look like. | 19:44 |
| mariojv | rloo: correct | 19:44 |
| jroll | leitan: yeah, that should work fine | 19:44 |
| mariojv | it'll have to be updated if/when spec merges | 19:44 |
| rloo | TheJulia: ++ kudos | 19:45 |
| leitan | jroll, thanks a lot ! | 19:46 |
| jroll | no problem | 19:47 |
| *** Goneri has quit IRC | 19:54 | |
| *** jcoufal_ has quit IRC | 19:56 | |
| rloo | mariojv: sorry, do you have an idea as to when some action 'starts', and when it 'ends'? Eg, does it start when we start handling the request at the API level? and does it end at the API level just before we issue a response? | 19:56 |
| rloo | mariojv: except for something not immediate, it can't end at the api level. | 19:57 |
| mariojv | rloo: that's a really good question. i think it will depend on the action itself | 19:57 |
| mariojv | there are some things that can be initiated at API level or within ironic itself | 19:57 |
| rloo | mariojv: if it depends on the action, we are going to have to document this. as an operator, what are you interested in? | 19:58 |
| mariojv | for example, power state changes | 19:58 |
| mariojv | rloo: exactly. i added a requirement that each notification be documented in the spec | 19:58 |
| mariojv | rloo: i'll have an example later today with the power state notif | 19:58 |
| mariojv | rloo: my interest is 1) tracking changes over time and persisting them in an external service and potentially in the future 2) using these to trigger anomaly resolution | 19:58 |
| rloo | mariojv: ok thx. cuz i was looking at the crud spec, and it said it'd do the notifications at the objects layer. which won't work for eg failures detected at the API layer. | 19:59 |
| mariojv | i.e. power state fails, i want to cold reset the bmc automatically | 19:59 |
| mariojv | rloo: np. it sounds like we should talk about that in that particular spec | 19:59 |
| mariojv | i think it'll be a bit clearer once the base class updates and doc updates for power state notifications are ready | 19:59 |
| rloo | mariojv: yup to talking in that spec. unless there are guidelines that the base spec can mention. if we/you have guidelines :) | 20:00 |
| rloo | mariojv: ok, maybe i should look at your power changes first before the crud spec | 20:00 |
| mariojv | alright. i'll ping you here when it's ready. rloo what time zone are you in? | 20:00 |
| mariojv | want to get this out before EoD | 20:01 |
| rloo | mariojv: ET (not like ET outer space). am here for 2 more hours, then not til Tuesday (yay) | 20:01 |
| mariojv | nice. i'm in mountain, moving to ET soon. | 20:01 |
| JayF | note to self: get rloo some reeses pieces at next summit | 20:01 |
| * rloo wonders why the rackers seem to be moving | 20:01 | |
| JayF | rloo: we can all work from home, and we have different taste in homes | 20:02 |
| rloo | JayF: no. don't like reeses pieces. they have peanut butter right? | 20:02 |
| JayF | well, brown-colored-vaguely-peanut-tasting-sugar-paste | 20:02 |
| TheJulia | chocolate? | 20:02 |
| JayF | but it was more an et joke :) | 20:02 |
| * rloo doesn't get the joke | 20:04 | |
| *** ijw has joined #openstack-ironic | 20:04 | |
| JayF | ET loved reeses pieces :) | 20:04 |
| rloo | JayF: ahhhh... :) | 20:05 |
| jlvillal | rloo: M&Ms? | 20:05 |
| jlvillal | JayF: rloo: In the book I think it was M&Ms. Then Reese's paid money to make it Reese's Pieces in the movie :) | 20:05 |
| *** derekh has joined #openstack-ironic | 20:06 | |
| rloo | jlvillal: sigh. i'm such a bore. i just like dark chocolate. not much of a sugar person. | 20:06 |
| jlvillal | Snopes says I am wrong and that Reeses did not pay. http://www.snopes.com/business/market/mandms.asp | 20:06 |
| devananda | rloo: dark chocolate ++ | 20:07 |
| TheJulia | I wanted chocolate before, now I really want chocolate. | 20:07 |
| * devananda will now eat chocolate after lunch is done | 20:08 | |
| * rloo knows why she likes TheJulia and devananda | 20:08 | |
| * jroll still has mackinac island fudge in the kitchen :D | 20:08 | |
| * rloo now takes a break to get some chocolate too | 20:09 | |
| *** rama_y has quit IRC | 20:09 | |
| *** ijw has quit IRC | 20:09 | |
| openstackgerrit | Fabio Dassan proposed openstack/pyghmi: Implement M4 and Kent support https://review.openstack.org/361457 | 20:14 |
| openstackgerrit | Nate House proposed openstack/ironic-python-agent: Partial-bug: #1611553 https://review.openstack.org/363959 | 20:16 |
| openstack | bug 1611553 in ironic-python-agent "[RFE] Metrics for IPA extensions and GenericHardwareManager" [Wishlist,Triaged] https://launchpad.net/bugs/1611553 | 20:16 |
| natorious | JayF: ^^ | 20:17 |
| *** dprince has quit IRC | 20:26 | |
| *** ijw has joined #openstack-ironic | 20:26 | |
| *** sabeen has quit IRC | 20:28 | |
| *** e0ne has quit IRC | 20:30 | |
| *** sabeen has joined #openstack-ironic | 20:31 | |
| *** ijw has quit IRC | 20:31 | |
| openstackgerrit | Mario Villaplana proposed openstack/ironic-specs: Update notifications spec https://review.openstack.org/364422 | 20:32 |
| natorious | JayF: do you think we should remove all of these mentions from the docs still? | 20:39 |
| natorious | metrics_utils.get_metrics_logger(__name__).timer | 20:39 |
| natorious | or send_gauge etc | 20:40 |
| JayF | natorious: the question to ask is: should anyone adding new metrics use that method vs the context mgr / decorator you're proposing | 20:40 |
| natorious | for consistency I'd added a metrics.send_gauge method so ppl wouldn't have to use metrics bits from different places | 20:40 |
| JayF | natorious: IMO The only case you'd wanna call it that way would be for the send_* class of metrics sending one | 20:40 |
| JayF | natorious: well, if you shortcut that, then maybe not even there :D | 20:40 |
| *** ijw has joined #openstack-ironic | 20:41 | |
| natorious | ok, so I'll remove that from the metrics docs then and only mention the new | 20:41 |
| *** ijw has quit IRC | 20:46 | |
| *** derekh has quit IRC | 20:47 | |
| *** mkam has quit IRC | 20:47 | |
| *** ijw has joined #openstack-ironic | 20:52 | |
| openstackgerrit | Nate House proposed openstack/ironic-python-agent: Provides convenience for logging metrics data. https://review.openstack.org/363959 | 20:52 |
| natorious | JayF: think *hopefully I got those changes right ^^ | 20:53 |
| *** baoli_ has quit IRC | 20:53 | |
| *** mkam has joined #openstack-ironic | 20:59 | |
| *** sabeen has quit IRC | 21:00 | |
| openstackgerrit | Michael Turek proposed openstack/python-ironicclient: [WIP] Refactor Deploy and Clean to use inherited take_action https://review.openstack.org/364552 | 21:00 |
| *** rama_y has joined #openstack-ironic | 21:01 | |
| openstackgerrit | Michael Turek proposed openstack/python-ironicclient: [WIP] Add --wait to OSC provisioning commands https://review.openstack.org/354806 | 21:02 |
| *** nicodemos has joined #openstack-ironic | 21:03 | |
| *** joprovost has quit IRC | 21:04 | |
| *** trown is now known as trown|outtypewww | 21:07 | |
| openstackgerrit | Mario Villaplana proposed openstack/ironic: Add power state change notifications https://review.openstack.org/321865 | 21:22 |
| openstackgerrit | Mario Villaplana proposed openstack/ironic: Replace "phase" with "status" in notification base https://review.openstack.org/364558 | 21:22 |
| mariojv | rloo: JayF: jroll: those are the notification updates that we wanted to get in this week | 21:22 |
| mariojv | there is also a spec update which i would recommend first: https://review.openstack.org/#/c/364422/ | 21:22 |
| rloo | thx mariojv | 21:23 |
| mariojv | np, thanks for reviewing | 21:23 |
| *** rcernin has quit IRC | 21:24 | |
| *** Goneri has joined #openstack-ironic | 21:26 | |
| mrda | Morning Ironic | 21:26 |
| JayF | o/ | 21:26 |
| mrda | o/. | 21:26 |
| rloo | morning mrda | 21:26 |
| mrda | hi rloo | 21:27 |
| mariojv | morning mrda | 21:27 |
| openstackgerrit | Stephanie Miller proposed openstack/ironic: Add volume_connector table to save connector information https://review.openstack.org/200983 | 21:28 |
| openstackgerrit | Stephanie Miller proposed openstack/ironic: Introduce VolumeConnector object https://review.openstack.org/214584 | 21:29 |
| *** rama_y has quit IRC | 21:32 | |
| *** wajdi_ has joined #openstack-ironic | 21:39 | |
| *** priteau has quit IRC | 21:42 | |
| *** cdearborn has quit IRC | 21:42 | |
| *** wajdi has quit IRC | 21:43 | |
| *** wajdi_ has quit IRC | 21:43 | |
| *** ijw has quit IRC | 21:44 | |
| openstackgerrit | Vladyslav Drok proposed openstack/python-ironicclient: Print pecan exceptions properly https://review.openstack.org/364566 | 21:44 |
| vdrok | morning mrda :) | 21:45 |
| vdrok | jroll: ^^ that's the fix, it worked when I tried at work, but right now I can't properly test it | 21:45 |
| vdrok | another possibility to fix things is not to validate in the _lookup, then the checks will have to be in a whole bunch of places in the controller code, but we won't have to change the client | 21:48 |
| mrda | vdrok: o/ | 21:51 |
| openstackgerrit | Vladyslav Drok proposed openstack/python-ironicclient: Negative tests for testing actions with node https://review.openstack.org/361583 | 21:51 |
| *** sabeen has joined #openstack-ironic | 21:57 | |
| *** Goneri has quit IRC | 21:57 | |
| jlvillal | vdrok: It is 1am where you live. Just an FYI ;) | 21:59 |
| vdrok | client is sad without a fix, I have to help! :D | 22:00 |
| rloo | vdrok: just don't emit any pecan exceptions :) | 22:01 |
| vdrok | rloo: It seems that pecan can't handle wsme style exceptions :( tried to return the dict that is constructed internally by wsme in case of exception, but was not able to make it work | 22:03 |
| rloo | vdrok: go to sleep :) | 22:04 |
| *** ijw has joined #openstack-ironic | 22:04 | |
| *** michauds has quit IRC | 22:04 | |
| vdrok | on my way, good night everyone! | 22:04 |
| rloo | good night vdrok! | 22:05 |
| *** joprovost has joined #openstack-ironic | 22:07 | |
| openstackgerrit | Stephanie Miller proposed openstack/ironic: Introduce new RPCs to support volume connector operation https://review.openstack.org/214585 | 22:10 |
| openstackgerrit | Stephanie Miller proposed openstack/ironic: Add volume_targets table to save target information https://review.openstack.org/285218 | 22:11 |
| openstackgerrit | Stephanie Miller proposed openstack/ironic: Introduce VolumeTarget object https://review.openstack.org/285219 | 22:11 |
| *** sabeen1 has joined #openstack-ironic | 22:14 | |
| *** sabeen has quit IRC | 22:17 | |
| *** hj-hpe has quit IRC | 22:18 | |
| *** thrash is now known as thrash|g0ne | 22:19 | |
| *** leitan has quit IRC | 22:23 | |
| *** joprovost has quit IRC | 22:24 | |
| *** mkam has quit IRC | 22:43 | |
| *** ijw has quit IRC | 22:56 | |
| *** joprovost has joined #openstack-ironic | 23:02 | |
| *** ijw has joined #openstack-ironic | 23:04 | |
| *** ijw has quit IRC | 23:09 | |
| *** joprovost has quit IRC | 23:09 | |
| openstackgerrit | Stephanie Miller proposed openstack/ironic: Introduce new RPCs to support volume target operations https://review.openstack.org/285220 | 23:11 |
| *** anush__ has quit IRC | 23:24 | |
| *** vishwanathj has quit IRC | 23:30 | |
| *** mkam has joined #openstack-ironic | 23:32 | |
| *** chlong has quit IRC | 23:33 | |
| *** mkam has quit IRC | 23:37 | |
| *** joprovost has joined #openstack-ironic | 23:43 | |
| *** joprovost has quit IRC | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!