| *** amotoki has joined #openstack-ironic | 00:05 | |
| *** [1]cdearborn has quit IRC | 00:07 | |
| *** [1]cdearborn has joined #openstack-ironic | 00:08 | |
| *** amotoki has quit IRC | 00:10 | |
| *** wajdi has joined #openstack-ironic | 00:25 | |
| *** bfournie has joined #openstack-ironic | 00:25 | |
| *** sdake has quit IRC | 00:36 | |
| mat128 | TheJulia: hmm, those are still out of tree, right? | 00:37 |
|---|---|---|
| mat128 | TheJulia: I'll take a look at them tomorrow | 00:37 |
| *** sdake has joined #openstack-ironic | 00:37 | |
| TheJulia | mat128: AFAIK yeah, I don't remember where things are at w/r/t getting one of the DIB issues fixed that impacts the ability to break the element out. :( | 00:38 |
| openstackgerrit | Mathieu Mitchell proposed openstack/ironic-python-agent: Enforce upper-constraints when building ramdisks https://review.openstack.org/358855 | 00:39 |
| mat128 | TheJulia: even if we arent dont breaking it out, it should respect upper-constraints | 00:40 |
| mat128 | otherwise it builds unpredictable ramdisks | 00:40 |
| TheJulia | agreed, and it might. I have officially hit scatter brained at this point. | 00:40 |
| TheJulia | would depend on pip I guess | 00:41 |
| mat128 | https://github.com/openstack/diskimage-builder/blob/master/elements/ironic-agent/install.d/ironic-agent-source-install/60-ironic-agent-install#L16 | 00:41 |
| mat128 | doesnt look like it honors it | 00:41 |
| mat128 | do we build official DIB images for IPA? | 00:41 |
| mat128 | trying to find how I can inject the UPPER_CONSTRAINT_FILE handling | 00:42 |
| mat128 | hmm | 00:42 |
| * mat128 was supposted to be here for a sec | 00:42 | |
| TheJulia | Not yet, but essentially some of the features don't work without a dib based image | 00:42 |
| mat128 | going offline now, must not look at dib | 00:42 |
| TheJulia | mat128: call it a night :) | 00:42 |
| mat128 | yeah | 00:42 |
| TheJulia | goodnight! | 00:42 |
| TheJulia | :) | 00:42 |
| mat128 | noted for tomorrow | 00:42 |
| mat128 | thanks for your hints :) | 00:42 |
| openstackgerrit | zhangyanxian proposed openstack/ironic: Replace assertEqual(None, *) with assertIsNone in test_nodestates.py https://review.openstack.org/356231 | 00:48 |
| *** hoangcx has joined #openstack-ironic | 00:48 | |
| *** adnavare has quit IRC | 00:59 | |
| *** amotoki has joined #openstack-ironic | 01:06 | |
| *** amotoki has quit IRC | 01:10 | |
| *** vishwanathj has quit IRC | 01:11 | |
| *** mtanin___ has quit IRC | 01:12 | |
| *** syed_ has quit IRC | 01:12 | |
| *** Sukhdev has quit IRC | 01:12 | |
| *** rloo has quit IRC | 01:13 | |
| *** thrash is now known as thrash|g0ne | 01:15 | |
| *** hoangcx has quit IRC | 01:17 | |
| *** chlong has joined #openstack-ironic | 01:20 | |
| *** baoli has joined #openstack-ironic | 01:23 | |
| *** baoli_ has joined #openstack-ironic | 01:24 | |
| *** hoangcx has joined #openstack-ironic | 01:26 | |
| *** syed_ has joined #openstack-ironic | 01:27 | |
| *** baoli has quit IRC | 01:28 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/359457 | 01:32 |
| openstackgerrit | Nguyen Hung Phuong proposed openstack/ironic: Merge imports in code https://review.openstack.org/357597 | 01:34 |
| openstackgerrit | Ruby Loo proposed openstack/python-ironicclient: Add --node arg to 'ironic port-list' https://review.openstack.org/359520 | 01:35 |
| *** moshele has joined #openstack-ironic | 01:43 | |
| *** sabeen has quit IRC | 01:46 | |
| *** adnavare has joined #openstack-ironic | 01:47 | |
| *** sabeen has joined #openstack-ironic | 01:49 | |
| openstackgerrit | hubian proposed openstack/ironic-specs: fix uderline too shirt and missed subsections https://review.openstack.org/359538 | 01:50 |
| *** amotoki has joined #openstack-ironic | 02:06 | |
| *** phuongnh has joined #openstack-ironic | 02:08 | |
| *** amotoki has quit IRC | 02:11 | |
| *** rybridges has joined #openstack-ironic | 02:12 | |
| rybridges | Hello | 02:14 |
| rybridges | I am having a look at _check_deploying_status in ironic/conductor/manager.py | 02:14 |
| rybridges | that's inside of the https://github.com/openstack/ironic repository | 02:15 |
| rybridges | And I think there may be a bug | 02:15 |
| rybridges | It looks as though self.iter_nodes is only returning nodes that belong to this conductor | 02:16 |
| rybridges | however, the whole point of the method is to find nodes that do not belong to this conductor, and instead belong to a conductor that is offline | 02:16 |
| rybridges | So the list of nodes that iter_nodes returns will never contain nodes that belong to an offline conductor, because it only returns nodes that belong to the conductor on which the method is run | 02:19 |
| rybridges | So the entire _check_deploying_status method will be broken since it does not actually check the deploying status of nodes that belong to offline conductors | 02:21 |
| *** [1]cdearborn has quit IRC | 02:21 | |
| rybridges | can anyone have a look and confirm? | 02:21 |
| *** adnavare has quit IRC | 02:35 | |
| *** moshele has quit IRC | 02:38 | |
| *** sdake has quit IRC | 02:46 | |
| *** sdake has joined #openstack-ironic | 02:49 | |
| *** yuanying_ has joined #openstack-ironic | 02:51 | |
| *** yuanying has quit IRC | 02:52 | |
| *** rbudden has quit IRC | 02:53 | |
| *** yuanying has joined #openstack-ironic | 02:53 | |
| *** yuanying has quit IRC | 02:55 | |
| *** yuanying has joined #openstack-ironic | 02:56 | |
| *** rybridges has quit IRC | 02:57 | |
| *** yuanying_ has quit IRC | 02:57 | |
| openstackgerrit | Merged openstack/ironic: Update api-ref for v1.22 https://review.openstack.org/353117 | 02:58 |
| openstackgerrit | hubian proposed openstack/ironic-specs: fix subsections missed problems https://review.openstack.org/359560 | 03:07 |
| *** amotoki has joined #openstack-ironic | 03:07 | |
| *** mtanino has joined #openstack-ironic | 03:08 | |
| *** amotoki has quit IRC | 03:11 | |
| *** vishwanathj has joined #openstack-ironic | 03:11 | |
| *** aswadr_ has joined #openstack-ironic | 03:17 | |
| *** vishwanathj is now known as vishwanathj_zzz | 03:19 | |
| *** amotoki has joined #openstack-ironic | 03:24 | |
| *** rajinir has quit IRC | 03:25 | |
| *** annp has joined #openstack-ironic | 03:29 | |
| *** penick has quit IRC | 03:32 | |
| *** amotoki has quit IRC | 03:38 | |
| *** hoangcx has quit IRC | 03:40 | |
| *** vikrant has joined #openstack-ironic | 03:41 | |
| *** Sukhdev has joined #openstack-ironic | 03:42 | |
| *** yuanying has quit IRC | 03:48 | |
| *** yuanying has joined #openstack-ironic | 03:51 | |
| *** amotoki has joined #openstack-ironic | 04:00 | |
| *** baoli_ has quit IRC | 04:01 | |
| *** syed_ has quit IRC | 04:02 | |
| openstackgerrit | Tuan proposed openstack/ironic: Fix incorrect order of params of assertEqual() https://review.openstack.org/359580 | 04:02 |
| *** amotoki has quit IRC | 04:09 | |
| *** sabeen has quit IRC | 04:10 | |
| *** amotoki has joined #openstack-ironic | 04:10 | |
| *** jraju has joined #openstack-ironic | 04:14 | |
| *** hoangcx has joined #openstack-ironic | 04:15 | |
| *** jraju has quit IRC | 04:15 | |
| openstackgerrit | Nguyen Phuong An proposed openstack/ironic: [api-ref]: Add response codes to chassis api https://review.openstack.org/359582 | 04:19 |
| openstackgerrit | avnish proposed openstack/ironic: Fix tempest.conf generation https://review.openstack.org/359583 | 04:21 |
| *** amotoki has quit IRC | 04:31 | |
| *** kalpase has joined #openstack-ironic | 04:31 | |
| *** amotoki has joined #openstack-ironic | 04:37 | |
| *** rama_y has joined #openstack-ironic | 04:38 | |
| *** mtanino has quit IRC | 04:42 | |
| *** ChubYann has quit IRC | 04:54 | |
| *** appu103_ has joined #openstack-ironic | 05:03 | |
| *** rama_y has quit IRC | 05:04 | |
| *** appu103_ is now known as aparnav | 05:04 | |
| *** ibmchas has joined #openstack-ironic | 05:06 | |
| *** ChubYann has joined #openstack-ironic | 05:07 | |
| *** jaosorior has joined #openstack-ironic | 05:07 | |
| *** jaosorior has quit IRC | 05:09 | |
| *** jaosorior has joined #openstack-ironic | 05:10 | |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Generic power interface for soft reboot and soft power off https://review.openstack.org/216730 | 05:11 |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic: Ipmitool power driver for soft reboot and soft power off https://review.openstack.org/216738 | 05:15 |
| *** Sukhdev has quit IRC | 05:27 | |
| *** moshele has joined #openstack-ironic | 05:32 | |
| *** _vsaienko has joined #openstack-ironic | 05:38 | |
| *** harlowja_at_home has joined #openstack-ironic | 05:39 | |
| *** mpsairam has quit IRC | 05:39 | |
| *** _vsaienko has quit IRC | 05:51 | |
| *** _vsaienko has joined #openstack-ironic | 05:55 | |
| *** zhangjl has joined #openstack-ironic | 05:55 | |
| *** mjura has joined #openstack-ironic | 05:55 | |
| *** hoangcx has quit IRC | 05:59 | |
| *** moshele has quit IRC | 06:06 | |
| *** hoangcx has joined #openstack-ironic | 06:10 | |
| *** mpsairam has joined #openstack-ironic | 06:19 | |
| *** rcernin has joined #openstack-ironic | 06:23 | |
| *** ccamacho has joined #openstack-ironic | 06:32 | |
| *** ccamacho has quit IRC | 06:32 | |
| *** ccamacho has joined #openstack-ironic | 06:36 | |
| *** syed_ has joined #openstack-ironic | 06:43 | |
| *** harlowja_at_home has quit IRC | 06:44 | |
| *** fragatina has quit IRC | 06:44 | |
| *** rbartal has joined #openstack-ironic | 06:44 | |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic: Add 'inspect_ports' as a 'driver_info' parameter to iLO drivers https://review.openstack.org/325230 | 06:46 |
| *** fragatina has joined #openstack-ironic | 06:46 | |
| *** ibmchas has quit IRC | 06:50 | |
| *** fragatina has quit IRC | 06:51 | |
| *** moshele has joined #openstack-ironic | 06:56 | |
| *** tesseract- has joined #openstack-ironic | 06:56 | |
| *** jhooker has quit IRC | 06:59 | |
| *** jhooker has joined #openstack-ironic | 06:59 | |
| *** rbrud has quit IRC | 07:02 | |
| lazy_prince | Anyone seen http://paste.openstack.org/show/562889/ looks like IPA heartbeat is broken with mitaka.. | 07:20 |
| *** ohamada has joined #openstack-ironic | 07:25 | |
| *** appu103_ has joined #openstack-ironic | 07:29 | |
| *** aparnav has quit IRC | 07:29 | |
| *** jpich has joined #openstack-ironic | 07:32 | |
| *** athomas has joined #openstack-ironic | 07:37 | |
| *** e0ne has joined #openstack-ironic | 07:40 | |
| *** ifarkas_afk is now known as ifarkas | 07:47 | |
| *** e0ne has quit IRC | 07:53 | |
| *** _vsaienko has quit IRC | 07:58 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** rama_y has joined #openstack-ironic | 08:00 | |
| *** zzzeek has joined #openstack-ironic | 08:00 | |
| *** amotoki has quit IRC | 08:01 | |
| *** _vsaienko has joined #openstack-ironic | 08:01 | |
| *** appu103_ has quit IRC | 08:04 | |
| openstackgerrit | hubian proposed openstack/ironic-specs: add link to the launchpad bug https://review.openstack.org/359670 | 08:06 |
| openstackgerrit | Milan Kováčik proposed openstack/ironic-inspector: Introducing node introspection state management https://review.openstack.org/348943 | 08:07 |
| milan | morning Ironic! :) | 08:08 |
| *** derekh has joined #openstack-ironic | 08:08 | |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/ironic: [WIP] Testing latest u-c https://review.openstack.org/318440 | 08:10 |
| *** rama_y has quit IRC | 08:11 | |
| *** lucas-afk is now known as lucasagomes | 08:12 | |
| lucasagomes | morning all | 08:12 |
| *** vmud213 has joined #openstack-ironic | 08:13 | |
| milan | morning lucasagomes ! :) | 08:15 |
| openstackgerrit | Shivanand Tendulker proposed openstack/proliantutils: Discover Storage details https://review.openstack.org/341370 | 08:16 |
| openstackgerrit | hubian proposed openstack/ironic-specs: fix line whitespace and too long problems https://review.openstack.org/359685 | 08:29 |
| *** pcaruana has joined #openstack-ironic | 08:29 | |
| *** stendulker has joined #openstack-ironic | 08:38 | |
| tangchen | lucasagomes, milan: morning :) | 08:39 |
| milan | morning tangchen :) | 08:39 |
| *** appu103_ has joined #openstack-ironic | 08:40 | |
| *** appu103_ is now known as aparnav | 08:41 | |
| *** electrofelix has joined #openstack-ironic | 08:45 | |
| *** Lord_Phoenix has joined #openstack-ironic | 08:50 | |
| *** milan has quit IRC | 08:52 | |
| *** syed_ has quit IRC | 08:52 | |
| openstackgerrit | hubian proposed openstack/ironic-specs: fix unexpected unindent problems https://review.openstack.org/359709 | 08:54 |
| *** Lord_Phoenix has quit IRC | 08:57 | |
| openstackgerrit | Ankit Kumar proposed openstack/ironic: Fix iLO drivers for local_gb https://review.openstack.org/341752 | 09:01 |
| openstackgerrit | hubian proposed openstack/ironic-specs: fix unexpected unindent problems https://review.openstack.org/359719 | 09:06 |
| *** amotoki has joined #openstack-ironic | 09:09 | |
| *** phuongnh has quit IRC | 09:11 | |
| *** _vsaienko has quit IRC | 09:16 | |
| *** amotoki has quit IRC | 09:17 | |
| *** sambetts|afk is now known as sambetts | 09:18 | |
| sambetts | rajinir: I had to switch to an ubuntu image from cirros because cirros didn't have the drivers required to boot on my hardware, I used disk-image-builder to create the image and I host it internally to my testbed so my slaves can download it during devstack install | 09:20 |
| *** jaosorior is now known as jaosorior_lunch | 09:21 | |
| *** _vsaienko has joined #openstack-ironic | 09:23 | |
| *** amotoki has joined #openstack-ironic | 09:26 | |
| *** eset has joined #openstack-ironic | 09:29 | |
| *** eset has quit IRC | 09:29 | |
| *** eset has joined #openstack-ironic | 09:29 | |
| *** tuanla_ has joined #openstack-ironic | 09:33 | |
| tuanla_ | Hello | 09:34 |
| *** kalpase has quit IRC | 09:37 | |
| openstackgerrit | avnish proposed openstack/ironic-inspector: Fix tempest.conf generation https://review.openstack.org/359109 | 09:41 |
| *** milan has joined #openstack-ironic | 09:44 | |
| *** sborkows has joined #openstack-ironic | 09:46 | |
| openstackgerrit | Sam Betts proposed openstack/ironic-specs: Add spec for VLAN Aware Baremetal Instances https://review.openstack.org/277853 | 09:47 |
| *** mgoddard_ has joined #openstack-ironic | 10:02 | |
| *** Lord_Phoenix has joined #openstack-ironic | 10:05 | |
| *** mgoddard has quit IRC | 10:05 | |
| *** dtantsur has joined #openstack-ironic | 10:07 | |
| dtantsur | morning/afternoon folks! my IRC bouncer has been down since yesterday evening, so please repeat any pings. | 10:08 |
| *** yogi has joined #openstack-ironic | 10:12 | |
| tuanla_ | Hello dtantsur: I have look at ironic-python-agent service: I wonder Why Log don't using oslo_i18n. I there any problem if i add oslo_i18n in IPA | 10:13 |
| dtantsur | tuanla_, hi! I don't see any problem with that, I guess our dependencies already pull oslo.i18n in, so it won't increase the ramdisk any further. | 10:13 |
| dtantsur | I think we haven't done it because nobody was interested enough | 10:14 |
| tuanla_ | Yes, I'm just have join in OpenStack. Thanks for your answer. | 10:16 |
| dtantsur | Welcome! :) | 10:16 |
| *** bks has joined #openstack-ironic | 10:18 | |
| *** hoangcx has quit IRC | 10:21 | |
| tuanla_ | dtantsur: Do you have free time now. I have problem with my lab. I try to deploy whole-disk-image in Ironic. I'm wonder I there any different in configuration between partition image vs whole-disk-image | 10:21 |
| tuanla_ | Because I done deploy partition image but have some problem when deploy whole-disk-image. | 10:22 |
| sambetts | tuanla_: Are you seeing an error in Ironic, if you have, copy the traceback into here paste.openstack.org, and we can see if we can help | 10:23 |
| *** _vsaienko has quit IRC | 10:23 | |
| tuanla_ | sambetts: Thanks you so much. it is my log http://paste.openstack.org/show/560758/ | 10:24 |
| *** _vsaienko has joined #openstack-ironic | 10:27 | |
| sambetts | tuanla_: is there an error message on the ironic conductor? This is failing trying to initally contact ironic, before any of the provisioning processing happens | 10:27 |
| tuanla_ | sambetts: No, | 10:28 |
| *** sdake has quit IRC | 10:28 | |
| tuanla_ | sambetts: Sometime i got node active in horizon | 10:28 |
| tuanla_ | but When it reboot in instance I have problem try to locate configuation file | 10:29 |
| *** jaosorior_lunch is now known as jaosorior | 10:30 | |
| *** chas has joined #openstack-ironic | 10:32 | |
| tuanla_ | sambetts: I realy the process deploy partition and whole-disk-image using PXE have different. When when deploy whole-disk-image every file created in deploy will remove such as conf. image, ramdisk | 10:32 |
| tuanla_ | in /tftpboot dir | 10:33 |
| *** chas is now known as ibmchas | 10:33 | |
| sambetts | tuanla_: if you are having issues on reboot, can I ask how big your harddrive is? | 10:34 |
| tuanla_ | I have beey try to change my hardisk | 10:35 |
| tuanla_ | Now it have 500GB | 10:35 |
| sambetts | actually 500GB or 500GB configured in Ironic? | 10:36 |
| tuanla_ | No, I just configured I see it near 571000MB | 10:36 |
| tuanla_ | sambetts: I see when computer start: I show capacity: 572325MB | 10:39 |
| sambetts | when the instance image boots, cloud-init expands the root partition to use up all the disk space avaible to it, if you have a harddrive >1TB then it'll expand the partition from 500GB to what ever the harddrive supports. This can break reboots if the partition grows bigger than the installed bootloader supports | 10:39 |
| sambetts | to force your node to use GPT a bootloader that can support big disks, follow this guide http://docs.openstack.org/developer/ironic/deploy/install-guide.html#when-used-with-compute-service | 10:41 |
| tuanla_ | Yes, I have try to configurate and re-deploy But I show message " ERROR ironic-python-agent LookupNodeError: Error getting configuration from Ironic: Could not look up node info. Check logs for details. | 10:42 |
| tuanla_ | " | 10:42 |
| sambetts | tuanla_: OOooh btw what version of Ironic and Ironic-python-agent are you using? | 10:44 |
| sambetts | dtantsur: http://paste.openstack.org/show/560758/ is this a case of IPA not falling back to the old URL again?? | 10:44 |
| tuanla_ | sambetts: I'm using devstack to enable ironic, ironic: 1.6.0 | 10:48 |
| tuanla_ | But I don't know how to find ironic-python-agent | 10:48 |
| tuanla_ | verison | 10:49 |
| *** athomas has quit IRC | 10:49 | |
| *** yogi has quit IRC | 10:49 | |
| *** bks has quit IRC | 10:50 | |
| tuanla_ | sambetts: I have plan try to deploy but don't using IPA | 10:51 |
| sambetts | tuanla_: we don't support deployment without IPA | 10:52 |
| tuanla_ | sambetts: I using devstack but now I don't see IPA in my /stack dir | 10:53 |
| sambetts | tuanla_: by default in devstack we download IPA images prebuilt by us, from tarballs.openstack.org | 10:54 |
| sambetts | so it doesn't have to build it everytime | 10:54 |
| *** athomas has joined #openstack-ironic | 10:54 | |
| sambetts | tuanla_: when did you last restack?? | 10:54 |
| tuanla_ | sambetts: Today | 10:54 |
| tuanla_ | about 8 hours ago | 10:55 |
| *** yogi has joined #openstack-ironic | 10:56 | |
| sambetts | ok, thats fine, there was a bug that caused lookup to fail but that was merged and built into IPA 4/5 days ago | 10:56 |
| dtantsur | sambetts, no, looks like connection refused | 10:56 |
| dtantsur | i.e. ironic API can't be reached at all | 10:56 |
| sambetts | yeah, just spotted that | 10:56 |
| sambetts | the URL looks wrong | 10:57 |
| sambetts | 10.0.0.1 is that your Ironic API service URL?? | 10:57 |
| tuanla_ | sambetts: yes | 10:57 |
| sambetts | so this failure isn't related to the type of image your deploying, has something happened to your Ironic API service? | 10:59 |
| *** d0ugal has quit IRC | 10:59 | |
| *** d0ugal has joined #openstack-ironic | 10:59 | |
| *** bks has joined #openstack-ironic | 11:02 | |
| tuanla_ | sambetts: Thanks for your help. | 11:03 |
| tuanla_ | sambetts: I using bios and PXE: is it good for whole disk image | 11:06 |
| sambetts | tuanla_: yup | 11:07 |
| dtantsur | folks, we have to release ironic-lib today. please review everything open there! | 11:08 |
| dtantsur | though lucasagomes' patches need updating | 11:08 |
| lucasagomes | dtantsur, on it right now | 11:09 |
| sambetts | https://review.openstack.org/#/q/project:openstack/ironic-lib+status:open | 11:09 |
| sambetts | for lazy people ^ | 11:09 |
| sambetts | :-P | 11:09 |
| dtantsur | thanks :) | 11:09 |
| *** tuanla_ has quit IRC | 11:10 | |
| mat128 | morning ironic | 11:12 |
| mat128 | brb vpn | 11:12 |
| *** jtomasek_ is now known as jtomasek | 11:13 | |
| wlfightup | good morning | 11:20 |
| *** _vsaienko has quit IRC | 11:23 | |
| *** thrash|g0ne is now known as thrash | 11:25 | |
| TheJulia | Good morning | 11:28 |
| *** _vsaienko has joined #openstack-ironic | 11:28 | |
| dtantsur | morning TheJulia, wlfightup | 11:28 |
| TheJulia | Regarding ironic-lib, I guess root device hints might not make it? | 11:29 |
| openstackgerrit | Julien proposed openstack/bifrost: Fix package installation failure https://review.openstack.org/359806 | 11:29 |
| *** mgoddard has joined #openstack-ironic | 11:31 | |
| *** mgoddard_ has quit IRC | 11:31 | |
| dtantsur | TheJulia, ideally they should... | 11:33 |
| dtantsur | but it depends on whether we are able to land them essentially today | 11:33 |
| *** _vsaienko has quit IRC | 11:36 | |
| dtantsur | TheJulia, your driver composition spec update is definitely on my radar (thanks!), I'm just a bit overwhelmed with urgent priorities both upstream and downstream | 11:37 |
| *** _vsaienko has joined #openstack-ironic | 11:37 | |
| openstackgerrit | Sam Betts proposed openstack/ironic-lib: Correct reraising of exception https://review.openstack.org/338797 | 11:38 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-lib: Add match_root_device_hints() to the utils.py module https://review.openstack.org/358000 | 11:38 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-lib: Extend parse_root_device_hints to support operators https://review.openstack.org/348953 | 11:38 |
| TheJulia | dtantsur: Okay, I'm in the same boat with tons of priorities right now too, just wanted to make sure it was in your review list :) | 11:38 |
| lucasagomes | vdrok, rloo ^ thanks a lot for the reviews! | 11:38 |
| lucasagomes | hope that address it | 11:38 |
| openstackgerrit | Shivanand Tendulker proposed openstack/ironic: Add 'inspect_ports' as a 'driver_info' parameter to iLO drivers https://review.openstack.org/325230 | 11:40 |
| *** ohamada_ has joined #openstack-ironic | 11:41 | |
| *** ohamada has quit IRC | 11:41 | |
| *** sean-k-mooney has quit IRC | 11:44 | |
| *** stendulker has quit IRC | 11:45 | |
| sambetts | lucasagomes: QQ, in https://review.openstack.org/#/c/348953/6/ironic_lib/utils.py you do re.split(...)[0::1] but [0::1] just gives you the same list back as you started with | 11:47 |
| lucasagomes | sambetts, oh | 11:48 |
| lucasagomes | sambetts, lemme check 1 sec | 11:48 |
| sambetts | lucasagomes: did you mean [1::1] | 11:48 |
| sambetts | ? | 11:48 |
| sambetts | to skip the operator? | 11:48 |
| sambetts | and even then you don't need the :1 part | 11:48 |
| *** ohamada_ has quit IRC | 11:48 | |
| *** ohamada_ has joined #openstack-ironic | 11:49 | |
| lucasagomes | sambetts, guess that can be removed then | 11:49 |
| lucasagomes | sambetts, yeah probably that it's a left over because I refactored the _extract_operators... method to extra both values and operator and be able to use it in more places | 11:50 |
| * lucasagomes fix it | 11:50 | |
| sambetts | :) | 11:50 |
| openstackgerrit | Dmitry Tantsur proposed openstack/python-ironicclient: [WIP] Add --wait to OSC provisioning commands https://review.openstack.org/354806 | 11:50 |
| *** jaosorior has quit IRC | 11:51 | |
| *** jaosorior has joined #openstack-ironic | 11:52 | |
| lucasagomes | sambetts, good catch, btw! Thanks | 11:54 |
| *** ohamada_ has quit IRC | 11:57 | |
| *** ohamada_ has joined #openstack-ironic | 11:57 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-lib: Add match_root_device_hints() to the utils.py module https://review.openstack.org/358000 | 12:00 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-lib: Extend parse_root_device_hints to support operators https://review.openstack.org/348953 | 12:00 |
| lucasagomes | sambetts, ^ I've added some tests to the normalize method separated that I forgot in the last review | 12:00 |
| *** annp has quit IRC | 12:00 | |
| * lucasagomes goes grab some food | 12:01 | |
| *** lucasagomes is now known as lucas-hungry | 12:01 | |
| sambetts | lucasagomes: another QQ, is it possible to have a hint with multiple different operators in, e.g. >= 5 < 10 ? because right now the extract_hint_operator_and_values function assumes 1 or more of the same operator | 12:03 |
| *** vmud213 has quit IRC | 12:04 | |
| *** zhangjl has left #openstack-ironic | 12:08 | |
| sambetts | man it is super hot in my office right now, I wish I had a fan | 12:09 |
| dtantsur | hot? in UK? | 12:10 |
| sambetts | dtantsur: I know its unnatural | 12:11 |
| dtantsur | well, they promise us +30C again :( but for now it's very nice: +25C and sunny | 12:11 |
| sambetts | dtantsur: we're at 27C today | 12:12 |
| *** dprince has joined #openstack-ironic | 12:12 | |
| *** ccamacho is now known as ccamacho|lunch | 12:12 | |
| dtantsur | yeah, this might be a problem without A/C | 12:13 |
| sambetts | my office is like a green house at the moment, I have 2 big windows behind me | 12:14 |
| dtantsur | ouch :( | 12:14 |
| *** bfournie has quit IRC | 12:15 | |
| openstackgerrit | Merged openstack/ironic-lib: Add developer documentation on metrics https://review.openstack.org/356796 | 12:15 |
| *** baoli has joined #openstack-ironic | 12:24 | |
| *** MattMan has quit IRC | 12:26 | |
| *** MattMan has joined #openstack-ironic | 12:26 | |
| *** bks has quit IRC | 12:38 | |
| *** wajdi has quit IRC | 12:42 | |
| *** rloo has joined #openstack-ironic | 12:44 | |
| *** bfournie has joined #openstack-ironic | 12:48 | |
| *** rbudden has joined #openstack-ironic | 12:54 | |
| *** lucas-hungry is now known as lucasagomes | 12:59 | |
| *** cdearborn has joined #openstack-ironic | 13:00 | |
| *** vishwanathj_zzz is now known as vishwanthj | 13:03 | |
| *** baoli_ has joined #openstack-ironic | 13:04 | |
| *** vikrant has quit IRC | 13:05 | |
| *** baoli has quit IRC | 13:06 | |
| sborkows | dtantsur: When you have a time, would you mind looking at my newest patch for PCI devices processing hook in ironic inspector? | 13:07 |
| dtantsur | sure, it's on my radar (not right this minute though) | 13:07 |
| *** dprince has quit IRC | 13:08 | |
| *** dprince has joined #openstack-ironic | 13:09 | |
| milan | sambetts, aarefiev hey guys :) could you please give a +1 to the Inspector states patch ;) https://review.openstack.org/#/c/348943/ | 13:15 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-lib: Add match_root_device_hints() to the utils.py module https://review.openstack.org/358000 | 13:15 |
| *** _vsaienko has quit IRC | 13:15 | |
| *** appu1034 has joined #openstack-ironic | 13:19 | |
| *** aparnav has quit IRC | 13:19 | |
| *** _vsaienko has joined #openstack-ironic | 13:24 | |
| sambetts | milan: are we going to apply the version_id replacing locking stuff as a separate patch? | 13:26 |
| *** openstackgerrit has quit IRC | 13:26 | |
| sambetts | milan: I would expect that to come before this one | 13:26 |
| *** openstackgerrit has joined #openstack-ironic | 13:27 | |
| *** causten has joined #openstack-ironic | 13:27 | |
| *** amotoki has quit IRC | 13:29 | |
| *** aparnav has joined #openstack-ironic | 13:31 | |
| *** appu1034 has quit IRC | 13:31 | |
| *** sdake has joined #openstack-ironic | 13:34 | |
| *** wajdi has joined #openstack-ironic | 13:35 | |
| *** aparnav has quit IRC | 13:36 | |
| *** sdake_ has joined #openstack-ironic | 13:37 | |
| *** thiagop has joined #openstack-ironic | 13:40 | |
| thiagop | good morning, Ironicers | 13:40 |
| *** sdake has quit IRC | 13:40 | |
| *** amotoki has joined #openstack-ironic | 13:44 | |
| *** _vsaienko has quit IRC | 13:44 | |
| *** sborkows has quit IRC | 13:46 | |
| *** _vsaienko has joined #openstack-ironic | 13:46 | |
| openstackgerrit | Imre Farkas proposed openstack/ironic: DRAC RAID configuration https://review.openstack.org/176319 | 13:48 |
| *** mgoddard_ has joined #openstack-ironic | 13:48 | |
| *** mgoddard has quit IRC | 13:52 | |
| milan | sambetts, what would be the benefit having that before the states rather than after? | 13:54 |
| sambetts | because states needs it, and we're adding work arounds in that patch that we don't need if we do the other thing first | 13:55 |
| milan | sambetts, right, not sure how fast we could land that, but I think we'd be able to land this one before feature freeze, wdyt? | 13:56 |
| sambetts | milan: well have the work is already in that patch though right? | 13:56 |
| sambetts | s/have/half | 13:56 |
| milan | sambetts, true, just I'm pessimistic about that we'd manage all the reviews etc.. this way we'd have states&the preparation as is more likely to get reviewers attention than a new one, isn't it? | 13:58 |
| sambetts | I think it'd only be 2 parts right? 1 patch adding version_id and removing existing locking, and the existing patch | 13:59 |
| milan | my point is if we manage to get it out of the doors in Newton we'd have some feedback before more HA bits land | 13:59 |
| sambetts | and most of the code we'll be putting in the new patch will be code that is already in the current patch | 14:00 |
| milan | I like the idea but there's just 1 week | 14:00 |
| milan | dtantsur, also wdyt ^ | 14:00 |
| *** sabeen has joined #openstack-ironic | 14:01 | |
| *** anush_ has joined #openstack-ironic | 14:01 | |
| milan | sambetts, but you're 100% right about the workaround @unlocks decorator :-/ | 14:01 |
| *** amotoki has quit IRC | 14:02 | |
| milan | sambetts, but the trick is the state is part of the locking mechanism with the @fsm_transition preventing actions in the wrong state, I need to think this through | 14:05 |
| *** mgoddard_ has quit IRC | 14:06 | |
| sambetts | milan: thats a good point :/ I was just thinking about that | 14:06 |
| *** mgoddard has joined #openstack-ironic | 14:06 | |
| milan | on the plus side, removing the locks patch will be quite simple one IMHO | 14:07 |
| sambetts | milan: without the states we have no way to determine what your allowed and not allowed to do to a node at the current time | 14:07 |
| *** rama_y has joined #openstack-ironic | 14:07 | |
| sambetts | milan: damn :/ | 14:07 |
| milan | sambetts, yeah, and we'd have to have some rudimentary version of the decorator checking the version_id anyway | 14:08 |
| sambetts | milan: yeah :/ | 14:08 |
| sambetts | I guess we can follow up with resolving our locking situation after this patch then | 14:09 |
| milan | sambetts, I'd like that | 14:10 |
| sambetts | we could still break the patch down (if we wanted for easier reviewing) but seeing as its only like me, dtantsur and aarefiev I think its ok like it is | 14:10 |
| * milan going to implement that right after the feature freeze | 14:10 | |
| rloo | hi and good morning everyone, milan, sambetts, lucasagomes, TheJulia, dtantsur | 14:11 |
| dtantsur | morning rloo | 14:11 |
| milan | morning rloo! :) | 14:11 |
| sambetts | milan: I see you resolved the state machine so finished -> error isn't a thing any more, thats much nicer | 14:11 |
| milan | sambetts, yeah | 14:11 |
| lucasagomes | rloo, good morning | 14:12 |
| rloo | dtantsur: as I look more into osc plugin, i keep thinking of other things we might want to change. sigh. eg 'passthru' -> 'passthrough'. am going to spend a bit of time thinking about it. | 14:12 |
| dtantsur | yeah, maybe | 14:13 |
| milan | sambetts, I'll be rebasing the OOL patch on this so I can remove the locking in there to track what it's going to take | 14:14 |
| milan | https://review.openstack.org/#/c/356982/ | 14:15 |
| milan | or I'll start a fresh one may | 14:15 |
| milan | maybe | 14:16 |
| sambetts | milan: that would be awesome :D btw I'm looking at the NodeInfo object right now, and I wonder if we should rename the self.finished() function to something that doesn't make it sound like we're moving into a finished state? | 14:16 |
| sambetts | milan: e.g. something like def set_finished_at(...) | 14:17 |
| milan | sambetts, it's a bit unfortunate w/r the naming | 14:17 |
| milan | yeah | 14:17 |
| milan | sambetts, but the change will grow, it's +980-75 :P | 14:20 |
| sambetts | yeah true :-P | 14:20 |
| *** amotoki has joined #openstack-ironic | 14:22 | |
| *** michauds has joined #openstack-ironic | 14:24 | |
| *** hjanssen-hpe has joined #openstack-ironic | 14:26 | |
| *** hpe-hj has joined #openstack-ironic | 14:26 | |
| sambetts | milan: I think I've spotted a point where the node might get stuck | 14:27 |
| sambetts | milan: and never move the Finished state | 14:27 |
| *** syed_ has joined #openstack-ironic | 14:27 | |
| sambetts | milan: https://review.openstack.org/#/c/348943/20/ironic_inspector/process.py if something goes wrong in the try in the _reapply function, before moving to _finish then we're stuck in reapplying I think | 14:28 |
| *** michauds_ has joined #openstack-ironic | 14:28 | |
| *** michauds has quit IRC | 14:28 | |
| * milan reads | 14:29 | |
| openstackgerrit | Markos Chandras proposed openstack/bifrost: Install the net-tools package in scripts/env-setup.sh https://review.openstack.org/359922 | 14:29 |
| *** rbrud has joined #openstack-ironic | 14:30 | |
| *** moshele has quit IRC | 14:30 | |
| sambetts | milan: I think we need to bubble the exception up the the decorator | 14:30 |
| *** ccamacho|lunch is now known as ccamacho | 14:30 | |
| milan | sambetts, seems you're right | 14:31 |
| *** michauds_ is now known as michauds__ | 14:31 | |
| *** mgoddard_ has joined #openstack-ironic | 14:31 | |
| *** michauds__ is now known as michauds | 14:32 | |
| milan | sambetts, so true: http://paste.openstack.org/show/563016/ | 14:34 |
| *** mgoddard has quit IRC | 14:34 | |
| * milan updated database to simulate it | 14:34 | |
| milan | fix should be to either make that reentrant or strictly non-reentrant | 14:35 |
| milan | sambetts, please -1 it, I shall fix that later, need to be running now | 14:35 |
| milan | sambetts, thanks a lot1 | 14:36 |
| *** amotoki has quit IRC | 14:36 | |
| * milan gone | 14:37 | |
| *** milan has quit IRC | 14:41 | |
| *** rajinir has joined #openstack-ironic | 14:48 | |
| *** amotoki has joined #openstack-ironic | 14:50 | |
| lucasagomes | rloo, re raising ParseException. I'm not sure it will actually raise it (looking at pyparsing code now) because we are converting all values to string before calling it | 14:51 |
| sambetts | lucasagomes: another QQ, is it possible to have a hint with multiple different operators in, e.g. >= 5 < 10 ? because right now the extract_hint_operator_and_values function assumes 1 or more of the same operator | 14:51 |
| lucasagomes | sambetts, it's not | 14:51 |
| lucasagomes | it's always one operator, we probably can improve the oslo utils module a lot in the future | 14:52 |
| lucasagomes | and make it a proper language, but atm it's quite limited | 14:52 |
| rloo | lucasagomes: then it is fine. I didn't know (and we can always address that later) | 14:52 |
| *** trandles has joined #openstack-ironic | 14:52 | |
| *** _vsaienko has quit IRC | 14:52 | |
| sambetts | I can't do >=5 <and> < 10 or something? | 14:52 |
| lucasagomes | sambetts, AFAIK nop, lemme check it | 14:52 |
| sambetts | it would be usful to be able to do something like size: >500GB <and> <2TB | 14:53 |
| lucasagomes | sambetts, yeah it does not | 14:54 |
| lucasagomes | http://paste.openstack.org/show/563083/ | 14:54 |
| lucasagomes | sambetts, indeed... there's a lot of things to improve on that syntax | 14:55 |
| lucasagomes | I would like to have size units as well like 20M 10GiB etc | 14:55 |
| sambetts | ++! | 14:55 |
| lucasagomes | currently we only support GiB for size | 14:55 |
| sambetts | lucasagomes: I guess thats because nova only supports GiB for size | 14:56 |
| sambetts | lucasagomes: :/ | 14:56 |
| lucasagomes | yeah, tho in Ironic we could support any size unit | 14:58 |
| sambetts | yup | 14:58 |
| *** jistr is now known as jistr|mtg | 15:00 | |
| *** appu103_ has joined #openstack-ironic | 15:02 | |
| rloo | lucasagomes: ping me when there's another revision to review | 15:02 |
| rloo | dtantsur: wrt 'ironic port-list --node'. the reason for it is because i'd like to have 'openstack baremetal port list --node' and NOT have 'openstack baremetal node port list' | 15:07 |
| dtantsur | rloo, well, but we already have node-port-list. do you want to deprecate it? | 15:07 |
| rloo | dtantsur: https://review.openstack.org/#/c/359520/ & the RFE: https://bugs.launchpad.net/python-ironicclient/+bug/1616242 | 15:08 |
| openstack | Launchpad bug 1616242 in python-ironicclient "[RFE] port list should have --node argument" [Wishlist,In progress] - Assigned to Ruby Loo (rloo) | 15:08 |
| rloo | dtantsur: I am hoping that at some future date, we won't support ironic CLI any more. I'm more concerned now with getting the OSC stuff 'right'. | 15:08 |
| dtantsur | fair | 15:08 |
| rloo | dtantsur: i don't think we want to provide more than one way in osc, to do something. | 15:09 |
| *** sabeen has quit IRC | 15:09 | |
| rloo | dtantsur: so if we want to support 'openstack baremetal node port list', we could add that later, but I think it does make sense to have a 'openstack baremetal port list --node' regardless, cuz --address is available | 15:10 |
| dtantsur | true | 15:11 |
| dtantsur | I'm just worried about 2 similar commands in the old CLI | 15:11 |
| rloo | dtantsur: OH. | 15:12 |
| rloo | dtantsur: I could just do it for osc, but there would be two ways in the API. We probably shouldn't have provided those two REST APIs. | 15:12 |
| dtantsur | yeah... | 15:13 |
| sambetts | *cough* working on it *cough* | 15:13 |
| rloo | sambetts: :) | 15:14 |
| lucasagomes | rloo, will do, in a meeting atm but I will put a new patch-set right after it | 15:15 |
| sambetts | rloo: these are a step in the right direction for the APIs https://review.openstack.org/#/q/topic:bug/1580997 because there are actually about a number of different ways in the API to node port list, some of which were never intended | 15:15 |
| rloo | sambetts: yeah, gotta clean that up. So much to do... and I want to get as much osc stuff into next week's client release. | 15:16 |
| openstackgerrit | Michael Turek proposed openstack/ironic: Add classes for Portgroups API enablement https://review.openstack.org/347549 | 15:16 |
| sambetts | totally :D /me is longing for not having to type set-provision-state any more | 15:16 |
| rloo | sambetts: ++, especially since i can never remember the actual verb we are using! | 15:17 |
| openstackgerrit | Mathieu Mitchell proposed openstack/ironic-python-agent: Enforce upper-constraints when building ramdisks https://review.openstack.org/358855 | 15:23 |
| openstackgerrit | Mathieu Mitchell proposed openstack/ironic-python-agent: Enforce upper-constraints when building ramdisks https://review.openstack.org/358855 | 15:23 |
| *** david-lyle_ has joined #openstack-ironic | 15:25 | |
| *** sdake_ has quit IRC | 15:25 | |
| *** david-lyle_ has quit IRC | 15:26 | |
| *** rybridges has joined #openstack-ironic | 15:27 | |
| *** rbartal has quit IRC | 15:29 | |
| *** _vsaienko has joined #openstack-ironic | 15:30 | |
| *** fragatina has joined #openstack-ironic | 15:31 | |
| *** fragatin_ has joined #openstack-ironic | 15:32 | |
| *** mgoddard_ has quit IRC | 15:34 | |
| *** mgoddard has joined #openstack-ironic | 15:34 | |
| *** rpioso has joined #openstack-ironic | 15:35 | |
| *** fragatina has quit IRC | 15:35 | |
| *** krtaylor has quit IRC | 15:35 | |
| *** hshiina has joined #openstack-ironic | 15:36 | |
| *** sdake has joined #openstack-ironic | 15:36 | |
| openstackgerrit | Merged openstack/ironic-lib: Correct reraising of exception https://review.openstack.org/338797 | 15:37 |
| dtantsur | folks, where are we with ironic-lib changes? | 15:38 |
| *** sukhdev_ has joined #openstack-ironic | 15:39 | |
| sambetts | dtantsur: two have merged | 15:39 |
| sambetts | https://review.openstack.org/#/q/project:openstack/ironic-lib+status:open | 15:39 |
| sambetts | dtantsur: its just lucasagomes's matcher stuff now | 15:39 |
| dtantsur | so, we're waiting for lucasagomes to update the other two, right? | 15:39 |
| *** ibmchas has quit IRC | 15:40 | |
| *** _vsaienko has quit IRC | 15:40 | |
| *** fxpester has quit IRC | 15:41 | |
| rloo | dtantsur: yup. is it OK to do the release tomorrow? | 15:41 |
| *** _vsaienko has joined #openstack-ironic | 15:42 | |
| rloo | dtantsur: i was thinking that ironic-lib is only used w/i ironic. i wonder if it really needs to follow the release schedule. but maybe it makes life easier for others. | 15:43 |
| dtantsur | rloo, all g-r bumps will be frozen soon, this is why.. but yes, we can do it tomorrow | 15:43 |
| openstackgerrit | Jay Faulkner proposed openstack/ironic: Add metrics to administrator guide https://review.openstack.org/356811 | 15:43 |
| rloo | dtantsur: ok thx. | 15:44 |
| *** _vsaienko has quit IRC | 15:45 | |
| lucasagomes | rloo, great suggestions btw | 15:51 |
| lucasagomes | rloo, the parse inside match() | 15:51 |
| lucasagomes | also, do you think it would be better to get a list of dictionaries instead of device objects ? To make it more generic | 15:51 |
| dtantsur | not sure what you're talking about, but if inspector will use root device hints, devices will be dicts | 15:52 |
| openstackgerrit | Michael Turek proposed openstack/ironic: Prevent URL collisions with sub-controllers: nodes/ports https://review.openstack.org/314514 | 15:53 |
| openstackgerrit | Michael Turek proposed openstack/ironic: Add classes for Portgroups API enablement https://review.openstack.org/347549 | 15:53 |
| *** rama_y has quit IRC | 15:53 | |
| *** penick has joined #openstack-ironic | 15:53 | |
| *** Sukhdev has joined #openstack-ironic | 15:54 | |
| *** vishwanthj has quit IRC | 15:56 | |
| lucasagomes | dtantsur, ++ I will make it into dicts | 15:56 |
| lucasagomes | and we can have a as_dict() method in IPA's BlockDevice class | 15:56 |
| rloo | lucasagomes: that's good with me | 15:57 |
| *** penick has quit IRC | 15:58 | |
| JayF | lucasagomes: I think it already exists, because it's serializable | 15:58 |
| * JayF looks | 15:59 | |
| lucasagomes | there we go :D | 15:59 |
| JayF | https://github.com/openstack/ironic-python-agent/blob/master/ironic_python_agent/encoding.py#L19 | 15:59 |
| openstackgerrit | Thiago Paiva Brito proposed openstack/ironic: OneView test nodes to use dynamic allocation https://review.openstack.org/359990 | 15:59 |
| openstackgerrit | Julia Kreger proposed openstack/ironic-python-agent: Fix IPA for stable/mitaka with noauth mode https://review.openstack.org/359991 | 16:00 |
| JayF | any of the HardwareType classes in IPA take a blah.serializable() | 16:00 |
| *** mgoddard_ has joined #openstack-ironic | 16:00 | |
| thiagop | folks, 359990 is a small add to the devstack plugin to allow oneview nodes to be enrolled in the new allocation model. Effectively, it is a 6-line patch, should be easy to review | 16:02 |
| *** krtaylor has joined #openstack-ironic | 16:02 | |
| openstackgerrit | Julia Kreger proposed openstack/ironic-python-agent: Fix IPA for stable/mitaka with noauth mode https://review.openstack.org/359991 | 16:03 |
| *** mgoddard has quit IRC | 16:04 | |
| *** ifarkas is now known as ifarkas_afk | 16:08 | |
| *** jistr|mtg is now known as jistr | 16:08 | |
| *** hshiina has quit IRC | 16:09 | |
| *** athomas has quit IRC | 16:21 | |
| *** fragatin_ has quit IRC | 16:21 | |
| openstackgerrit | Michael Turek proposed openstack/ironic: Add portgroups to support LAG interfaces - API https://review.openstack.org/332177 | 16:24 |
| *** hjanssen-hpe has quit IRC | 16:27 | |
| NobodyCam | Good Morning Ironic'ers | 16:28 |
| *** hpe-hj has quit IRC | 16:28 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic: Fix issues in dev-quickstart and index https://review.openstack.org/360003 | 16:28 |
| JayF | rloo: ^ that should fix your post-review comments for the doc update | 16:28 |
| *** jpich has quit IRC | 16:29 | |
| rloo | thx JayF. Although no time to review; hopefully others will. I want to focus on ironic-lib & client stuff for the next week or so. (I shouldn't have looked at the doc and api-ref changes.) | 16:29 |
| *** rcernin has quit IRC | 16:29 | |
| NobodyCam | JayF: +2 from me :) | 16:30 |
| JayF | rloo: that's fine; I just wanted you to know I did take action, the comments/post merge review wasn't in vain :) | 16:31 |
| JayF | rloo: I pushed for that to get merged yesterday so I wanted to make sure any issues you still had were addressed | 16:31 |
| dtantsur | morning NobodyCam | 16:31 |
| NobodyCam | Morning dtantsur :) | 16:31 |
| rloo | JayF: thx. which reminds me, I should check something... | 16:31 |
| NobodyCam | morning rloo and JayF too :) | 16:32 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-lib: Add match_root_device_hints() to the utils.py module https://review.openstack.org/358000 | 16:32 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-lib: Extend parse_root_device_hints to support operators https://review.openstack.org/348953 | 16:32 |
| lucasagomes | rloo, ^ | 16:32 |
| rybridges | Hello, I was in here last night asking about '_check_deploying_status' in the https://github.com/openstack/ironic repository, but I closed my irc client and went to bed before i saw a response last night. | 16:32 |
| lucasagomes | hope I've addressed everything | 16:32 |
| NobodyCam | morning lucasagomes | 16:32 |
| lucasagomes | NobodyCam, good morning | 16:32 |
| NobodyCam | :) | 16:32 |
| rloo | JayF: yeah, that's what i was wondering. The console doc is missing :-( | 16:32 |
| rloo | JayF: I think that's why you had to rebase yesterday | 16:33 |
| lucasagomes | rloo, urgh forgot a docstring | 16:33 |
| JayF | rloo: uh oh, like missing-missing or just missing a link? | 16:33 |
| rloo | JayF: do you have time to fix. I think we need to add a pointer/link from http://docs.openstack.org/developer/ironic/#configuration | 16:33 |
| JayF | rloo: absolutely I can fix, will find the link and add it | 16:34 |
| rloo | JayF: I didn't look, but I hope the actual page is there :) | 16:34 |
| rybridges | Basically, I think that 'iter_nodes' in the '_check_deploying_status' method is only returning nodes that belong to the current conductor. However this defeats the point of the method because the '_check_deploying_status' is supposed to look at nodes that belong to other (offline) conductors | 16:34 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-lib: Add match_root_device_hints() to the utils.py module https://review.openstack.org/358000 | 16:34 |
| rloo | rybridges: each conductor has their own periodic tasks; each conductor will check the status for the nodes they are responsible for | 16:34 |
| rybridges | however, since iter_nodes is only returning nodes that belong to the current conductor, that method will never see nodes that belong to offline conductors. | 16:35 |
| *** tesseract- has quit IRC | 16:35 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic: Fix issues in dev-quickstart and index https://review.openstack.org/360003 | 16:35 |
| rloo | rybridges: if a node belongs to an offline conductor, at some point it will get reassigned to an online conductor | 16:35 |
| rybridges | @rloo ohhh | 16:36 |
| rloo | lucasagomes: is https://review.openstack.org/#/c/348953/ ready to be reviewed? | 16:37 |
| lucasagomes | rloo, yeah hope I've addressed everything | 16:38 |
| rloo | lucasagomes: ok, gonna look now :) | 16:38 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-lib: Add match_root_device_hints() to the utils.py module https://review.openstack.org/358000 | 16:38 |
| lucasagomes | rloo, thanks, and sorry for the annoyance :D | 16:39 |
| lucasagomes | brute forcing for getting it in | 16:39 |
| rloo | lucasagomes: what annoyance? don't worry, dtantsur said he could do the release tomorrow. | 16:39 |
| lucasagomes | right on | 16:39 |
| rybridges | rloo: so if a conductor goes down while a node is in the deploying state, then the node gets reassigned to a new conductor, will the new conductor at that point try to continue deploying that node or will it try to put it into the fail state? | 16:40 |
| rloo | rybridges: it should work (new conductor should continue with the deploy), but I haven't tested it myself. | 16:40 |
| lucasagomes | rybridges, the new conductor will fail the deployment AFAIK | 16:41 |
| dtantsur | yeah, I can request it tomorrow morning, if we land all the stuff by then | 16:41 |
| rloo | lucasagomes: oh? | 16:41 |
| lucasagomes | rloo, cause when it's in deploying means the previous conductor was working on it (!= deploy wait) | 16:41 |
| dtantsur | I think DEPLOYING is not recoverable, DEPLOY WAIT is | 16:41 |
| lucasagomes | yeah | 16:41 |
| lucasagomes | the new conductor won't know which stage the node was when it was DEPLOYING | 16:41 |
| rloo | lucasagomes: right, if the conductor that goes offline is doing something, then all bets are off | 16:42 |
| lucasagomes | rloo, yeah, unless the node was in deploy wait. Then the lookup/heartbeat method will just redirect the requests to a new conductor and the deployment will continue | 16:42 |
| lucasagomes | I mean, the API will redirect it | 16:42 |
| * lucasagomes looks at the code | 16:43 | |
| *** _vsaienko has joined #openstack-ironic | 16:43 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic: Fix issues in dev-quickstart and index https://review.openstack.org/360003 | 16:44 |
| rybridges | okay, so if we are assuming that nodes who are associated with offline conductors are automatically reassigned, then what is the point of _check_deploying_status? Theoretically, if nodes are always re-assigned to online conductors, there should never be a scenario in which a node is sitting in the DEPLOY state but still assigned to an offline conductor | 16:44 |
| openstackgerrit | Jay Faulkner proposed openstack/ironic: Add metrics to administrator guide https://review.openstack.org/356811 | 16:45 |
| *** derekh has quit IRC | 16:45 | |
| JayF | NobodyCam: ^ fixed the issue rloo pointed out in 360003, also rebased my metrics doc patch on that, would love a re-review (and a first-review of the metrics patch if you don't mind) | 16:45 |
| openstackgerrit | Merged openstack/ironic: Add notification base classes and docs https://review.openstack.org/298461 | 16:45 |
| rloo | lucasagomes: so the operators are always separated from the values by a space? | 16:45 |
| rybridges | The description of the method says the following "Periodically checks the status of nodes in DEPLOYING state. Periodically checks the nodes in DEPLOYING and the state of the conductor deploying them. If we find out that a conductor that was provisioning the node has died we then break release the node and gracefully mark the deployment as failed." | 16:46 |
| NobodyCam | :) | 16:46 |
| lucasagomes | rloo, it doesn't really matter it can be in both way | 16:46 |
| lucasagomes | rloo, spaces just makes it looks better | 16:46 |
| lucasagomes | e.g | 16:46 |
| rloo | lucasagomes: ok, i was just wondering why you changed it :) | 16:46 |
| *** jaosorior has quit IRC | 16:47 | |
| lucasagomes | rloo, http://paste.openstack.org/show/563101/ | 16:47 |
| rloo | thx lucasagomes | 16:47 |
| lucasagomes | rybridges, so it's a bit complicated. Because there's also a lock mechanism which makes sure that if the ring is rebalanced the new conductor won't do anything with the node while there's another conductor using it | 16:48 |
| lucasagomes | rybridges, that's the "reserved" field in the node, the value of it is the hostname of the conductor that is actively using it atm | 16:48 |
| *** Sukhdev has quit IRC | 16:48 | |
| *** sukhdev_ is now known as sukhdev | 16:48 | |
| lucasagomes | rybridges, also, currently this lock is a DB lock. So it is still possible I believe to have a node stuck with it | 16:49 |
| *** mtanino has joined #openstack-ironic | 16:50 | |
| lucasagomes | rybridges, we plan to use something like tooz (it's a library abstracting locking mechanisms e.g zookeeper) to handle it for us | 16:50 |
| *** appu103_ has quit IRC | 16:50 | |
| rybridges | lucasagomes: thanks for the response! :) so if i understand correctly, when the a conductor goes offline, the node could still be stuck with that lock in the db? | 16:51 |
| *** ohamada_ has quit IRC | 16:52 | |
| *** vishwanathj has joined #openstack-ironic | 16:52 | |
| lucasagomes | rybridges, yes | 16:52 |
| lucasagomes | there are some edge cases yet | 16:52 |
| lucasagomes | tho many have been improved from few releases ago | 16:53 |
| *** rama_y has joined #openstack-ironic | 16:54 | |
| dtantsur | g'night, see you tomorrow | 16:55 |
| *** dtantsur is now known as dtantsur|afk | 16:55 | |
| NobodyCam | night dtantsur|afk | 16:55 |
| *** aNupoisc has joined #openstack-ironic | 16:56 | |
| *** phschwartz has quit IRC | 16:56 | |
| *** hj-hpe has joined #openstack-ironic | 16:57 | |
| *** mgoddard_ has quit IRC | 16:57 | |
| *** _vsaienko has quit IRC | 16:57 | |
| *** mgoddard has joined #openstack-ironic | 16:57 | |
| *** phschwartz has joined #openstack-ironic | 16:58 | |
| rybridges | lucasagomes: okay i see. this still seems a off though. _check_deploying_status first gets a list of the offline conductors. then iter_nodes returns a list of nodes that belong to the current conductor. then we check if any of those nodes that were returned by iter_nodes belong to an offline conductor. How could a node returned by iter_nodes in this scenario ever belong to an offline conductor though since they always get | 16:58 |
| rybridges | lucasagomes: nodes associated with the current conductor (which is online) | 16:59 |
| lucasagomes | rybridges, sorry I've jumped in the middle of the conversation heh... lemme look at the code | 17:01 |
| rybridges | lucasagomes: oh no worries. thanks so much for taking the time lucasagomes i really appreciate it | 17:02 |
| openstackgerrit | Mathieu Mitchell proposed openstack/ironic-python-agent: Enforce upper-constraints when building ramdisks https://review.openstack.org/358855 | 17:02 |
| lucasagomes | rybridges, I think there's two things: 1. the API will route requests for that node to a specific conductor, this is done by the hash_ring.py module | 17:05 |
| lucasagomes | 2. a node being reserved by a conductor | 17:05 |
| lucasagomes | I think what happens there is, a conductor was deploying a node and suddenly died (leaving it in DEPLOYING state) | 17:06 |
| rybridges | right | 17:06 |
| rybridges | thats the idea | 17:06 |
| lucasagomes | the hash_ring will identify it because the conductor does not heartbeat anymore | 17:06 |
| lucasagomes | so it will rebalance the ring and the new conductor now will "own" that node | 17:06 |
| rybridges | okay. | 17:06 |
| lucasagomes | but the reservation field still points to the previous conductor that was deploying the node | 17:07 |
| rybridges | right! | 17:07 |
| rybridges | okay i was just thinking that | 17:07 |
| lucasagomes | and that periodic task clean that reservation field | 17:07 |
| rybridges | okay that makes much more sense now | 17:07 |
| rybridges | the key to this whole thing was understanding how that reservation field works | 17:07 |
| lucasagomes | and mark the deploy as failed because it can't continue (it doesn't know where the previous conductor was in the deployment phase) | 17:08 |
| lucasagomes | rybridges, yeah | 17:08 |
| rybridges | so, the fact that the reservation field is still set to some other (offline) node will not prevent this conductor from changing the state, right? | 17:08 |
| lucasagomes | rybridges, and the reason to have a reservation field is because, some actions (e.g updating the firwmare) can't be interrupted | 17:08 |
| rybridges | err, i meant, some other (offline) conductor | 17:09 |
| rybridges | right, that makes sense | 17:09 |
| lucasagomes | if we didn't have it, another conductor would own that node in case of hash ring rebalance and power off it | 17:09 |
| lucasagomes | while it was updating the firmware, and that could brick the node | 17:09 |
| lucasagomes | so we need to be careful | 17:09 |
| rybridges | yes that makes sense | 17:09 |
| lucasagomes | rybridges, it will prevent changes yes | 17:09 |
| lucasagomes | rybridges, to change the state of a conductor, the new conductor needs to acquire the lock | 17:10 |
| rybridges | i see, thats why we call objects.Node.release | 17:10 |
| lucasagomes | and the code will prevent a new conductor from doing it while the reservation field is set | 17:10 |
| rybridges | to try and release that lock | 17:10 |
| lucasagomes | rybridges, yes | 17:10 |
| lucasagomes | exactly yes | 17:10 |
| rybridges | ohhkay | 17:11 |
| rybridges | you have convinced me that this is correct. so one more question | 17:11 |
| rybridges | it was my understanding that we were also supposed to have similar logic for nodes that get stuck in the 'CLEANING` state when their conductor goes down | 17:11 |
| rybridges | but i dont see that anywhere.. did i miss it? | 17:11 |
| lucasagomes | rybridges, idk off the top of my head, but it wouldn't surprise me | 17:12 |
| rybridges | so nodes that are in CLEANING when their conductor goes down will get set to CLEAN FAIL | 17:12 |
| rybridges | okay | 17:12 |
| lucasagomes | and yes, we should have such mechanisms to all *ING states | 17:12 |
| rybridges | right | 17:12 |
| *** jcoufal has joined #openstack-ironic | 17:12 | |
| rybridges | well i dont see it in the upstream repo, so I may start work on it this week. | 17:12 |
| *** yogi has quit IRC | 17:12 | |
| lucasagomes | rybridges, https://bugs.launchpad.net/ironic/+bug/1461937 | 17:13 |
| openstack | Launchpad bug 1461937 in Ironic "Node stuck on DEPLOYING (potentially all *ING) state(s)" [High,Fix released] - Assigned to Lucas Alvares Gomes (lucasagomes) | 17:13 |
| rybridges | this all makes a lot more sense now. thanks so much for the help lucasagomes | 17:13 |
| lucasagomes | I think it fits in the potentially all ING states | 17:13 |
| lucasagomes | so yeah if CLEANING is not there, we should have something | 17:13 |
| lucasagomes | rybridges, no worries. Thank YOU for looking into fixing it | 17:13 |
| rybridges | yea I actually work at Yahoo and was working on creating a patch for us for that exact ticket so we could pull that fix for deploying states into our codebase :) | 17:14 |
| rybridges | anyways, thanks again. I'm sure we will meet again soon | 17:15 |
| lucasagomes | rybridges, cool | 17:15 |
| lucasagomes | definitely | 17:15 |
| *** sabeen has joined #openstack-ironic | 17:18 | |
| openstackgerrit | Julia Kreger proposed openstack/ironic-python-agent: Re-use API client for Heartbeat operations https://review.openstack.org/360018 | 17:21 |
| lucasagomes | folks, I'm calling it a day | 17:22 |
| lucasagomes | have a great evening all. Talk to you all tomorrow | 17:22 |
| TheJulia | goodnight lucasagomes | 17:23 |
| *** rcernin has joined #openstack-ironic | 17:23 | |
| lucasagomes | TheJulia, g'night! Thax again for that fix ^ | 17:23 |
| lucasagomes | added to my list to review tomorrow | 17:23 |
| TheJulia | No problem, just wish I could have figured out a unit test to raise it easily :) | 17:24 |
| * TheJulia may have spent a good chunk of the morning on trying to do that | 17:24 | |
| *** electrofelix has quit IRC | 17:29 | |
| NobodyCam | night lucasagomes | 17:30 |
| *** lucasagomes is now known as lucas-dinner | 17:35 | |
| *** sukhdev has quit IRC | 17:44 | |
| *** aswadr_ has quit IRC | 17:51 | |
| *** jhooker has quit IRC | 17:52 | |
| *** jhooker has joined #openstack-ironic | 17:52 | |
| *** rybridges has quit IRC | 17:53 | |
| *** sabeen has quit IRC | 17:59 | |
| *** sabeen has joined #openstack-ironic | 18:00 | |
| * sambetts -> home | 18:13 | |
| *** sambetts is now known as sambetts|afk | 18:13 | |
| sambetts|afk | night all! | 18:13 |
| *** amotoki has quit IRC | 18:17 | |
| *** dims has quit IRC | 18:18 | |
| *** dims has joined #openstack-ironic | 18:18 | |
| *** Lord_Phoenix has quit IRC | 18:24 | |
| *** vishwanathj has quit IRC | 18:25 | |
| *** vishwanathj has joined #openstack-ironic | 18:26 | |
| *** Sukhdev has joined #openstack-ironic | 18:28 | |
| *** intr1nsic has quit IRC | 18:28 | |
| *** intr1nsic has joined #openstack-ironic | 18:29 | |
| NobodyCam | night sambetts|afk | 18:42 |
| *** sabeen has quit IRC | 18:44 | |
| *** fragatina has joined #openstack-ironic | 18:46 | |
| *** thiagop has quit IRC | 18:55 | |
| *** mjturek1 has quit IRC | 18:55 | |
| *** jcoufal has quit IRC | 18:59 | |
| *** causten has quit IRC | 19:00 | |
| *** thiagop has joined #openstack-ironic | 19:02 | |
| *** adnavare has joined #openstack-ironic | 19:03 | |
| *** aNupoisc has left #openstack-ironic | 19:03 | |
| *** mjturek has joined #openstack-ironic | 19:04 | |
| *** Goneri has joined #openstack-ironic | 19:05 | |
| *** ccamacho has quit IRC | 19:07 | |
| *** _vsaienko has joined #openstack-ironic | 19:15 | |
| JayF | I'm reviewing smoe client patches for the first time, really ever honestly | 19:17 |
| JayF | and I'm finding that if you do any command that doesn't exist, the error is kinda terrible | 19:17 |
| JayF | openstack: 'baremetal' is not an openstack command. See 'openstack --help'. | 19:17 |
| JayF | baremetal does exist, just the 3 item down the command line was screwed up | 19:17 |
| openstackgerrit | Merged openstack/ironic: Updated from global requirements https://review.openstack.org/359457 | 19:19 |
| *** sdake has quit IRC | 19:20 | |
| rloo | JayF: welcome to osc-land :-( | 19:21 |
| rloo | JayF: those errors are from the openstackclient. | 19:21 |
| JayF | I tried to test it against my preproduction actual ironic | 19:21 |
| JayF | and it won't auth properly for that either | 19:21 |
| JayF | so I gotta nail that down some other time, until then, running against devstack | 19:22 |
| rloo | JayF: I'm testing against a devstack setup | 19:22 |
| rloo | JayF: no time to play around with it | 19:22 |
| JayF | yeah; I'm just saying OSC is different there too :( | 19:22 |
| JayF | I think jroll figured it out at some point, for our downstream ironic, I'll bug him | 19:22 |
| JayF | but I am reviewing all these OSC patches | 19:22 |
| JayF | bug/1526479 topic | 19:23 |
| rloo | JayF: the coding isn't the difficult part for me, i'm rethinking some of the osc commands we came up with :-( | 19:23 |
| JayF | I really dislike that baremetal list ~= baremetal node list | 19:24 |
| *** rama_y has quit IRC | 19:26 | |
| mat128 | I felt like it should have been "openstack node list" | 19:27 |
| mat128 | but was turned down when I mentioned it :P | 19:27 |
| JayF | that works too | 19:27 |
| JayF | well openstack baremetal node list | 19:27 |
| JayF | but openstack baremetal list does too | 19:27 |
| mat128 | do we have openstack instance list | 19:27 |
| rloo | JayF: sorry, in a conversation right now | 19:27 |
| mat128 | "openstack server list" aka nova list | 19:28 |
| JayF | yeah; i don't know | 19:28 |
| JayF | to be blunt I kinda hate the idea of a unified client in general | 19:28 |
| JayF | but once you concede that; I want all ironic things namespaced under openstack baremetal; which is what seems to be the case | 19:29 |
| mat128 | hmm | 19:29 |
| mat128 | so where do you put chassis? :) | 19:29 |
| mat128 | or ports | 19:29 |
| JayF | right now openstack baremetal chassis [blah] (at least in proposed patches | 19:30 |
| mat128 | openstack baremetal ports list | 19:30 |
| mat128 | hmm | 19:30 |
| JayF | openstack baremetal node [blah] works too | 19:30 |
| JayF | but openstack baremetal list implies node | 19:30 |
| JayF | so *shrug* | 19:30 |
| mat128 | baremetal = node | 19:30 |
| *** Goneri has quit IRC | 19:30 | |
| mat128 | by that logic, nothing else should go under baremetal | 19:30 |
| JayF | well that's why I said I disliked that `openstack baremetal list` does something and it's not always required to be `openstack baremetal node list` | 19:30 |
| mat128 | because you see baremetal as a group, not as being equal to node | 19:31 |
| mat128 | we have to be consistent otherwise I'd rather keep using python-ironicclient :) | 19:31 |
| mat128 | btw, were you saying all of OSC for ironic doesnt work with your setup, or only code you're reviewing? | 19:31 |
| JayF | I mean, that's kinda why I dislike the idea of a unified client | 19:31 |
| JayF | all of OSC doesn't work *using our existing automation for setting vars to auth* | 19:32 |
| JayF | I suspect it's all downstream config bs + rackspace auth not being real keystone | 19:32 |
| mat128 | ah | 19:32 |
| *** _vsaienko has quit IRC | 19:33 | |
| mat128 | OSC works here, but the output is terrible when your term isn't wide enough | 19:33 |
| mat128 | old client just dumped it, and I could resize after issuing a command (given you have a good term) | 19:33 |
| JayF | does osc have a --json equivalent like ironic cli does? | 19:33 |
| mat128 | osc: https://www.dropbox.com/s/8ouw584spt57e1q/Screen%20Shot%202016-08-24%20at%203.33.49%20PM.png?dl=0 | 19:34 |
| mat128 | ironicclient: https://www.dropbox.com/s/wll33ughpbjoplx/Screen%20Shot%202016-08-24%20at%203.34.00%20PM.png?dl=0 | 19:34 |
| *** jvgrant has quit IRC | 19:35 | |
| JayF | egad | 19:35 |
| *** e0ne has joined #openstack-ironic | 19:35 | |
| *** jhooker has quit IRC | 19:37 | |
| mat128 | btw, yes it has a -f json option | 19:37 |
| *** jhooker has joined #openstack-ironic | 19:37 | |
| JayF | then I'd use that + jq for everything | 19:37 |
| JayF | json is a hell of a lot easier than the console table output anyway | 19:37 |
| mat128 | TIL jq | 19:38 |
| mat128 | wow | 19:38 |
| JayF | seriously? | 19:38 |
| JayF | awk = old busted | 19:38 |
| JayF | jq = new hotness | 19:38 |
| mat128 | I've been using python -m json.tool | 19:38 |
| JayF | at least if you're using modern apis that output json :) | 19:38 |
| mat128 | and inline python -c 'import sys; open(sys.stdin) stuff | 19:38 |
| JayF | not saying that's completely replaced --json > output.json and then doing stuff with it | 19:38 |
| mat128 | so thats like xpath for json | 19:39 |
| JayF | xpath? | 19:39 |
| mat128 | xml xpath | 19:39 |
| *** _vsaienko has joined #openstack-ironic | 19:39 | |
| JayF | oh; I don't use a lot of xml apis ;) | 19:39 |
| mat128 | https://en.wikipedia.org/wiki/XPath#Abbreviated_syntax | 19:39 |
| mat128 | query language for xml trees | 19:39 |
| JayF | docs jobs for ironic-lib just merged in project-config \o/ | 19:40 |
| *** Goneri has joined #openstack-ironic | 19:43 | |
| *** rama_y has joined #openstack-ironic | 19:43 | |
| JayF | either way, all this bikeshedding about UX | 19:43 |
| JayF | I'm going to just review these for consistency with the spec | 19:43 |
| JayF | and make sure they work | 19:43 |
| rloo | JayF, mat128: sorry, just got back. wrt osc, did you see our spec? it provides a picture of the way the commands are going | 19:43 |
| *** _vsaienko has quit IRC | 19:44 | |
| rloo | JayF, mat128: http://specs.openstack.org/openstack/ironic-specs/specs/not-implemented/ironicclient-osc-plugin.html | 19:44 |
| mat128 | sorry, was playing with my new tool, jq | 19:44 |
| *** _vsaienko has joined #openstack-ironic | 19:44 | |
| mat128 | openstack baremetal list -f json | jq '.[].UUID' | 19:44 |
| mat128 | :D | 19:44 |
| mat128 | rloo: checking | 19:44 |
| rloo | JayF, mat128: and I'm in the middle of changing some of that. | 19:44 |
| rloo | mat128: don't use openstack baremetal list. it has been deprecated. | 19:44 |
| JayF | rloo: ah, I see that we're obsoleting ^ yeah | 19:44 |
| mat128 | rloo: thats all my help has | 19:44 |
| rloo | mat128: openstack baremetal node list | 19:44 |
| mat128 | ooh | 19:45 |
| mat128 | I know someone who is happy now | 19:45 |
| JayF | mat128: review and download these patches implementnig this spec :) | 19:45 |
| rloo | mat128: oh, older client then. | 19:45 |
| JayF | https://review.openstack.org/#/q/status:open+project:openstack/python-ironicclient+branch:master+topic:bug/1526479 | 19:45 |
| JayF | mat128: ^ go therefore and review | 19:45 |
| *** e0ne has quit IRC | 19:45 | |
| mat128 | on my way | 19:45 |
| rloo | mat128: we just made changes in newton. the problem with the 'openstack baremetal list' is that it was implemented/merged before we had a chance (as a group) to decide on the commands. | 19:45 |
| mat128 | do we have to be backwards compatible? I remember deva saying the client isnt a bash SDK therefore we should be backwards compatible | 19:46 |
| mat128 | I kind of disagree, but OSC is so new that I dont think much will be impacted | 19:46 |
| JayF | I mean, if you toss out what we "can" do | 19:46 |
| JayF | deprecating it slowly is probably the right thing to do | 19:47 |
| JayF | so I worry less about if we can just pull it | 19:47 |
| mat128 | if there a provision for deprecations in OSC? | 19:47 |
| mat128 | like oslo.config ? | 19:47 |
| rloo | mat128: yup, we need to be backwards compatible. deprecation period. hope to remove those in ocata. soon... | 19:47 |
| mat128 | ok | 19:47 |
| *** Goneri has quit IRC | 19:48 | |
| JayF | rloo: any reason I shouldn't land these if I'm the second +2 afayk? | 19:49 |
| rloo | JayF: I think it is fine. which ones are you referring to? | 19:49 |
| JayF | bug/1526479 | 19:49 |
| JayF | right now, specifically, node validate being added | 19:50 |
| rloo | JayF: which patches? there are a lot of them. | 19:50 |
| JayF | and I think I was the second +2 on node adopt as well | 19:50 |
| rloo | JayF: validate is good to go. | 19:50 |
| JayF | landing it | 19:50 |
| JayF | yeah, node adopt is the other | 19:50 |
| rloo | JayF: like i said, i'm taking another look. so eg, want to change 'passthru' to 'passthrough'. | 19:50 |
| JayF | +2 from me and dmitry | 19:50 |
| rloo | i think validate might depend on some other patches that haven't merged yet. adopt might be standalone :) | 19:51 |
| JayF | you're right, it's adopt -> console -> boot device + passthru -> validate | 19:52 |
| mat128 | https://review.openstack.org/#/c/357400/2/ironicclient/osc/v1/baremetal_node.py line 82 | 19:52 |
| mat128 | did we have that boot devices list hardcoded in our old client too? | 19:52 |
| rloo | mat128: yes | 19:53 |
| *** causten has joined #openstack-ironic | 19:53 | |
| mat128 | ok, then it's fine :) | 19:53 |
| mat128 | thank you rloo | 19:53 |
| JayF | mat128: https://github.com/openstack/python-ironicclient/blob/5e04d15a5a10587d15dab211689532bb4d45b77b/ironicclient/v1/node_shell.py#L583 | 19:53 |
| *** causten_ has joined #openstack-ironic | 19:53 | |
| rloo | mat128: there's lots of room for improvement, i have to tell myself to keep it simple for now. otherwise, too many patches/changes. | 19:54 |
| mat128 | yes, "porting" ironicclient to osc is the best move, then we can improve | 19:54 |
| mat128 | agreed | 19:54 |
| mat128 | thanks Jay | 19:54 |
| rloo | mat128: right. but it means we have to try to get the osc commands right the first time. | 19:54 |
| openstackgerrit | Thiago Paiva Brito proposed openstack/ironic: OneView test nodes to use dynamic allocation https://review.openstack.org/359990 | 19:54 |
| rloo | mat128: i just finished mulling over this -- let me know what your thoughts are. https://bugs.launchpad.net/python-ironicclient/+bug/1609338 | 19:55 |
| openstack | Launchpad bug 1609338 in python-ironicclient "Wrong layout of driver-related commands in OSC-plugin spec" [High,Triaged] | 19:55 |
| rloo | JayF: ^^ you too if you are interested | 19:55 |
| JayF | I'm kinda -0 on s/passthru/passthrough/, but I'm heavily biased due to my constant ability to misspell "through" | 19:56 |
| JayF | lol | 19:56 |
| mat128 | thr<tab> | 19:57 |
| mat128 | :D | 19:57 |
| rloo | JayF: English is through. Is 'thru' used in ops-land though? | 19:57 |
| JayF | I mean, drive-thru is common parlance | 19:57 |
| *** causten has quit IRC | 19:58 | |
| *** causten_ has joined #openstack-ironic | 19:58 | |
| JayF | Honestly I think it could go either way | 19:58 |
| JayF | but as someone who knows the api methods are called passthru, I'll personally screw it up a bit for at least a while | 19:58 |
| mat128 | rloo: I agree 1c + 2b make more sense to me | 19:58 |
| mat128 | rloo: and I think action should be nothing | 19:58 |
| rloo | JayF: I think we're supposed to think of osc as 'getting it right', being 'english' blah blah, and not think of our rest API or CLI. | 19:58 |
| JayF | 1c+2b agreed | 19:59 |
| mat128 | so it'd be something like "openstack baremetal driver passthrough agent_ipmitool bleh" | 19:59 |
| rloo | JayF: ^^ you agree? | 19:59 |
| JayF | passthru vs passthrough, I prefer passthru but I don't care that much | 19:59 |
| *** jhooker has quit IRC | 19:59 | |
| JayF | +1 to 1c+2b and the format mat128 lays out | 19:59 |
| JayF | http://www.dictionary.com/browse/thru?s=t thru is in the dictionary, but is listed as informal | 20:00 |
| JayF | *shrug* | 20:00 |
| JayF | I sit beside a giant physical set of OED books | 20:00 |
| * JayF looks it up | 20:00 | |
| JayF | (my wife has a Masters in English and is a teacher, hence the OED) | 20:01 |
| mat128 | TBH, we called it "passthru" in our docs / release notes: https://github.com/openstack/ironic/blob/7c855d5868e999dedf471246e93df5e5ac00e936/releasenotes/notes/vendor-passthru-shared-lock-6a9e32952ee6c2fe.yaml | 20:01 |
| mat128 | passthrough, 2 matches: https://github.com/openstack/ironic/search?utf8=%E2%9C%93&q=passthrough | 20:01 |
| rloo | JayF: wow, i doubt the younger generation will even know what OED is, or that it is a set of books. | 20:01 |
| JayF | this is actually the last ever paper edition of the OED | 20:01 |
| JayF | was a present to my wife when we were engaged :) | 20:02 |
| JayF | we also have a hardcopy of world book encyclopedia | 20:02 |
| JayF | (just in case the apocalypse comes, I have some knowledge on paper, lol) | 20:02 |
| rloo | mat128, JayF: I don't actually care if we use thru or through. Just that osc seems to prefer English. but if we leave it passthru is it less work for me! | 20:02 |
| JayF | thru is in OED as well, listed similarly as on the dictionary.com (syn for through, see through) | 20:02 |
| rloo | JayF: WHOA. i've got an old set of encyclopaedia, keep thinking i should throw them out. | 20:03 |
| rloo | JayF: ok, let's leave it as thru then. | 20:03 |
| *** mjturek has quit IRC | 20:03 | |
| JayF | I was homeschooled growing up and used to read our encyclopedia cover to cover when I was bored | 20:03 |
| JayF | so I have a bit of an affection for them :P | 20:03 |
| mat128 | guess what the word of the day on dictionary.com is? | 20:03 |
| rloo | JayF: cool! | 20:03 |
| mat128 | drum roll...... thrummy | 20:03 |
| JayF | lol | 20:03 |
| mat128 | http://www.dictionary.com/misspelling?term=passthru&s=t | 20:04 |
| mat128 | :( | 20:04 |
| mat128 | the hyphen would kill me | 20:04 |
| JayF | I don't always think it's evil for the api to be leaky, when it's also easier to understand | 20:04 |
| rloo | mat128: we can't use hyphens in the command so don't worry | 20:04 |
| JayF | is anyone going to doubt what passthru means (at least vs passthrough) | 20:05 |
| JayF | probably not | 20:05 |
| JayF | so why be different from the api? | 20:05 |
| mat128 | #startvote passthru | 20:05 |
| mat128 | =) | 20:05 |
| JayF | I'm honestly very curious what someone who speaks english as a second language would thing | 20:05 |
| JayF | *think | 20:05 |
| JayF | because that's my only concern | 20:05 |
| JayF | (about keeping thru) | 20:05 |
| mat128 | JayF: do I count? | 20:05 |
| mat128 | french is my first language | 20:05 |
| JayF | mat128: TIL :) | 20:06 |
| mat128 | :P | 20:06 |
| *** hj-hpe has quit IRC | 20:06 | |
| mat128 | asking folks around | 20:07 |
| *** _vsaienko has quit IRC | 20:08 | |
| mat128 | I'll be 100% honest with you... https://www.dropbox.com/s/jdabkplwqvfn1d7/Screen%20Shot%202016-08-24%20at%204.08.35%20PM.png?dl=0 | 20:09 |
| *** dprince has quit IRC | 20:09 | |
| JayF | ghetto :) | 20:09 |
| mat128 | his words | 20:10 |
| mat128 | lol | 20:10 |
| mat128 | late voter: [16:10:07] <glonlas> passtrough | 20:10 |
| mat128 | he even spelt it wrong | 20:10 |
| JayF | glorious | 20:10 |
| mat128 | prior art counts? http://php.net/manual/fr/function.passthru.php | 20:10 |
| rloo | ha ha. in this case, i think we could do an alias and satisfy everyone :) | 20:11 |
| JayF | I'd be very +1 to an alias | 20:11 |
| mat128 | same | 20:11 |
| JayF | thaat's the best of all worlds | 20:11 |
| mat128 | make the official one "passthrough" | 20:11 |
| JayF | like I said in the bug and here though: my preference is very, very small. I'd rather do the one that's quickest to get consensus and merged :) | 20:12 |
| rloo | JayF: do you create a console? open a console? what's the right verb? | 20:12 |
| mat128 | rloo: in Ironic you turn it on then get it's details | 20:12 |
| mat128 | enable/disable? | 20:12 |
| JayF | rloo: probably enable? | 20:12 |
| rloo | mat128, JayF: ok, enable/disable. i had it as set/unset, dtantsur|afk suggested create/delete. | 20:13 |
| mat128 | can you create more than one? | 20:13 |
| mat128 | (I very much know the answer to this one) | 20:13 |
| mat128 | let me rephrase that | 20:13 |
| JayF | can you enable for the same node multiple times? | 20:13 |
| mat128 | I prefer create/delete when you can have more than one | 20:13 |
| JayF | is that what you mean | 20:13 |
| mat128 | you probably can | 20:14 |
| mat128 | let me check | 20:14 |
| JayF | if I have an enabled console, and wanna re-get the details, do I do a node show, or do I do a node console enable? | 20:14 |
| JayF | (man, words are hard :D) | 20:14 |
| mat128 | the current command accepts turning it "on" multiple times | 20:14 |
| rloo | JayF: 'show' just shows information, doesn't *do* anything | 20:14 |
| mat128 | node console show | 20:15 |
| JayF | so you can console enable multiple times, but if already enabled it's a noop and essentially works like show/ | 20:15 |
| JayF | if that's valid, I like enable/disable | 20:15 |
| mat128 | enabling it doesnt even return anything | 20:15 |
| JayF | oh, well then that makes even *more* sense | 20:15 |
| JayF | perfect! | 20:15 |
| mat128 | http://paste.openstack.org/show/563128/ | 20:16 |
| JayF | I swear, the term "passthrough" has showed up in 4 other IRC channels since we just got done bikeshedding about it | 20:16 |
| mat128 | (this aint the new serial stuff, but behaves the same) | 20:16 |
| *** Goneri has joined #openstack-ironic | 20:16 | |
| JayF | the world is trolling me | 20:16 |
| mat128 | lol | 20:16 |
| rloo | JayF: ha ha | 20:17 |
| wajdi | Hey guys...just passing through | 20:17 |
| JayF | I like enable/disable or on/off then for that | 20:17 |
| wajdi | :) | 20:17 |
| rloo | hi wajdi! | 20:17 |
| wajdi | through? thru? :P | 20:17 |
| wajdi | hey rloo :) | 20:17 |
| mat128 | openstack baremetal console enable UUID | 20:17 |
| mat128 | ? | 20:17 |
| *** krtaylor has quit IRC | 20:18 | |
| mat128 | brb | 20:18 |
| rloo | mat128: I'm checking. enable/disable aren't action verbs for whatever reason: http://docs.openstack.org/developer/python-openstackclient/commands.html#actions | 20:19 |
| JayF | wajdi: someone was trolling me, but I threw them out the window ;) | 20:19 |
| JayF | if we have to pick from that list, start/stop are the best fits | 20:20 |
| JayF | but the description obviously isn't what we're looking for | 20:20 |
| JayF | can we not add actions to that list? enable/disable seem like something that would be a good fit in that list | 20:21 |
| rloo | JayF: yeah, i don't think we have to only use what is there. i was just wondering why enable/disable wasn't there. | 20:22 |
| JayF | because we haven't added it yet? | 20:22 |
| rloo | JayF: heh. on/off isn't there either. Maybe ironic is more complex than other projects :) | 20:23 |
| JayF | rloo: I knew that from day one ;) | 20:23 |
| *** causten_ has quit IRC | 20:30 | |
| mat128 | start/stop ? | 20:31 |
| mat128 | ah | 20:31 |
| mat128 | I should read all of the scrollback before sending anything :P | 20:31 |
| JayF | rloo: I've looked at all the pathces in that topic, feel free to ping me when you update for a re-review, or lmk if there's anything I can do to help | 20:32 |
| mat128 | that list was mostly written around nova if you ask me | 20:32 |
| rloo | thx JayF | 20:32 |
| jlvillal | rloo: On that operators patch. Is there a doc which explains operators? I'm trying to review it and getting confused. | 20:32 |
| rloo | mat128: well, around the 'core' projects | 20:32 |
| *** syed_ has quit IRC | 20:32 | |
| JayF | rloo: thank jim by proxy, he's the one who told me to pay more attention to OSC patches :) ... BTW that's always true for me, if you nneed reviews on a thing, point me at a topic and I'll look at them | 20:33 |
| rloo | jlvillal: lucas' patches? no doc. only code in oslo.utils. unless nova has some doc somewhere. | 20:33 |
| * JayF likes helping grease the rails to get stuff merged | 20:33 | |
| jlvillal | rloo: Yeah lucas' patch. | 20:33 |
| jlvillal | Thanks | 20:33 |
| rloo | thx JayF. you could look at the two ironic-lib patches if you want. we want those in for a release tomorrow. | 20:33 |
| JayF | rloo: is it a specific topic? or just pretty obvious which two | 20:34 |
| JayF | it's only two open | 20:34 |
| JayF | lol | 20:34 |
| rloo | JayF: should be obvious. sec. | 20:34 |
| rloo | JayF: yeah, those two :) | 20:34 |
| JayF | we have docs jobs running now on ironic-lib, btw | 20:35 |
| JayF | that merged earlier today | 20:35 |
| rloo | wrt 'openstack baremetal driver passthru list' and 'openstack baremetal driver passthru'. can't do it (with existing code). i just remembered, that's why that bug was opened. | 20:35 |
| mat128 | ann | 20:36 |
| mat128 | ahh* | 20:36 |
| rloo | ^ JayF. So we need some action/verb. 'openstack baremetal driver passthru ACTION' | 20:36 |
| mat128 | is it the parser thing? | 20:36 |
| mat128 | ah | 20:36 |
| rloo | mat128: yeah, the parser thing. | 20:36 |
| rloo | i can't wait to go back to coding. stay away from users and api's... | 20:36 |
| rloo | JayF: cool about the docs. | 20:36 |
| JayF | rloo: openstack baremetal driver passthru call <blah> ? | 20:36 |
| JayF | or do I misunderstand what we're talking about | 20:37 |
| JayF | I'm going to reopen the bug | 20:37 |
| mat128 | thats exactly it | 20:37 |
| rloo | JayF: so you like 'call'? I think I had 'run', 'execute'. | 20:37 |
| *** syed_ has joined #openstack-ironic | 20:37 | |
| mat128 | passthru list | 20:37 |
| mat128 | rloo: you were suggesting one of the following: 'call', 'run', 'execute', 'invoke' | 20:37 |
| JayF | execute is fine too, I don't really like run | 20:37 |
| rloo | mat128: yeah, unless you had something else. those were what came to mind. | 20:37 |
| JayF | +1 [call, execute] -1 [invole, run] | 20:38 |
| JayF | *invoke | 20:38 |
| JayF | whatever it is, people are going to have to read the docs or run help to know what's up | 20:38 |
| JayF | so I'm less concerned about what we call it, as long as it makes sense at all | 20:39 |
| mat128 | openstack baremetal driver passthru call agent_ipmitool lookup | 20:40 |
| mat128 | openstack baremetal driver passthru execute agent_ipmitool lookup | 20:40 |
| mat128 | openstack baremetal driver passthru run agent_ipmitool lookup | 20:40 |
| mat128 | openstack baremetal driver passthru invoke agent_ipmitool lookup | 20:40 |
| openstackgerrit | Ramamani Yeleswarapu proposed openstack/ironic: Centralize Config Options - patch merge, cleanup https://review.openstack.org/330160 | 20:40 |
| mat128 | if we go back to http://docs.openstack.org/developer/python-openstackclient/commands.html#actions ... | 20:41 |
| mat128 | none of them work | 20:41 |
| JayF | yep | 20:41 |
| *** sabeen has joined #openstack-ironic | 20:41 | |
| JayF | I don't think the name we pick there matters much at all | 20:41 |
| rloo | mat128: ha ha, i thought you found something there. yup, none of them work. | 20:41 |
| JayF | because nothing is going to be "obvious" | 20:41 |
| JayF | which means I'd say use something short like "call" or "run" (even though run wasn't my favorite originally) | 20:42 |
| wajdi | I would go with short as possible and choose between call or run | 20:42 |
| JayF | wajdi: you just thru me a bone! /s | 20:43 |
| rloo | ok, I'm going to use call then. I don't like run for a method. | 20:43 |
| *** ricardoas has joined #openstack-ironic | 20:43 | |
| wajdi | I came thru right on time ;) | 20:44 |
| mat128 | hahahaha | 20:44 |
| rloo | :D | 20:44 |
| mat128 | the problem we have is that passthru actions are actions themselves | 20:44 |
| mat128 | they could be top level and make sense | 20:44 |
| mat128 | call works for me | 20:45 |
| wajdi | between call/execute/run/invoke => +1 call | 20:46 |
| *** Goneri has quit IRC | 20:52 | |
| *** jhooker has joined #openstack-ironic | 20:56 | |
| mat128 | I have to go, will continue reviewing the OSC stuff tomorrow | 20:59 |
| rloo | thx mat128! | 21:00 |
| openstackgerrit | Merged openstack/ironic: Fix issues in dev-quickstart and index https://review.openstack.org/360003 | 21:01 |
| *** causten has joined #openstack-ironic | 21:02 | |
| jlvillal | rloo: So I see parse_root_device_hints() in both ironic-lib and in ironic. | 21:04 |
| jlvillal | rloo: Is that intentional? | 21:05 |
| rloo | jlvillal: i think ironic is supposed to call the one in ironic-lib. | 21:05 |
| rloo | jlvillal: sec, let me look. | 21:05 |
| jlvillal | rloo: I see it defined in: ironic/drivers/modules/deploy_utils.py | 21:06 |
| JayF | jlvillal: we're migrating it out of ironic into ironic-lib | 21:06 |
| JayF | jlvillal: so ipa can share the code | 21:06 |
| JayF | jlvillal: so we have to add it to -lib before it can go outta ironic aiui | 21:06 |
| jlvillal | JayF: Yeah, but these changes seem breaking changes. | 21:06 |
| rloo | jlvillal: I think lucas added it to the lib recently. july 22. | 21:06 |
| rloo | jlvillal: so i think his intent is for the code in ironic to call the lib. | 21:07 |
| jlvillal | I guess I don't understand why we aren't switching over first. and then make the changes. | 21:07 |
| rloo | jlvillal: that's a good question. i haven't been paying any attention to the order of things. | 21:07 |
| rloo | jlvillal: it might be cuz the ironic-lib deadline is this week so lucas has to get this done sooner rather than later. | 21:07 |
| rloo | jlvillal: well, if no one is using the ironic-lib version, then we don't have to worry about breakage. | 21:08 |
| rloo | jlvillal: yet. | 21:08 |
| jlvillal | rloo: True | 21:08 |
| openstackgerrit | Ruby Loo proposed openstack/ironic-specs: Changes to some OSC commands https://review.openstack.org/357505 | 21:10 |
| rloo | JayF, mat128: ^^ hopefully captures what we discussed :) | 21:12 |
| JayF | Yeah; I'm looking | 21:12 |
| JayF | really confused by LN118-120 | 21:13 |
| JayF | oh; I understand now. | 21:13 |
| jlvillal | Interesting to read that the OSIC cluster for gate testing is IPv6 only. I wonder if that will affect us? | 21:13 |
| JayF | it's probably public-ipv6-only | 21:13 |
| JayF | which means it shouldn't. | 21:13 |
| rloo | JayF: let me know if I can word it differently. i wanted to mention it so someone doesn't say, hey, where' the equivalent command... | 21:13 |
| * jlvillal thought someone (thiagop) said Ironic had problems with IPv6 for the gate | 21:14 | |
| jlvillal | JayF: Ah okay. | 21:14 |
| * rloo wonder what jlvillal is reading | 21:14 | |
| JayF | CLI users who want to see a list of nodes belonging to a given chassis should use `openstack baremetal node list --chassis`, since there is no equivalent to `ironic chassis-node-list` in OSC. | 21:14 |
| JayF | rloo: ^ wdyt? | 21:14 |
| jlvillal | rloo: http://lists.openstack.org/pipermail/openstack-dev/2016-August/102211.html | 21:14 |
| * jlvillal has been trying hard to at least glance at every email that goes to openstack-dev | 21:15 | |
| rloo | jlvillal: thx | 21:15 |
| rloo | JayF: well, what i want to say is that we intentionally don't want to provide an equivalent. | 21:15 |
| JayF | I mean, I'm going to +2 it anyway (barring any issue I haven't seen yet) | 21:16 |
| JayF | so feel free to ignore my comment as well :) | 21:16 |
| rloo | JayF: I can use what you wrote, with a 'since we are not providing an equivalent to ... :) | 21:16 |
| JayF | yeah; that's the only comment I have | 21:17 |
| JayF | I'm going to +2 it and you can just leave it be ;) | 21:17 |
| JayF | or change it and lmk and I'll revote | 21:17 |
| *** adnavare has quit IRC | 21:18 | |
| JayF | ...and gerrit is down | 21:18 |
| rloo | JayF: sigh | 21:18 |
| JayF | well, I'll vote when it works, at least :) | 21:19 |
| *** baoli_ has quit IRC | 21:23 | |
| *** aNupoisc has joined #openstack-ironic | 21:23 | |
| *** thiagop has quit IRC | 21:32 | |
| rloo | JayF: did you want to add your comments to that patch? Or I can update based on what you said here. | 21:33 |
| JayF | comments added and +2'd | 21:34 |
| jlvillal | JayF: So what kind of "rolling upgrade" capability do we have for IPA and Ironic. Like if Ironic gets upgraded do they require IPA to be updated? | 21:34 |
| jlvillal | I'm thinking about that operator patch and wonder what happens if Ironic or IPA has the change but not the other. What will happen. | 21:34 |
| JayF | jlvillal: Today, Ironic works with any release of IPA ever. | 21:34 |
| JayF | jlvillal: which is not sustainable, but sure is easy | 21:34 |
| jlvillal | I'm not sure if those changes would break that or not | 21:35 |
| JayF | for me it's hard to review stuff like that in the lib, right? | 21:35 |
| jlvillal | Me too | 21:35 |
| JayF | because that backwards compat is /not/ the job of ironic-lib | 21:35 |
| JayF | it's the job of the code in ipa/ironic | 21:35 |
| JayF | and the "right way" should be in i-l and the "old way" should be in the actual projects, with a deprecation period .... right? | 21:36 |
| JayF | jlvillal: actually, we do deprecate IPA stuff. We just did with the officialization of lookup/heartbeat | 21:36 |
| jlvillal | I think that ipa-lib should consider breaking compatibility | 21:36 |
| JayF | ipa-lib ? not a thing ? | 21:36 |
| jlvillal | Sorry. ironic-lib :) | 21:37 |
| openstackgerrit | Ruby Loo proposed openstack/ironic-specs: Changes to some OSC commands https://review.openstack.org/357505 | 21:40 |
| JayF | I'm not sure exactly what you mean then | 21:40 |
| JayF | either way; I'd put comments on the patch | 21:40 |
| JayF | lucas is obviously the expert on that | 21:40 |
| *** wajdi_ has joined #openstack-ironic | 21:40 | |
| rloo | JayF: I updated it. still more verbose than what you suggested, but hopefully clearer. | 21:40 |
| JayF | I like verbosity | 21:40 |
| jlvillal | JayF: I added some comments. Thanks. | 21:40 |
| rloo | jlvillal: ironic-lib itself doesn't *have* to consider breaking compatibility, whatever uses it though, should. | 21:41 |
| JayF | lgtm +2 | 21:41 |
| rloo | thx JayF. Will see what dtantsur|afk thinks. | 21:41 |
| jlvillal | rloo: I guess I'll be in the minority here, and disagree. If the library provides functions and then the signatures or return values change. That doesn't seem acceptable. | 21:42 |
| *** wajdi__ has joined #openstack-ironic | 21:42 | |
| JayF | jlvillal: that's explicitly why we say it's only usable by bare metal projects | 21:42 |
| JayF | jlvillal: so we can use the question of "does it actually break" instead of having to follow "real" library compat principles | 21:42 |
| rloo | jlvillal: yes, in that case. do we use upper constraints for ironic-lib or do we specify an exact version? | 21:43 |
| JayF | rloo: we bump the version in g-r when we use functionality in the new version | 21:43 |
| JayF | rloo: today it's =>2.0.0 in reqs.txt | 21:43 |
| *** wajdi has quit IRC | 21:44 | |
| jlvillal | JayF: Yeah. But in this case the code will be used in both Ironic and IPA. And to me seems like it will be breaking change if one gets updated and not the other. | 21:44 |
| rloo | JayF: right, but if we don't have upper constraints, an old version of ironic could potentially use a newer version of ironic-lib, right? | 21:44 |
| JayF | egad | 21:44 |
| jlvillal | rloo: upper-constraints has it as: ironic-lib===2.0.0 | 21:44 |
| JayF | rloo: you mean do we constrain it in stable/* branches, right? | 21:45 |
| rloo | jlvillal: in that case, should be ok. | 21:45 |
| jlvillal | In openstack/requirements | 21:45 |
| rloo | jlvillal: but good practice to do as you say. | 21:45 |
| jlvillal | JayF: We do | 21:45 |
| *** wajdi_ has quit IRC | 21:45 | |
| JayF | https://github.com/openstack/ironic/blob/stable/mitaka/requirements.txt#L16 | 21:45 |
| JayF | yes, we do :) | 21:45 |
| rloo | jlvillal: what happens in reality, someone running ironic. how does openstack/requirements fit in there? are they used? | 21:45 |
| jlvillal | JayF: And I think it is constrained in the gate also. | 21:45 |
| JayF | rloo: yes | 21:45 |
| JayF | rloo: we sync them over into requirements.txt in ironic | 21:46 |
| JayF | jlvillal: rloo actually... no | 21:46 |
| JayF | https://github.com/openstack/ironic/blob/stable/mitaka/requirements.txt#L16 | 21:46 |
| JayF | that is >= not == | 21:46 |
| JayF | which means we could break stable branches with new ironic-lib releases | 21:46 |
| JayF | egad | 21:46 |
| jlvillal | rloo: I am not sure. If someone just git clones. I don't they get constrained. | 21:46 |
| jlvillal | s/don't they/don't think they/ | 21:46 |
| *** wajdi__ has quit IRC | 21:47 | |
| rloo | jlvillal: then it might be a problem. eg, if you grab ubuntu's packaging of ironic. | 21:47 |
| JayF | yeah, rloo ^ the situation I lay out is the problem | 21:47 |
| jlvillal | But I think the stable branches expect people to abide by the upper-constraints file | 21:47 |
| JayF | if we ship ironic without == in the requiremenst for ironic-lib | 21:47 |
| JayF | then old releases installed from pip could pull in newer ironic-lib, and break things | 21:47 |
| rloo | ok, we need to fix or be aware of that... | 21:47 |
| JayF | yeah, I'm going to file a bug about it | 21:48 |
| rloo | thx JayF | 21:48 |
| jlvillal | You have to scroll to the right to see how it works: https://github.com/openstack/ironic/blob/stable/mitaka/tox.ini#L9 | 21:48 |
| jlvillal | But they use the stable/mitaka version of upper-constraints.txt | 21:48 |
| JayF | jlvillal: nobody uses tox except developers. | 21:48 |
| *** sdake has joined #openstack-ironic | 21:48 | |
| JayF | jlvillal: so that's not really valid to the question of "do we break users" | 21:48 |
| JayF | jlvillal: right now, requirements.txt says any version if ironic-lib 1.1.0 or newer works with stable/mitaka | 21:49 |
| JayF | that may not always be true. | 21:49 |
| jlvillal | JayF: I believe (don't know for sure) that they have guidelines on how to run a stable/release and use upper-constraints.txt | 21:49 |
| *** sdake_ has joined #openstack-ironic | 21:50 | |
| *** sdake has quit IRC | 21:54 | |
| JayF | https://bugs.launchpad.net/ironic/+bug/1616648 | 22:00 |
| openstack | Launchpad bug 1616648 in Ironic "Newer ironic-lib can be installed in stable branches" [Undecided,New] | 22:00 |
| JayF | jroll: you should look at https://bugs.launchpad.net/ironic/+bug/1616648 tomorrow | 22:01 |
| *** sdake_ has quit IRC | 22:01 | |
| *** vishwanathj has quit IRC | 22:02 | |
| *** jhooker has quit IRC | 22:03 | |
| *** sdake has joined #openstack-ironic | 22:03 | |
| *** hj-hpe has joined #openstack-ironic | 22:04 | |
| *** cdearborn has quit IRC | 22:07 | |
| openstackgerrit | Oleksandr Berezovskyi proposed openstack/ironic-inspector: [WIP] Implement rules for ports https://review.openstack.org/360104 | 22:13 |
| openstackgerrit | Oleksandr Berezovskyi proposed openstack/ironic-inspector: [WIP] Implement rules for ports https://review.openstack.org/360106 | 22:22 |
| jlvillal | JayF: Thanks for filing that bug | 22:23 |
| JayF | np | 22:23 |
| rajinir | krtaylor: The test_baremetal_server_ops is run twice by our CI. One from the tempest folder and other from ironic plugin. How can I make only one run? | 22:24 |
| rajinir | The test_baremetal_server_ops is run twice by our CI. One from the tempest folder and other from ironic plugin. How can I make only one run? | 22:24 |
| *** michauds has quit IRC | 22:30 | |
| *** jrist has quit IRC | 22:33 | |
| *** trandles has quit IRC | 22:34 | |
| openstackgerrit | Jay Faulkner proposed openstack/ironic-python-agent: Add metrics for IPA extensions https://review.openstack.org/360111 | 22:34 |
| rloo | jlvillal: what's the breaking change wrt https://review.openstack.org/#/c/348953/8/ironic_lib/utils.py? | 22:36 |
| jlvillal | rloo: What it returns. | 22:36 |
| jlvillal | rloo: Look at the unit tests | 22:37 |
| rloo | jlvillal: but nothing is using this method. | 22:37 |
| jlvillal | rloo: https://review.openstack.org/#/c/348953/8/ironic_lib/tests/test_utils.py@295 | 22:37 |
| jlvillal | rloo: Yes. But the code is copied from Ironic. And IPA expects to get it in that format. | 22:38 |
| jlvillal | rloo: So if Ironic starts using this new code and then run an old IPA. I think it will not work. | 22:38 |
| jlvillal | I could be wrong. | 22:38 |
| JayF | jlvillal: IMO that should be handled /in IPA and Ironic/ | 22:39 |
| JayF | jlvillal: not in ironic-lib | 22:39 |
| rloo | jlvillal: oh. maybe be more explicit in your review then. i don't know what lucas is intending and/or maybe he didn't think of that. | 22:39 |
| jlvillal | I'll still continue to disagree. | 22:39 |
| jlvillal | rloo: Okay | 22:39 |
| JayF | IMO we shouldn't " introduce code into ironic-lib to | 22:39 |
| JayF | blah, [enter] mid-edit | 22:39 |
| *** anush_ has quit IRC | 22:39 | |
| JayF | IMO we shouldn't introduce "old"code into ironic-lib to | 22:40 |
| JayF | IMO we shouldn't introduce "old" code into ironic-lib just to ensure compatibility | 22:40 |
| rloo | well, i can't agree or disagree w/o knowing how it will be used. i asked at some point but i know he's in a hurry to get it into the library. | 22:40 |
| JayF | there we go, I need to avoid the enter key better, haha | 22:40 |
| *** causten has quit IRC | 22:40 | |
| jlvillal | JayF: The old code is already there. | 22:41 |
| JayF | And it's compataible with the preexisting stuff in ipa/ironic/ | 22:41 |
| jlvillal | The parse_root_device_hints() is in both ironic-lib and ironic | 22:41 |
| jlvillal | Currently not used in ironic-lib from what I can tell. | 22:41 |
| *** bfournie has quit IRC | 22:42 | |
| jlvillal | What is currently there works with IPA. | 22:42 |
| jlvillal | My concern is that the changes will break IPA. First because the IPA code expects it to be in the format that it is currently. | 22:42 |
| jlvillal | Currently it is like: 'size': 12345 | 22:43 |
| jlvillal | new way would be: 'size': '== 12345' | 22:44 |
| *** rpioso has quit IRC | 22:44 | |
| jlvillal | And strings like: 'wnm': "foo" become 'wnm': "s== foo" | 22:44 |
| jlvillal | I would think IPA will need to be changed to understand that. | 22:44 |
| JayF | maybe so, I'm curious what lucas-dinner will say | 22:45 |
| jlvillal | Me too | 22:45 |
| *** jrist has joined #openstack-ironic | 22:45 | |
| *** rcernin has quit IRC | 22:53 | |
| *** zigo has quit IRC | 22:56 | |
| *** sdake has quit IRC | 22:57 | |
| *** zigo has joined #openstack-ironic | 22:57 | |
| *** sdake has joined #openstack-ironic | 23:07 | |
| jroll | <tonyb:#openstack-ironic> jroll: we shoudl chat about that constraints/dockerfile review See if we can muster someone on the ironic side to run with it <- tonyb, I mentioned in the review "mat128 volunteered to do the dockerfile bits this week." and I guess it's in progress here https://review.openstack.org/#/c/358855/ | 23:08 |
| jroll | tonyb: I see you saw that now :D | 23:08 |
| jroll | anteaya: some of us have seen that user story. no surprises in there. things that bend the laws of physics, maybe, but no surprises :) | 23:15 |
| jroll | dtantsur|afk: we don't use i18n in IPA because it isn't user facing. open to changing that, but it's been the reason in the past | 23:15 |
| anteaya | jroll: thank you | 23:15 |
| anteaya | jroll: mostly my point was don't let the term 'user story' dissuage operators from joining project channels and weekly meetings and speaking up about their needs | 23:16 |
| tonyb | jroll: Yeah jlvillal pointed me at it :) | 23:16 |
| tonyb | jroll: thanks for following up. | 23:16 |
| anteaya | there seems to be a feeling that a user story prevents folks from joining in | 23:16 |
| jroll | anteaya: +1 | 23:18 |
| jroll | tonyb: no problem | 23:18 |
| jroll | I only skimmed scrollback, so if I didn't respond to something that folks super care about please ping me again | 23:18 |
| jroll | JayF: on that bug... that's an openstack-wide problem, generally regarded as fixed by upper-constraints (as in, we assume people are supposed to use them, and people know to use them) | 23:19 |
| jroll | jlvillal: ^^ | 23:20 |
| jlvillal | jroll: Thanks | 23:20 |
| jroll | patches to g-r to cap things will almost certainly be -2'd | 23:20 |
| jroll | but you can confirm with tonyb, requirements, ptl, if you like | 23:20 |
| tonyb | jroll: Yeah probably but each one is assessed on merits, but you're right if it's somethign that's covered by upper-constraints then it'll probably get -2'd | 23:21 |
| tonyb | jroll: which bug/review? | 23:21 |
| jroll | tonyb: https://bugs.launchpad.net/ironic/+bug/1616648 | 23:22 |
| openstack | Launchpad bug 1616648 in Ironic "Newer ironic-lib can be installed in stable branches" [Undecided,New] | 23:22 |
| tonyb | jroll: Thanks. | 23:26 |
| jroll | np, thanks for jumping in :P | 23:26 |
| tonyb | jroll, JayF: I added my $0.02 | 23:26 |
| jroll | ty sir | 23:26 |
| tonyb | jroll: always happy to help. Most of the time I'm just a nick highlight away (if I'm awake) | 23:27 |
| jroll | tonyb: that's why we love you ;) | 23:27 |
| tonyb | :) | 23:27 |
| tonyb | awww shucks | 23:27 |
| *** krtaylor has joined #openstack-ironic | 23:33 | |
| *** rama_y has quit IRC | 23:33 | |
| *** rama_y has joined #openstack-ironic | 23:36 | |
| *** jcoufal has joined #openstack-ironic | 23:36 | |
| *** jcoufal has quit IRC | 23:44 | |
| *** rama_y has quit IRC | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
