| openstackgerrit | Xavier proposed openstack/ironic: Add Dynamic Allocation feature for the OneView drivers https://review.openstack.org/286192 | 00:00 |
|---|---|---|
| *** rama_y has quit IRC | 00:04 | |
| *** spandhe has joined #openstack-ironic | 00:04 | |
| *** sacharya has quit IRC | 00:09 | |
| *** serverascode has quit IRC | 00:17 | |
| *** evgenyl has quit IRC | 00:17 | |
| *** agordeev has quit IRC | 00:17 | |
| *** cppforlife_ has quit IRC | 00:17 | |
| *** NobodyCam has quit IRC | 00:17 | |
| *** sergek has quit IRC | 00:17 | |
| *** yonglihe has quit IRC | 00:18 | |
| *** agordeev has joined #openstack-ironic | 00:18 | |
| *** BadCub has quit IRC | 00:19 | |
| *** patrickeast has quit IRC | 00:19 | |
| *** serverascode has joined #openstack-ironic | 00:19 | |
| openstackgerrit | Merged openstack/ironic-lib: Remove deprecated disk util configs https://review.openstack.org/318834 | 00:20 |
| *** yonglihe has joined #openstack-ironic | 00:20 | |
| *** cppforlife_ has joined #openstack-ironic | 00:20 | |
| *** amotoki has quit IRC | 00:20 | |
| *** NobodyCam has joined #openstack-ironic | 00:20 | |
| *** patrickeast has joined #openstack-ironic | 00:22 | |
| *** evgenyl has joined #openstack-ironic | 00:26 | |
| *** BadCub has joined #openstack-ironic | 00:26 | |
| *** sergek has joined #openstack-ironic | 00:27 | |
| *** zigo has quit IRC | 00:37 | |
| *** amotoki has joined #openstack-ironic | 00:41 | |
| *** hoangcx has joined #openstack-ironic | 00:44 | |
| *** sacharya has joined #openstack-ironic | 00:44 | |
| *** mtanino has joined #openstack-ironic | 00:50 | |
| *** aNupoisc has quit IRC | 00:51 | |
| *** amotoki has quit IRC | 00:55 | |
| *** baoli has joined #openstack-ironic | 01:00 | |
| *** zigo has joined #openstack-ironic | 01:03 | |
| *** ayoung has joined #openstack-ironic | 01:07 | |
| *** chlong has joined #openstack-ironic | 01:16 | |
| *** sdake has joined #openstack-ironic | 01:25 | |
| *** PollyZ has joined #openstack-ironic | 01:54 | |
| *** ijw has quit IRC | 01:57 | |
| *** PollyZ has quit IRC | 02:00 | |
| *** PollyZ has joined #openstack-ironic | 02:05 | |
| *** mtanino has quit IRC | 02:05 | |
| *** mtanino has joined #openstack-ironic | 02:10 | |
| *** PollyZ has quit IRC | 02:13 | |
| *** Goneri has joined #openstack-ironic | 02:14 | |
| *** PollyZ has joined #openstack-ironic | 02:18 | |
| *** Nisha_away has joined #openstack-ironic | 02:24 | |
| *** PollyZ has quit IRC | 02:35 | |
| *** Nisha_brb has joined #openstack-ironic | 02:39 | |
| *** Nisha_away has quit IRC | 02:43 | |
| *** yuanying has quit IRC | 02:51 | |
| *** PollyZ has joined #openstack-ironic | 02:51 | |
| *** PollyZ has quit IRC | 02:55 | |
| *** amotoki has joined #openstack-ironic | 02:57 | |
| *** vishwanathj has joined #openstack-ironic | 02:58 | |
| *** Nisha_brb has quit IRC | 03:00 | |
| *** rbudden has quit IRC | 03:04 | |
| *** rloo has quit IRC | 03:12 | |
| *** Nisha_away has joined #openstack-ironic | 03:12 | |
| *** Nisha_away has quit IRC | 03:16 | |
| *** jaybeale has joined #openstack-ironic | 03:18 | |
| *** PollyZ has joined #openstack-ironic | 03:23 | |
| *** jaybeale has quit IRC | 03:24 | |
| *** jaybeale has joined #openstack-ironic | 03:30 | |
| *** spandhe has quit IRC | 03:37 | |
| *** PollyZ has quit IRC | 03:38 | |
| *** sdake has quit IRC | 03:41 | |
| *** PollyZ has joined #openstack-ironic | 03:41 | |
| *** jaybeale has quit IRC | 03:41 | |
| *** PollyZ has quit IRC | 03:46 | |
| *** yuanying has joined #openstack-ironic | 03:48 | |
| *** Goneri has quit IRC | 03:49 | |
| *** mtanino has quit IRC | 03:58 | |
| *** baoli has quit IRC | 04:00 | |
| *** links has joined #openstack-ironic | 04:00 | |
| *** vishwanathj has quit IRC | 04:10 | |
| *** vishwanathj has joined #openstack-ironic | 04:10 | |
| *** vmud213 has joined #openstack-ironic | 04:14 | |
| *** vmud213 has quit IRC | 04:19 | |
| *** vmud213 has joined #openstack-ironic | 04:20 | |
| *** sdake has joined #openstack-ironic | 04:46 | |
| *** spandhe has joined #openstack-ironic | 04:58 | |
| openstackgerrit | Moshe Levi proposed openstack/ironic-python-agent: Add vendor id and vendor name to interface information https://review.openstack.org/279940 | 04:58 |
| *** vmud213 has quit IRC | 05:13 | |
| *** sabeen has quit IRC | 05:22 | |
| *** wajdi has quit IRC | 05:51 | |
| *** chlong has quit IRC | 06:03 | |
| *** yolanda has joined #openstack-ironic | 06:06 | |
| *** Fdaisuke has quit IRC | 06:07 | |
| *** fragatina has quit IRC | 06:13 | |
| *** Fdaisuke has joined #openstack-ironic | 06:13 | |
| *** hoangcx has quit IRC | 06:19 | |
| *** hoangcx has joined #openstack-ironic | 06:19 | |
| *** chlong has joined #openstack-ironic | 06:20 | |
| openstackgerrit | Yibo Cai proposed openstack/ironic: Support multi arch deployment https://review.openstack.org/322632 | 06:22 |
| *** openstackgerrit has quit IRC | 06:32 | |
| *** openstackgerrit has joined #openstack-ironic | 06:33 | |
| *** fragatina has joined #openstack-ironic | 06:33 | |
| *** chlong has quit IRC | 06:35 | |
| *** fragatina has quit IRC | 06:36 | |
| *** fragatina has joined #openstack-ironic | 06:37 | |
| *** pcaruana has joined #openstack-ironic | 06:40 | |
| *** links has quit IRC | 06:55 | |
| *** hparekh has quit IRC | 06:58 | |
| *** hparekh has joined #openstack-ironic | 07:02 | |
| *** links has joined #openstack-ironic | 07:10 | |
| *** spandhe has quit IRC | 07:10 | |
| *** rcernin has joined #openstack-ironic | 07:17 | |
| *** hoangcx has quit IRC | 07:19 | |
| *** hoangcx has joined #openstack-ironic | 07:20 | |
| *** sacharya has quit IRC | 07:22 | |
| *** dtantsur|afk is now known as dtantsur | 07:25 | |
| dtantsur | Morning Ironic | 07:26 |
| *** tesseract has joined #openstack-ironic | 07:27 | |
| dtantsur | our last grenade patch was approved \o/ | 07:28 |
| *** openstackgerrit has quit IRC | 07:33 | |
| *** openstackgerrit has joined #openstack-ironic | 07:33 | |
| *** ccamacho has joined #openstack-ironic | 07:38 | |
| ccamacho | Hey folks!!! good morning! Quick question, he are hitting this issue randomly in TripleO https://bugs.launchpad.net/tripleo/+bug/1546749 would you mind to briefly read it?? just to see if you have some extra clues for the fix | 07:42 |
| openstack | Launchpad bug 1546749 in tripleo "CI testenv vm stuck in boot loop trying to boot rescue kernel" [Critical,Invalid] - Assigned to James Slagle (james-slagle) | 07:42 |
| dtantsur | ccamacho, hi, reading | 07:46 |
| ccamacho | dtantsur thanks man! | 07:47 |
| *** vishwanathj has quit IRC | 07:49 | |
| dtantsur | ccamacho, so, are you still hitting it? slagle wrote that he no longer faces it | 07:49 |
| dtantsur | it does not look familiar, but I'm well aware that our bootloader installation is fragile | 07:50 |
| ccamacho | yesterday, thanks the thing there should be a kind of race condition and doesn't happens all the time | 07:50 |
| *** vishwanathj has joined #openstack-ironic | 07:50 | |
| ccamacho | The vms start, they load the initial image, they do "something" and after the reboot... Dead.. | 07:51 |
| ccamacho | I hit that when working with https://bugs.launchpad.net/tripleo/+bug/1590755 | 07:52 |
| openstack | Launchpad bug 1590755 in tripleo "TripleO deployment fails using Centos7 + stable/liberty" [Undecided,Fix committed] - Assigned to Carlos Camacho (ccamacho) | 07:52 |
| dtantsur | aha, so it's liberty, right? | 07:56 |
| dtantsur | ccamacho, do you see it on mitaka and/or master? | 07:56 |
| ccamacho | both :$ it is everywhere... | 07:57 |
| *** zzzeek has quit IRC | 08:00 | |
| *** zzzeek has joined #openstack-ironic | 08:00 | |
| *** jtomasek_ has joined #openstack-ironic | 08:03 | |
| openstackgerrit | Christopher Brown proposed openstack/ironic-specs: LVM/Software RAID support in ironic-python-agent https://review.openstack.org/327769 | 08:03 |
| *** e0ne has joined #openstack-ironic | 08:05 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Imported Translations from Zanata https://review.openstack.org/328147 | 08:05 |
| *** ohamada has joined #openstack-ironic | 08:05 | |
| *** ohamada has quit IRC | 08:07 | |
| dtantsur | ccamacho, do you have an example handy? | 08:07 |
| *** ohamada has joined #openstack-ironic | 08:09 | |
| ccamacho | dtantsur, I was able so see that behavior by connecting to the vms with virt-manager to the hypervisor, then when creating the overcloud you can see how they boot up, restart and die... But the actual problem is that is random.. you dont have the error all the time :$ Ill try to re-execute it and record the screen... | 08:10 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/ironic: [WIP] Testing latest u-c https://review.openstack.org/318440 | 08:11 |
| openstackgerrit | Davanum Srinivas (dims) proposed openstack/ironic: [WIP] Testing latest u-c https://review.openstack.org/318440 | 08:11 |
| *** athomas has joined #openstack-ironic | 08:11 | |
| dtantsur | ccamacho, yeah, it would be valuable to see what ironic is doing with the machine during deployment.. unfortunately we don't gather ramdisk logs for now.. | 08:11 |
| dtantsur | lucas-afk, a perfect example why we have to be able to collect logs even on success ^^^ | 08:11 |
| ccamacho | dtantsur, would you point me to the things i need to save when this appears again? | 08:12 |
| *** e0ne has quit IRC | 08:12 | |
| alineb | morning | 08:12 |
| dtantsur | morning alineb | 08:12 |
| *** e0ne has joined #openstack-ironic | 08:12 | |
| dtantsur | ccamacho, screencast of the node boot during deployment, probably. also ironic-conductor logs | 08:13 |
| alineb | dtantsur o/ | 08:13 |
| ccamacho | dtantsur, Ill record it, one more thing, this is making fail lot of CI jobs due a timeout in the deployment, so if you see a timeout this is a probable cause.. Thanks for the help! | 08:16 |
| dtantsur | got it, thanks | 08:17 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Move create-node.sh to create_vms function https://review.openstack.org/322850 | 08:17 |
| *** jtomasek_ has quit IRC | 08:20 | |
| *** sdake has quit IRC | 08:21 | |
| *** sacharya has joined #openstack-ironic | 08:22 | |
| *** kromanenko_ has joined #openstack-ironic | 08:26 | |
| *** Romanenko_K has joined #openstack-ironic | 08:26 | |
| *** kromanenko__ has joined #openstack-ironic | 08:26 | |
| *** hoangcx_ has joined #openstack-ironic | 08:26 | |
| *** hoangcx has quit IRC | 08:26 | |
| *** sacharya has quit IRC | 08:27 | |
| *** ifarkas has joined #openstack-ironic | 08:37 | |
| openstackgerrit | Christopher Brown proposed openstack/ironic-specs: LVM/Software RAID support in ironic-python-agent https://review.openstack.org/327769 | 08:51 |
| *** Fdaisuke_ has joined #openstack-ironic | 08:54 | |
| *** Fdaisuke has quit IRC | 08:55 | |
| *** daemontool has joined #openstack-ironic | 08:59 | |
| vdrok | morning ironic, dtantsur alineb | 09:00 |
| dtantsur | morning vdrok | 09:00 |
| *** Fdaisuke_ has quit IRC | 09:01 | |
| *** Fdaisuke has joined #openstack-ironic | 09:01 | |
| alineb | vdrok, morning | 09:01 |
| sergek | morning alineb , vdrok , dtantsur | 09:07 |
| dtantsur | morning sergek | 09:07 |
| vdrok | sergek: morning :) | 09:07 |
| alineb | morning sergek | 09:08 |
| *** SiRiuS__ has joined #openstack-ironic | 09:09 | |
| sambetts | Morning all TGIF!!!!!!! | 09:12 |
| pas-ha | morning sambetts and all | 09:13 |
| sambetts | o/ pas-ha | 09:14 |
| dtantsur | morning sambetts, pas-ha | 09:17 |
| dtantsur | yeah, TGIF finally :) | 09:17 |
| pas-ha | :) | 09:17 |
| openstackgerrit | Szymon Borkowski proposed openstack/ironic-python-agent: Add PCI devices collector to inspector https://review.openstack.org/326318 | 09:18 |
| vdrok | morning sambetts and pas-ha ! | 09:20 |
| openstackgerrit | Dmitry Tantsur proposed openstack/python-ironicclient: Add --wait flag for provision actions and wait_for_provision_state function https://review.openstack.org/328191 | 09:26 |
| dtantsur | hmm, do unit tests for ironicclient pass for anyone on master? | 09:27 |
| sambetts | dtantsur: give me a minute and I'll run them :) | 09:27 |
| sambetts | dtantsur: pulled latest, and ran tox -r, and I'll seeing all green here | 09:29 |
| sambetts | I'm seeing* | 09:29 |
| dtantsur | weird | 09:29 |
| sambetts | dtantsur: Are you seeing a failure | 09:30 |
| sambetts | ? | 09:30 |
| dtantsur | sambetts, yes, a lot of "TypeError: 'self' parameter lacking default value" | 09:30 |
| dtantsur | essentially the whole ironicclient.tests.unit.test_client.ClientTest fails | 09:31 |
| sambetts | :/ oh really weird | 09:32 |
| openstackgerrit | Dmitry Tantsur proposed openstack/python-ironicclient: [WIP] Add --wait flag for provision actions and wait_for_provision_state function https://review.openstack.org/328191 | 09:32 |
| openstackgerrit | Szymon Borkowski proposed openstack/ironic-python-agent: Add PCI devices collector to inspector https://review.openstack.org/326318 | 09:33 |
| *** dtantsur is now known as dtantsur|lunch | 09:42 | |
| *** mbound has joined #openstack-ironic | 09:48 | |
| *** marksei has joined #openstack-ironic | 09:54 | |
| *** athomas has quit IRC | 09:57 | |
| *** athomas has joined #openstack-ironic | 10:03 | |
| sambetts | sigh... why is there no freakin horizontal scroll in gerrit... I can't review a patch properly if I cant see the end of a line... | 10:03 |
| *** sacharya has joined #openstack-ironic | 10:05 | |
| *** hoangcx has joined #openstack-ironic | 10:09 | |
| *** sacharya has quit IRC | 10:10 | |
| *** hoangcx_ has quit IRC | 10:11 | |
| *** daemontool has quit IRC | 10:17 | |
| openstackgerrit | Andrey Shestakov proposed openstack/ironic: [WIP] [POC] Torrent based image provisioning https://review.openstack.org/278469 | 10:18 |
| aarefiev | morning all! | 10:22 |
| aarefiev | dtantsur: morning, about this one https://review.openstack.org/#/c/326620 | 10:23 |
| aarefiev | not get it, how does iscsi disk related to local_gb property | 10:23 |
| *** mgould|afk is now known as mgould | 10:24 | |
| mgould | morning Ironic | 10:33 |
| sambetts | Hi aarefiev, mgould | 10:33 |
| aarefiev | hey sambetts | 10:33 |
| aarefiev | mgould o/ | 10:34 |
| mgould | morning sambetts aarefiev dtantsur|lunch vdrok | 10:34 |
| vdrok | morning mgould :) | 10:34 |
| *** cdearborn has joined #openstack-ironic | 10:34 | |
| aarefiev | sambetts: please take a look on https://review.openstack.org/#/c/327492/ it block tempest tests | 10:36 |
| *** daemontool has joined #openstack-ironic | 10:38 | |
| sambetts | aarefiev: sure | 10:39 |
| aarefiev | thanks | 10:40 |
| *** vishwanathj has quit IRC | 10:45 | |
| *** vishwanathj has joined #openstack-ironic | 10:48 | |
| openstackgerrit | Christopher Brown proposed openstack/ironic-specs: LVM/Software RAID support in ironic-python-agent https://review.openstack.org/327769 | 10:49 |
| *** marksei_ has joined #openstack-ironic | 10:50 | |
| *** marksei_ has quit IRC | 10:51 | |
| ccamacho | dtantsur, in the bug comments, at the end of https://bugs.launchpad.net/tripleo/+bug/1546749 I have pushed the screencasts and log files from undercloud. As you can see there, I just have re-executed the command, and the second time went fine. | 10:52 |
| openstack | Launchpad bug 1546749 in tripleo "CI testenv vm stuck in boot loop trying to boot rescue kernel" [Critical,Invalid] - Assigned to James Slagle (james-slagle) | 10:52 |
| xavierr | good morning Ironic | 10:52 |
| xavierr | TGIF | 10:52 |
| sambetts | o/ xavierr | 10:52 |
| xavierr | \o sambetts | 10:52 |
| * xavierr needs coffee | 10:53 | |
| mgould | morning xavierr | 10:53 |
| * mgould too | 10:53 | |
| xavierr | mgould morning :) | 10:54 |
| *** marksei has quit IRC | 10:54 | |
| TheJulia | good morning everyone | 10:54 |
| xavierr | hey TheJulia ;) | 10:55 |
| *** hoangcx has quit IRC | 10:57 | |
| *** dtantsur|lunch is now known as dtantsur | 10:58 | |
| ccamacho | o/ | 10:58 |
| dtantsur | aarefiev, sorry, I don't get the question. an iscsi disk is a disk, it can have size... | 10:58 |
| dtantsur | also morning aarefiev, mgould, xavierr, TheJulia | 10:59 |
| *** vishwanathj has quit IRC | 10:59 | |
| *** vishwanathj has joined #openstack-ironic | 11:00 | |
| aarefiev | dtantsur: yeah, but local_gb shows hypervisor local disk size, why ironic should report info about iscsi disk here | 11:00 |
| dtantsur | aarefiev, local_gb shows the root disk, it's not required to be local | 11:01 |
| dtantsur | aarefiev, on the other hand, maybe we'd better support such disks in inspection, like in https://bugs.launchpad.net/ironic-python-agent/+bug/1590606 | 11:01 |
| openstack | Launchpad bug 1590606 in ironic-python-agent "[RFE] Support remote iSCSI shares as root devices via iscsistart" [Wishlist,In progress] - Assigned to Yossi Ovadia (jabadia) | 11:01 |
| dtantsur | so I'm not sure what to think about this patch | 11:01 |
| * dtantsur W-1 | 11:02 | |
| mgould | morning TheJulia | 11:03 |
| dtantsur | also, is our gate broken? I see strange failures in the IPA gate | 11:03 |
| xavierr | morning dtantsur o/ | 11:04 |
| sambetts | dtantsur: the error to do with horzion not startig? | 11:04 |
| dtantsur | sambetts, maybe, did not dig in.. but it's not on devstack stage, rather later | 11:05 |
| *** cdearborn has quit IRC | 11:05 | |
| aarefiev | dtantsur: yeah, I need to think about this more too, but imo it's not right way to add iscsi to local_gb | 11:05 |
| sambetts | dtantsur: oh, was seeing a weird error on one of the patches, where devstack crashed out when trying to start horizon | 11:05 |
| sambetts | dtantsur: I thought we had horizon disabled in our gate... | 11:05 |
| dtantsur | aarefiev, sorry, a confusion here. if iscsi is your root device you must put it in local_gb | 11:06 |
| dtantsur | aarefiev, however, we should be able to transparently handle it in IPA, see the bug I've linked it | 11:06 |
| dtantsur | sambetts, I would expect it too | 11:06 |
| aarefiev | dtantsur: should I, why | 11:06 |
| dtantsur | aarefiev, you need local_gb of you root device for nova to work | 11:06 |
| dtantsur | otherwise it won't be possible to schedule an instance on this node | 11:07 |
| aarefiev | dtantsur: why not local_gb=0, and boot from disk | 11:08 |
| dtantsur | aarefiev, local_gb=0 won't work, nova won't use it for scheduling. also it's not true, your disk is not 0 GiB, it's remote | 11:08 |
| dtantsur | it has a size etc. it's just get set by open-iscsi on system start up | 11:09 |
| sambetts | dtantsur: how does this work for nodes that are truly diskless, e.g. don't have a disk until Ironic attaches the one requested via nova boot | 11:10 |
| *** dmk0202 has joined #openstack-ironic | 11:10 | |
| dtantsur | sambetts, the thing I'm talking about is not about ironic attaching anything. they configure the disk in BIOS, then it gets passed to kernel (somehow), then 'iscsistart -b' command attaches it to the node | 11:10 |
| dtantsur | again, please see https://bugs.launchpad.net/ironic-python-agent/+bug/1590606 | 11:11 |
| openstack | Launchpad bug 1590606 in ironic-python-agent "[RFE] Support remote iSCSI shares as root devices via iscsistart" [Wishlist,In progress] - Assigned to Yossi Ovadia (jabadia) | 11:11 |
| aarefiev | sambetts: +1, also from nova side it's more related to instance, not hypervisor | 11:11 |
| dtantsur | I was not aware of such thing before yesterday, so I may be not the best source of truth :) | 11:11 |
| sambetts | dtantsur: I assumed this patch is about diskless nodes that already have an out of band disk attached | 11:11 |
| sambetts | dtantsur: but the boot from volume on truly diskless nodes is going to be interesting | 11:11 |
| dtantsur | sambetts, "already have" - the parameters are provided by BIOS. so from out PoV it should be transparent | 11:12 |
| sambetts | because then local_gb really will be 0 | 11:12 |
| dtantsur | yeah, true boot from volume is tricky :) | 11:12 |
| dtantsur | but this one is kind of a "fake", as ironic and nova are not involved in attaching a remove volume | 11:12 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: WIP: patch to test tempest-full job https://review.openstack.org/328242 | 11:12 |
| dtantsur | and I'd like to support it already, because it seems like we can make it work pretty easily | 11:12 |
| sambetts | right, the connected to an out of band iscsi volume, case this patch makes sense to me | 11:13 |
| sambetts | because from Ironics perspective its just another type of local disk | 11:13 |
| dtantsur | right | 11:14 |
| sambetts | dtantsur: can we deploy successfully to that kind of device? | 11:14 |
| sambetts | dtantsur: I've never tried, iscsi deploy would be weird right? because it would be an iscsi mount to an iscsi mount | 11:14 |
| aarefiev | what if disk attached to multiple nodes? | 11:14 |
| dtantsur | sambetts, I don't know. probably with some limitations, it needs further checking. unfortunately, I don't have access to such hardware | 11:14 |
| dtantsur | aarefiev, that's operator's problem really :) | 11:15 |
| openstackgerrit | Szymon Borkowski proposed openstack/ironic-python-agent: Add PCI devices collector to inspector https://review.openstack.org/326318 | 11:15 |
| sambetts | dtantsur: I think my hardware can do it, but my main development testbed is being borrowed for something else at the moment so I can't confirm | 11:15 |
| dtantsur | sambetts, would be interesting indeed | 11:15 |
| sambetts | I can definatly connect to an iscsi volume oob | 11:15 |
| sambetts | dtantsur: but I don't know if that is transparent to the OS or not | 11:16 |
| dtantsur | yeah, that's interesting. | 11:16 |
| dtantsur | the approach in this bug is not entirely transparent, but it seems to use some well defined and easy to use protocol | 11:16 |
| aarefiev | neverless I don't think iscsi disk should be in node property | 11:16 |
| aarefiev | only as separate resource | 11:17 |
| dtantsur | aarefiev, why? it's just a disk. remove iscsi from your sentence and explain why plesae :) | 11:17 |
| dtantsur | sambetts, aha: https://en.wikipedia.org/wiki/ISCSI_Boot_Firmware_Table | 11:17 |
| sambetts | aarefiev: in this case from Ironics perspective, iscsi is no different to it being connected via a sata cable | 11:18 |
| sambetts | aarefiev: because Ironic shouldn't know about it | 11:18 |
| sambetts | its just knows that node has access to N gb of disk space | 11:19 |
| sambetts | and thats all local_gb cares about | 11:19 |
| dtantsur | and from man iscsistart "[-b|--fwparam_connect] Create a session to the target using the iBFT or OF info" | 11:19 |
| dtantsur | that's impressive, I didn't know about that | 11:19 |
| sambetts | dtantsur: yeah, thats pretty neat, I believe in hardware.py we run iscsistart -b | 11:20 |
| dtantsur | UCSM 2.0 iSCSI boot: https://supportforums.cisco.com/sites/default/files/legacy/6/9/0/59096-iSCSI_Boot_updated.pdf ;) | 11:20 |
| aarefiev | yeah but we attach disk to instance, and ironic is like hypervisor | 11:20 |
| dtantsur | aarefiev, who "we"? we don't, an operator does. | 11:21 |
| dtantsur | I think there is a bit confusion here on which iSCSI method we're talking about | 11:21 |
| aarefiev | ok, operator | 11:21 |
| aarefiev | clarify please | 11:21 |
| dtantsur | the iBFT thing is about telling the OS to attach a remove disk. we don't have control on it. | 11:21 |
| sambetts | aarefiev: in the boot from volume case Ironic will connect disk, in this case the disk is connected out of band by the operator, before being given to Ironic | 11:22 |
| dtantsur | it just happened to didn't work right now, but e.g. our downstream ramdisk did it automatically and we never knew | 11:22 |
| openstackgerrit | Sam Betts proposed openstack/ironic: Fail early if ramdisk type is dib, and not building https://review.openstack.org/328246 | 11:22 |
| aarefiev | sambetts: operator will connect disk? I thought operator should say, hey deploy this node with this disk | 11:23 |
| *** dmk0202 has quit IRC | 11:23 | |
| sambetts | aarefiev: thats a different use case | 11:23 |
| dtantsur | yeah, it's the same result, but totally different approach | 11:23 |
| *** bcornec1 has joined #openstack-ironic | 11:23 | |
| sambetts | aarefiev: this use case, is about a diskless node, that is already configured to connect to an remote volume | 11:23 |
| sambetts | aarefiev: outside of Ironic | 11:24 |
| sambetts | Ironic isn't even aware that the node is diskless/has a remote disk | 11:24 |
| aarefiev | sambetts: how ironic will manage that node in future if it isn't aware about remote disk | 11:25 |
| sambetts | aarefiev: just like it does with a node that has a real local disk | 11:26 |
| *** jcoufal has joined #openstack-ironic | 11:26 | |
| aarefiev | sambetts: but we need somehow to pass iBFT data to target instance | 11:26 |
| aarefiev | via ipxe for example | 11:27 |
| dtantsur | aarefiev, seems like BIOS does it. I was incorrectly assuming that it's passed via kernel command line, seems like it's a separate ACPI table | 11:28 |
| *** daemontool has quit IRC | 11:28 | |
| *** narasimha_SV has joined #openstack-ironic | 11:28 | |
| narasimha_SV | Hi | 11:28 |
| *** daemontool has joined #openstack-ironic | 11:28 | |
| aarefiev | hmm, but ironic should configure it? | 11:29 |
| narasimha_SV | whenever I am trying to boot a baremetal from nova | 11:29 |
| narasimha_SV | i am stuck with nova-scheduler not select the baremetal host created | 11:29 |
| *** ccamacho is now known as ccamacho|lunch | 11:29 | |
| sambetts | aarefiev: no its configured by the operater before being given to Ironic | 11:29 |
| narasimha_SV | Filtering removed all hosts for the request with instance ID 'c8a68435-e50f-4b1d-bce8-c25648fb80eb'. Filter results: ['RetryFilter: (start: 1, end: 1)', 'AvailabilityZoneFilter: (start: 1, end: 1)', 'ComputeFilter: (start: 1, end: 1)', 'ComputeCapabilitiesFilter: (start: 1, end: 0)'] | 11:29 |
| narasimha_SV | 'ComputeCapabilitiesFilter: (start: 1, end: 0)' every time due to this it is going to error state | 11:30 |
| narasimha_SV | can anyone help me in understanding what exactly is this | 11:30 |
| aarefiev | ok, for sure, I need to make more research here | 11:30 |
| snecklifter | ugh, what is the problem here exactly? https://review.openstack.org/#/c/327769/ | 11:30 |
| sambetts | narasimha_SV: does the flavor you've configured exactly match your Ironic node? | 11:31 |
| narasimha_SV | yes | 11:31 |
| narasimha_SV | 1a0c8383-559c-474e-aa8a-72577a4bbfcc | my-baremetal-flavor1 | 1024 | 10 | 0 | | 1 | 1.0 | True | | 11:31 |
| aarefiev | dtantsur, sambetts: lets reach consensus on api versioning | 11:31 |
| narasimha_SV | Final resource view: name=742309b5-24f7-4f1f-a104-543172d3e79e phys_ram=1024MB used_ram=0MB phys_disk=10GB used_disk=0GB total_vcpus=1 used_vcpus=0 pci_stats=[] | 11:32 |
| narasimha_SV | this is the log from compute and above is the flavor which I created | 11:32 |
| narasimha_SV | ironic node-validate is also being successful | 11:33 |
| dtantsur | narasimha_SV, please take a look at extra_specs of the flavor | 11:33 |
| sambetts | narasimha_SV: did you configure capabilites on your ironic node, and your flavor extra specs | 11:33 |
| narasimha_SV | yes I just added cpu_arch as the extra_specs to flavor | 11:33 |
| narasimha_SV | apart from that do I need to add any other ??? | 11:33 |
| aarefiev | dtantsur, sambetts: I think we need bump version and return 200 on previous as Sam suggested | 11:35 |
| dtantsur | aarefiev, I don't see point in it | 11:36 |
| *** dmk0202 has joined #openstack-ironic | 11:36 | |
| aarefiev | dtantsur: if users expected 200 before, we don't break them | 11:37 |
| sambetts | dtantsur: well otherwise we might as well not be bumping, because if someone is pinning to an old API we'll break them anyways | 11:37 |
| dtantsur | aarefiev, well, if users check for success like code == 200, then yeah.. even though it's totally insane. | 11:37 |
| aarefiev | but possible :) | 11:38 |
| dtantsur | I'm just saying that nearly nobody pins ironic-inspector version, so our "fix" won't affect anyone | 11:38 |
| dtantsur | including our own gates btw | 11:38 |
| narasimha_SV | sambetts: yes i have added whatever extra specs of flavor I set to ironic node too | 11:39 |
| narasimha_SV | ironic node-update $NODE add instance_info/cpu_arch=$ARCH instance_info/capabilities='{"boot_option": "local"}' | 11:39 |
| narasimha_SV | properties | capabilities:boot_option='local', cpu_arch='{i686|x86_64}' | 11:39 |
| openstackgerrit | Christopher Brown proposed openstack/ironic-specs: LVM/Software RAID support in ironic-python-agent https://review.openstack.org/327769 | 11:39 |
| narasimha_SV | flavor properties I added | 11:40 |
| dtantsur | narasimha_SV, you should not update instance_info | 11:40 |
| sambetts | narasimha_SV: cpu_arch should be one value, e.g. x86_64 or i686, it can't be both | 11:40 |
| dtantsur | also yeah, it's one value | 11:40 |
| dtantsur | you should update properties/cpu_arch and properties/capabilities | 11:40 |
| narasimha_SV | ok thank you | 11:40 |
| dtantsur | I wonder if we have docs on this... | 11:41 |
| *** baoli has joined #openstack-ironic | 11:41 | |
| sambetts | dtantsur: we do in the install and devstack guides | 11:41 |
| dtantsur | narasimha_SV, we do have docs on it: http://docs.openstack.org/developer/ironic/deploy/install-guide.html#enrollment-process | 11:41 |
| aarefiev | dtantsur: I personally don't pin :), but if somebody does, who knows? so are you against return 200 on <1.6 ? | 11:41 |
| dtantsur | aarefiev, well, not critically against.. I'm just saying that it won't fix our tempest :) | 11:41 |
| sambetts | dtantsur: although everyone seems to think {i686|x86_64} is valid | 11:41 |
| dtantsur | sambetts, everyone? | 11:41 |
| sambetts | dtantsur: we don't pin in the gate so it would fix it right because it would use the latest? | 11:42 |
| *** baoli_ has joined #openstack-ironic | 11:42 | |
| sambetts | dtantsur: might lead to problems in grenade though | 11:43 |
| dtantsur | sambetts, yeah, inspector versioning is different from ironic's. | 11:43 |
| dtantsur | we never had a breaking change before, so this is the first time | 11:43 |
| snecklifter | dtantsur: finally got the LVM/software RAID spec passing :) | 11:44 |
| sambetts | dtantsur: I'm happy to fix it as a bug in across all API versions | 11:44 |
| *** dprince has joined #openstack-ironic | 11:44 | |
| sambetts | snecklifter: was it a problem with the symlink> | 11:44 |
| sambetts | ? | 11:44 |
| dtantsur | snecklifter, cool! | 11:44 |
| snecklifter | sambetts: ye | 11:44 |
| mat128 | good morning ironic! | 11:45 |
| aarefiev | sambetts: so you fine with version bump | 11:45 |
| dtantsur | sambetts, aarefiev, currently our versioning means "be compatible with X.Y", while ironic's is "use strictly X.Y" | 11:45 |
| aarefiev | yeah | 11:45 |
| dtantsur | so by default we're using 1.0 essentially. which does not mean you can't use features from 1.1 at your risk | 11:46 |
| *** baoli has quit IRC | 11:46 | |
| *** wajdi has joined #openstack-ironic | 11:49 | |
| narasimha_SV | is setting properties/capabilities compulsion for both flavor and ironic node??? | 11:50 |
| *** wajdi_ has joined #openstack-ironic | 11:50 | |
| narasimha_SV | I mean if this is not set for both flavor and node will this start booting the baremetal ? | 11:50 |
| mat128 | narasimha_SV: compulsion? | 11:50 |
| sambetts | dtantsur: so inspectors versioning says this is the version I'm running, its not backwards compatible so you need to be compatible with me? | 11:51 |
| narasimha_SV | mat128: i mean do we need to set properties/capabilities for flavor and also node? | 11:51 |
| *** pcaruana has quit IRC | 11:51 | |
| dtantsur | narasimha_SV, if it's not set on ironic, it should not be set on flavor | 11:52 |
| mat128 | narasimha_SV: everything in your flavor's extra specs that starts with capabilities: or that does not contain another "namespace" will be matched to nodes capabilities by the nova capabilities filter | 11:52 |
| dtantsur | sambetts, not sure I got the question. supplying inspector with version 1.1 will mean 'please serve API backward compatible with API 1.1' | 11:52 |
| mat128 | narasimha_SV: memory_mb, root_gb and vcpus are special cases | 11:53 |
| *** wajdi has quit IRC | 11:53 | |
| narasimha_SV | dtantsur: mat128: I didn't set properties/capabilities to flavor as well as node. but now booting is failed with Filtering removed all hosts for the request with instance ID '1c83edae-37f4-4063-86cb-a2e9afd2fbe1'. Filter results: ['RetryFilter: (start: 0, end: 0)'] | 11:53 |
| dtantsur | narasimha_SV, please paste somewhere output of 'nova flavor-show <flavor>' | 11:54 |
| sambetts | dtantsur: right, if thats the case then we need to make inspector return 200 for cases when it supplied with version 1.5 and 204 when supplied with 1.6 | 11:55 |
| dtantsur | sambetts, we can do that (we need to make sure our default version is 1.0, it's not set explicitly now iirc) | 11:55 |
| dtantsur | sambetts, but it won't fix tempest, because tempest does not pin any version (= request 1.0) and pinning will mean we won't test what many people use... | 11:56 |
| narasimha_SV | dtantsur: ]# nova flavor-show 10 +----------------------------+------------------------+ | Property | Value | +----------------------------+------------------------+ | OS-FLV-DISABLED:disabled | False | | OS-FLV-EXT-DATA:ephemeral | 0 | | disk | 10 | | extra_specs | {"cpu_arch": "x86_64"} | 11:56 |
| mat128 | narasimha_SV: paste.openstack.org please | 11:56 |
| sambetts | dtantsur: IMO no pin should == latest not first right? | 11:57 |
| dtantsur | sambetts, that's a question to discuss :) | 11:57 |
| dtantsur | sambetts, well, I like it, except for we'll get users all breaking changes by default | 11:58 |
| dtantsur | maybe it's fine though. people really caring should request an explicit version | 11:58 |
| dtantsur | so yeah, then it's even easier: return 200 if version < 1.6 else 201 | 11:58 |
| sambetts | dtantsur: exactly, although the tempest test for mitaka will still be broken when running against mitaka doesn't have v1.6 (but I'm not sure if we'll be doing that or not) | 11:59 |
| dtantsur | sambetts, I'm not sure if we plan running tempest against mitaka | 12:00 |
| sambetts | dtantsur: we'll need it for mitaka to newton grenade I think | 12:01 |
| dtantsur | sambetts, lets talk about it when we have at least master to master grenade :D | 12:01 |
| sambetts | ;) | 12:01 |
| narasimha_SV | dtantsur: http://paste.openstack.org/raw/510514/ | 12:01 |
| sambetts | narasimha_SV: can you paste an Ironic node show too | 12:02 |
| dtantsur | narasimha_SV, ok. now please 'ironic node-show <NODE>' | 12:02 |
| narasimha_SV | ok | 12:03 |
| dtantsur | :) | 12:03 |
| narasimha_SV | sambetts,dtantsur: http://paste.openstack.org/raw/510515/ | 12:04 |
| narasimha_SV | I am using Vbox driver | 12:04 |
| dtantsur | narasimha_SV, your properties lack cpu_arch | 12:04 |
| sambetts | ironic node-update <node-uuid> add properties/cpu_arch=x86_64 | 12:06 |
| *** pcaruana has joined #openstack-ironic | 12:07 | |
| narasimha_SV | sambetts, dtantsur: ok. as told i have updated the cpu_arch to the node. http://paste.openstack.org/raw/510517/ | 12:07 |
| narasimha_SV | but still ] Filtering removed all hosts for the request with instance ID '7b4e2781-95a3-47f1-8ebd-c3a5662c5d5a'. Filter results: ['RetryFilter: (start: 1, end: 1)', 'AvailabilityZoneFilter: (start: 1, end: 1)', 'ComputeFilter: (start: 1, end: 1)', 'ComputeCapabilitiesFilter: (start: 1, end: 0)'] | 12:08 |
| dtantsur | narasimha_SV, please show output of 'nova hypervisor-stats' | 12:09 |
| narasimha_SV | sambetts, dtantsur: is this anything related to root_gb and local_gb ???? | 12:09 |
| *** SiRiuS__ has quit IRC | 12:09 | |
| dtantsur | I don't think so. lets check the hypervisor stats first | 12:10 |
| narasimha_SV | http://paste.openstack.org/raw/510522/ | 12:10 |
| jroll | grenade \o/ \o\ /o/ \o\ /o/ | 12:10 |
| dtantsur | hmm, everything looks fine.. | 12:10 |
| dtantsur | morning jroll, yeah :) | 12:10 |
| jroll | morning :) | 12:10 |
| sambetts | narasimha_SV: can you paste a full snippet for the n-sch logs | 12:11 |
| * dtantsur rechecks his patch | 12:11 | |
| narasimha_SV | ok | 12:11 |
| sambetts | jroll morning and \o/ for grenade | 12:11 |
| jroll | dtantsur: feel free to finish other conversations, but I see you reviewing the patch here so are we planning to approve this rfe? https://bugs.launchpad.net/ironic-python-agent/+bug/1590606 | 12:12 |
| openstack | Launchpad bug 1590606 in ironic-python-agent "[RFE] Support remote iSCSI shares as root devices via iscsistart" [Wishlist,In progress] - Assigned to Yossi Ovadia (jabadia) | 12:12 |
| dtantsur | jroll, AJaeger was asking if we still need https://review.openstack.org/#/c/312596/ and I think the answer is yes | 12:12 |
| dtantsur | jroll, I was reviewing the patch to actually understand the RFE :) this is a question to discuss. I personally like it. | 12:12 |
| *** pcaruana has quit IRC | 12:13 | |
| sambetts | dtantsur, jroll: I like it, we just need to make sure people understand the difference between this and boot from volume | 12:13 |
| *** pcaruana has joined #openstack-ironic | 12:13 | |
| dtantsur | jroll, I mean, that's a lightweight approach to boot-from-volume, right? we don't care how it gets configured, just make sure we can use it | 12:13 |
| jroll | dtantsur: okay, cool, just double-checking | 12:13 |
| jroll | yep | 12:13 |
| sambetts | its manual bfv | 12:13 |
| sambetts | :-P | 12:14 |
| narasimha_SV | sambetts, dtantsur: thanks guys this time it is able to boot but I think there is some issue with my PXE | 12:14 |
| sambetts | narasimha_SV: :D | 12:14 |
| jroll | dtantsur: yes, we still need ipmitool, I want to cut everything over to that this cycle and think about moving ssh out | 12:14 |
| * jroll leaves a note in gerrit | 12:14 | |
| * sambetts starts chanting "death to ssh, death to ssh" | 12:14 | |
| dtantsur | oh hell yes | 12:15 |
| dtantsur | what a cycle: remove old ramdisk, remove ssh, add grenade! | 12:15 |
| *** ayoung has quit IRC | 12:15 | |
| jroll | and we're just pass n-1 :) | 12:15 |
| dtantsur | yep! | 12:15 |
| dtantsur | speaking of which, I'm looking for volunteers to review the driver composition spec ;) | 12:16 |
| dtantsur | which is https://review.openstack.org/188370 | 12:16 |
| sambetts | jroll: I'm worried about the nova freeze on june 30th for the network stuff | 12:16 |
| dtantsur | sambetts, I think we should start landing stuff finally | 12:16 |
| jroll | sambetts: yeah, we need to move on the networking | 12:16 |
| dtantsur | sambetts, but I would love it to not diverge too much from driver composition, hence my review request | 12:16 |
| dtantsur | jroll, ^^^ | 12:16 |
| sambetts | dtantsur: I keep meaning to do it review it but I keep getting distracted by other stuff | 12:16 |
| jroll | dtantsur: ++ | 12:16 |
| * jroll will try to review driver comp today | 12:17 | |
| jroll | dang, dtantsur beat me to rechecking the bash ramdisk patch | 12:18 |
| dtantsur | hehehe | 12:19 |
| openstackgerrit | Andre Aranha proposed openstack/python-oneviewclient: Extend python-oneviewclient to Uplink Set https://review.openstack.org/295981 | 12:24 |
| *** cdearborn has joined #openstack-ironic | 12:24 | |
| *** dmk0202 has quit IRC | 12:25 | |
| openstackgerrit | Andre Aranha proposed openstack/python-oneviewclient: Extend python-oneviewclient to Uplink Set https://review.openstack.org/295981 | 12:27 |
| *** [1]cdearborn has joined #openstack-ironic | 12:27 | |
| mgould | woo, grenade-dsvm-ironic is now -nv on grenade | 12:32 |
| mgould | so if the grenade folks break our CI, at least they were warned :-) | 12:33 |
| jroll | oh good, sean approved that, he's the grenade king | 12:33 |
| sambetts | \o/ | 12:33 |
| jroll | I was worried about people that actually work on grenade getting ninja'd | 12:33 |
| mgould | ninja'd? | 12:34 |
| jroll | like, others approving it before they get to it | 12:35 |
| dtantsur | mgould, we have to be non-voting there, otherwise every time we get broken, we break essentially them completely | 12:35 |
| dtantsur | mgould, it's unfortunate, but the only thing we can do about non-ironic projects is to run non-voting jobs on them | 12:35 |
| *** Fdaisuke has quit IRC | 12:35 | |
| mgould | jroll: ah, gotcha | 12:36 |
| mgould | dtantsur: that shouldn't happen if we're gating on grenade-dsvm-ironic ourselves, though, right? | 12:37 |
| jroll | mgould: because nothing outside of ironic has ever broken us, right? :) | 12:37 |
| dtantsur | mgould, that=what? we can get broken for many-many reasons, not limited to grenade | 12:37 |
| mgould | oh, right | 12:38 |
| mgould | project X breaks us => grenade-dsvm-ironic fails, through no fault of ours => grenade is also broken | 12:38 |
| dtantsur | correct | 12:38 |
| mgould | :-( | 12:38 |
| jroll | it's okay, we're also non-voting on tempest/devstack/etc | 12:39 |
| dtantsur | some time ago we didn't have a job on neutron (even non-voting like now) and they used to break us every 2 months or so :) | 12:39 |
| jroll | we've been broken before but it isn't often | 12:39 |
| jroll | and the results recording at least helps track down which patch broke it | 12:39 |
| *** dmk0202 has joined #openstack-ironic | 12:39 | |
| dtantsur | yeah, that's the most important feature of even a non-voting gate | 12:39 |
| * mgould currently has a couple of trivial docs-only patches for tripleo that have been failing CI consistently for weeks :-/ | 12:40 | |
| jroll | O_o | 12:40 |
| sambetts | I guess the real trick is we should have voting jobs for all the projects at might affect us so that none of them can merge something that breaks Ironic | 12:40 |
| mgould | holy s**t, this one's passing now: https://review.openstack.org/#/c/321116/ | 12:40 |
| mgould | anyone who's a core on tripleo, please +2 it before it breaks again | 12:41 |
| dtantsur | no tripleo cores here | 12:42 |
| * mgould will ask in #tripleo :-) | 12:42 | |
| dtantsur | but yeah, I'm impressed they keep rechecking the tripleo CI for a docs-only patch | 12:42 |
| mgould | I'm mostly annoyed that it keeps failing | 12:42 |
| *** cdearborn has quit IRC | 12:43 | |
| *** SiRiuS__ has joined #openstack-ironic | 12:46 | |
| *** alaski is now known as lascii | 12:48 | |
| *** ccamacho|lunch is now known as ccamacho | 12:52 | |
| *** links has quit IRC | 12:54 | |
| *** wajdi_ has quit IRC | 12:56 | |
| vdrok | morning jroll | 12:56 |
| jroll | \o vdrok | 12:57 |
| *** dmk0202 has quit IRC | 12:58 | |
| *** dmk0202 has joined #openstack-ironic | 12:59 | |
| openstackgerrit | Moshe Levi proposed openstack/ironic-python-agent: Add vendor id and vendor name to interface information https://review.openstack.org/279940 | 13:00 |
| openstackgerrit | Moshe Levi proposed openstack/ironic-python-agent: tinyipa: Fix git clone tags when 1.6.5 < git < 1.7.10 https://review.openstack.org/327825 | 13:00 |
| *** lindycoder has joined #openstack-ironic | 13:00 | |
| *** ppiela has joined #openstack-ironic | 13:06 | |
| *** cdearborn has joined #openstack-ironic | 13:07 | |
| *** rbudden has joined #openstack-ironic | 13:09 | |
| *** derekh has joined #openstack-ironic | 13:11 | |
| openstackgerrit | Moshe Levi proposed openstack/ironic-python-agent: Add vendor id and vendor name to interface information https://review.openstack.org/279940 | 13:15 |
| openstackgerrit | Moshe Levi proposed openstack/ironic-python-agent: tinyipa: Fix git clone tags when 1.6.5 <= git < 1.7.10 https://review.openstack.org/327825 | 13:15 |
| openstackgerrit | Bruno Cornec proposed openstack/ironic: Fix JSON error in documentation https://review.openstack.org/326079 | 13:15 |
| *** krtaylor has quit IRC | 13:15 | |
| *** amotoki has quit IRC | 13:15 | |
| *** ayoung has joined #openstack-ironic | 13:16 | |
| openstackgerrit | Szymon Borkowski proposed openstack/ironic-python-agent: Add PCI devices collector to inspector https://review.openstack.org/326318 | 13:18 |
| *** jistr is now known as jistr|mtg | 13:20 | |
| *** pcaruana has quit IRC | 13:20 | |
| *** [1]cdearborn has quit IRC | 13:22 | |
| *** dmk0202 has quit IRC | 13:29 | |
| *** krtaylor has joined #openstack-ironic | 13:31 | |
| *** links has joined #openstack-ironic | 13:36 | |
| *** vishwanathj has quit IRC | 13:39 | |
| *** pcaruana has joined #openstack-ironic | 13:39 | |
| aarefiev | dtantsur, sambetts: guys, do you want all api changes in one patch ? | 13:41 |
| dtantsur | is it so big? | 13:41 |
| aarefiev | no | 13:41 |
| aarefiev | bug fix and switch on current api | 13:42 |
| dtantsur | switch on current API? | 13:46 |
| *** joprovost has joined #openstack-ironic | 13:46 | |
| sambetts | dtantsur: default to current instead of minimum | 13:47 |
| dtantsur | sambetts, I think we already default to current | 13:48 |
| sambetts | nope :-P | 13:48 |
| sambetts | _DEFAULT_API_VERSION = _format_version(MINIMUM_API_VERSION) | 13:48 |
| sambetts | thats why unpinned == 1.0 as far as I can tell | 13:49 |
| aarefiev | yes | 13:49 |
| aarefiev | also we need to change client accordingly | 13:50 |
| dtantsur | aarefiev, client, why? | 13:50 |
| aarefiev | dtantsur: you think client should default to min? | 13:51 |
| dtantsur | aarefiev, client defaults to not sending anything, meaning using server default | 13:51 |
| dtantsur | and I'd prefer it continue doing the same | 13:51 |
| sambetts | ++ | 13:51 |
| aarefiev | yeah, I thought we send min by default | 13:52 |
| *** Goneri has joined #openstack-ironic | 13:56 | |
| *** ifarkas has quit IRC | 13:56 | |
| jlvillal | dtantsur, You also reviewed https://review.openstack.org/#/c/322850/ Not sure if you had comments about the environment variable in patchset 3. | 13:58 |
| dtantsur | I don't quite remember and gerrit does not show it :) | 13:59 |
| dtantsur | also morning jlvillal | 13:59 |
| jlvillal | dtantsur, You had done a +2 in a previous version :) | 13:59 |
| jlvillal | dtantsur, Good morning! | 13:59 |
| jlvillal | dtantsur, But yeah it was in May | 14:00 |
| *** ametts has joined #openstack-ironic | 14:00 | |
| *** links has quit IRC | 14:01 | |
| *** joprovost1 has joined #openstack-ironic | 14:02 | |
| jlvillal | Good morning aarefiev sambetts jroll vdrok mgould mat128 vsaienko NobodyCam and everyone else | 14:03 |
| aarefiev | morning jlvillal | 14:03 |
| *** SiRiuS__ has quit IRC | 14:03 | |
| *** joprovost has quit IRC | 14:03 | |
| *** joprovost1 is now known as joprovost | 14:03 | |
| vdrok | good morning jlvillal ! | 14:04 |
| jlvillal | :) | 14:04 |
| mgould | morning jlvillal | 14:05 |
| * mgould decides today is fired and goes home for a nap | 14:05 | |
| mgould | have a good weekend, everyone! | 14:05 |
| sambetts | night mgould | 14:05 |
| jlvillal | mgould, Ciao! | 14:05 |
| *** mgould is now known as mgould|afk | 14:05 | |
| *** sacharya has joined #openstack-ironic | 14:06 | |
| openstackgerrit | Anton Arefiev proposed openstack/ironic-inspector: Fix response code for rule creating API https://review.openstack.org/327492 | 14:06 |
| *** jistr|mtg is now known as jistr | 14:09 | |
| *** sacharya has quit IRC | 14:11 | |
| *** joprovost has quit IRC | 14:23 | |
| *** joprovost has joined #openstack-ironic | 14:24 | |
| *** wajdi has joined #openstack-ironic | 14:25 | |
| *** Jabadia has joined #openstack-ironic | 14:26 | |
| sambetts | mat128: You around? | 14:32 |
| *** ifarkas has joined #openstack-ironic | 14:32 | |
| dtantsur | is anybody tracking the horizon problem? sambetts? | 14:33 |
| *** joprovost1 has joined #openstack-ironic | 14:33 | |
| dtantsur | jroll, https://review.openstack.org/327083 passed grenade, the only failure is transient (horizon failed to start) | 14:33 |
| dtantsur | jroll, I think you can recheck + approve or at least remove -2 :) | 14:34 |
| jroll | dtantsur: there's a bug for the horizon thing, JayF submitted an e-r check but I don't think it's merged yet | 14:34 |
| *** joprovost has quit IRC | 14:35 | |
| *** joprovost1 is now known as joprovost | 14:35 | |
| jroll | +2'd | 14:35 |
| dtantsur | awesome! who else wants to review removing the old ramdisk? -1000 LoC, worth additional eyes :) | 14:35 |
| dtantsur | patch is https://review.openstack.org/327083 | 14:36 |
| dtantsur | jlvillal, maybe ^^ | 14:37 |
| *** SiRiuS__ has joined #openstack-ironic | 14:38 | |
| jlvillal | dtantsur, Getting ready to go to work. Will check it out in the next two hours. Thanks | 14:39 |
| dtantsur | cool | 14:39 |
| jroll | dtantsur: can I call dibs on +A there? :D | 14:39 |
| * jroll is proud of IPA | 14:39 | |
| * jlvillal groans | 14:39 | |
| jlvillal | :) | 14:39 |
| * jroll should promote russell_h to core so he can +A that, lol | 14:40 | |
| jlvillal | jroll, JayF would be proud of you! | 14:40 |
| dtantsur | LOL | 14:40 |
| jlvillal | and your pun | 14:40 |
| jroll | I don't see the pun | 14:40 |
| jlvillal | Ah. I was thinking DIB and dibs | 14:41 |
| jroll | heh, yeah just realized it | 14:41 |
| *** ChrisAusten has joined #openstack-ironic | 14:41 | |
| jroll | maybe if I liked DIB that would be intentional :) | 14:41 |
| *** PollyZ has joined #openstack-ironic | 14:42 | |
| vdrok | jroll: dtantsur a couple more of one-line patches to make partition images job actually do what it should, could you +1 these too? :) https://review.openstack.org/#/q/topic:ipa-partition-images | 14:45 |
| jroll | those are more than one line :P | 14:45 |
| dtantsur | jlvillal, lol, I also was at first like "wtf about DIB"? :D then I googled the expression | 14:46 |
| vdrok | :) | 14:46 |
| jroll | +1 for both | 14:46 |
| *** Goneri has quit IRC | 14:47 | |
| vdrok | thanks | 14:47 |
| dtantsur | it looks pretty controversial to me to set image based on ephemeral size.. do you think we could at least add a comment why it is done? | 14:47 |
| dtantsur | vdrok, ^^ | 14:47 |
| vdrok | dtantsur: yeah, sure, will do now | 14:48 |
| *** jvgrant_ has joined #openstack-ironic | 14:54 | |
| *** mtanino has joined #openstack-ironic | 14:55 | |
| NobodyCam | Good morning Ironicers | 14:55 |
| NobodyCam | morning jlvillal | 14:55 |
| vdrok | morning NobodyCam | 14:56 |
| NobodyCam | and happy Friday from the zombie behind the keyboard | 14:56 |
| NobodyCam | morning vdrok :) | 14:56 |
| sambetts | o/ NobodyCam TGIF | 14:56 |
| NobodyCam | hey sambetts and for sure TGIF :) | 14:57 |
| *** daemontool has quit IRC | 14:57 | |
| dtantsur | morning NobodyCam, TGIF!! | 14:57 |
| dtantsur | NobodyCam, happy grenade also :) | 14:57 |
| NobodyCam | morning dtantsur and w00t :) (zombieHappyDance) | 14:58 |
| dtantsur | morning JayF | 15:00 |
| JayF | hm | 15:00 |
| JayF | gm | 15:00 |
| jroll | heh, "good morning" "hmmmmmmmmmmmMM" | 15:00 |
| openstackgerrit | Yossi Ovadia proposed openstack/ironic-python-agent: Handle diskless hardware connected to remote iscsi. https://review.openstack.org/327807 | 15:01 |
| NobodyCam | hehehe Morning JayF jroll :) | 15:01 |
| JayF | gm | 15:01 |
| jroll | hey NobodyCam :) | 15:01 |
| vdrok | morning JayF ! | 15:01 |
| NobodyCam | :) | 15:03 |
| *** Jabadia_ has joined #openstack-ironic | 15:03 | |
| JayF | sambetts: any idea why that project-config change didn't pass tests? I thought I did define those jobs... | 15:04 |
| *** Jabadia has quit IRC | 15:04 | |
| *** jaybeale has joined #openstack-ironic | 15:04 | |
| dtantsur | damn, it seems like the only option to get rid of disk_available_least=-160 is to delete nova database... >_< | 15:04 |
| *** spandhe has joined #openstack-ironic | 15:04 | |
| jroll | JayF: there's another piece you need e.g. https://review.openstack.org/#/c/314966/1/jenkins/jobs/projects.yaml | 15:05 |
| sambetts | JayF: You added the templates but not the jobs themselves | 15:06 |
| JayF | hah | 15:06 |
| dtantsur | JayF, jroll, and everyone: lets stop using pxe_ipa in job names, now that we're removing the old ramdisk and its jobs are no longer there | 15:06 |
| JayF | dtantsur: -0.5 | 15:06 |
| sambetts | \o/ | 15:06 |
| dtantsur | JayF, mmm? | 15:07 |
| jroll | let's change them all to pxe_ipmitool ;) | 15:07 |
| *** spandhe_ has joined #openstack-ironic | 15:07 | |
| JayF | dtantsur: I don't like changing job names unless there's a really compelling reason, b/c it basically invalidates any statistics about the job | 15:07 |
| dtantsur | JayF, I rather mean lets stop creating new ones with such names | 15:07 |
| Jabadia_ | OK, time has come for me to write unitest. Where can I learn mocks quickly ? | 15:07 |
| JayF | dtantsur: and these exp jobs are 100% just to test different settings on the existing, so I want them to be parallel to existing names | 15:07 |
| dtantsur | Jabadia_, their official documentation is not so bad: http://www.voidspace.org.uk/python/mock/ | 15:08 |
| Jabadia_ | tnx dtantsur | 15:08 |
| dtantsur | JayF, well, if you're fine with job names that don't make any sense to anyone who does not know the background... | 15:08 |
| dtantsur | and yeah, ++ for ipmitool somewhere soon | 15:09 |
| *** spandhe has quit IRC | 15:09 | |
| *** spandhe_ is now known as spandhe | 15:09 | |
| jroll | everywhere | 15:09 |
| TheJulia | yes yes | 15:09 |
| *** tesseract has quit IRC | 15:09 | |
| jroll | I'd like to move all of our CI to ipmitool before n-2 | 15:09 |
| *** piet has joined #openstack-ironic | 15:09 | |
| dtantsur | jroll, lucas is on pto, so I'll rebase https://review.openstack.org/#/c/312596/ wdyt? | 15:10 |
| jroll | dtantsur: ++ | 15:10 |
| * jroll steps away for a few | 15:10 | |
| dtantsur | if we get it merged, we should switch to ipmitool-only pretty soon | 15:12 |
| dtantsur | otherwise we'll have too many voting jobs :) | 15:12 |
| *** pcaruana has quit IRC | 15:13 | |
| dtantsur | just for clarity: we probably need to replace pxe_ipa and agent_ssh jobs on Mitaka and Newton with pxe_ipmitool and agent_ipmitool, right? | 15:17 |
| jroll | yeah, I think so? | 15:19 |
| jroll | maybe just newton | 15:19 |
| jroll | we'll want to keep testing ssh driver on mitaka | 15:19 |
| *** rcernin has quit IRC | 15:19 | |
| dtantsur | jroll, but run these ipmitool on mitaka too, right? | 15:22 |
| jroll | dtantsur: yep | 15:22 |
| JayF | I need to run another of those giant scale tests on those jobs | 15:23 |
| JayF | since lucas-afk fixed the bugs we found last time doing | 15:23 |
| JayF | *doing it | 15:23 |
| * jroll actually steps away now | 15:23 | |
| *** klindgren has joined #openstack-ironic | 15:24 | |
| klindgren | Under libierty has anyone ran into an issue that when ironic asks neutron to create a port that it doesn't add the pxe bits? Keep running into this where the netutron port is missing the pxe information since we upgraded to liberty. | 15:25 |
| *** sacharya has joined #openstack-ironic | 15:25 | |
| *** Goneri has joined #openstack-ironic | 15:29 | |
| sambetts | klindgren: are you seeing any error in the Ironic logs related to it failed to do it/ | 15:30 |
| sambetts | ? | 15:30 |
| klindgren | nope - but if i remove/red-add the dhcp agent the pxe bits get added | 15:31 |
| sambetts | klindgren: that sounds like a bug in neutron liberty, it sounds like Ironic is setting the port information correctly, but the dhcp agent isn't picking it up | 15:32 |
| klindgren | yea - its weird. Going to debug it some more | 15:32 |
| *** Jabadia_ has quit IRC | 15:32 | |
| sambetts | klindgren: you can check if everything in Ironic is working correctly by checking the neutron ports and looking at dhcp_extra_ops, if everything is there correctly then its a issues on the neutron side | 15:33 |
| klindgren | kk - thanks. I assume its on the neutron side since it gets re-added when I remove/readd the dhcp mapping | 15:34 |
| *** jvgrant_ has quit IRC | 15:41 | |
| *** SiRiuS__ has quit IRC | 15:42 | |
| *** jvgrant_ has joined #openstack-ironic | 15:43 | |
| *** jvgrant_ has quit IRC | 15:46 | |
| dtantsur | please review carefully: https://review.openstack.org/#/c/312596/ | 15:50 |
| openstackgerrit | Xavier proposed openstack/ironic: Add Dynamic Allocation feature for the OneView drivers https://review.openstack.org/286192 | 15:54 |
| *** ametts has quit IRC | 15:55 | |
| *** SiRiuS__ has joined #openstack-ironic | 15:55 | |
| jlvillal | FYI: http://www.openstack.org/blog/2016/05/faq-evolving-the-openstack-design-summit/ If people hadn't seen this | 15:58 |
| *** PollyZ has quit IRC | 16:01 | |
| *** SiRiuS__ has quit IRC | 16:01 | |
| openstackgerrit | sean mooney proposed openstack/bifrost: split ironic install role into install,bootstrap,start phases https://review.openstack.org/318423 | 16:03 |
| *** PollyZ has joined #openstack-ironic | 16:06 | |
| openstackgerrit | sean mooney proposed openstack/bifrost: split ironic install role into install,bootstrap,start phases https://review.openstack.org/318423 | 16:06 |
| *** ametts has joined #openstack-ironic | 16:08 | |
| *** jvgrant_ has joined #openstack-ironic | 16:08 | |
| *** ifarkas has quit IRC | 16:10 | |
| *** PollyZ has quit IRC | 16:11 | |
| *** jvgrant_ has quit IRC | 16:12 | |
| *** rama_y has joined #openstack-ironic | 16:13 | |
| jlvillal | dtantsur: jroll: As a note, I noticed three jobs in project-config that have: deploy-with-ipa: 0 | 16:17 |
| jlvillal | Likely they aren't used, but just defined | 16:17 |
| dtantsur | jlvillal, deploy-with-ipa is not used with agent_ssh drivers iirc | 16:18 |
| *** Romanenko_K has quit IRC | 16:19 | |
| *** kromanenko_ has quit IRC | 16:19 | |
| *** kromanenko__ has quit IRC | 16:19 | |
| jlvillal | dtantsur: Thanks. The one job I checked quickly didn't appear to be used anymore. I'll try to look through more closely and clean-up unused things. Add it to my TODO list | 16:20 |
| *** amotoki has joined #openstack-ironic | 16:21 | |
| *** athomas has quit IRC | 16:24 | |
| *** piet has quit IRC | 16:28 | |
| *** Jabadia has joined #openstack-ironic | 16:32 | |
| *** PollyZ has joined #openstack-ironic | 16:33 | |
| *** ayoung has quit IRC | 16:33 | |
| *** vishwanathj has joined #openstack-ironic | 16:34 | |
| *** ccamacho is now known as ccamacho|away | 16:34 | |
| *** PollyZ has quit IRC | 16:37 | |
| openstackgerrit | Moshe Levi proposed openstack/ironic-python-agent: Add vendor id and vendor name to interface information https://review.openstack.org/279940 | 16:41 |
| openstackgerrit | Moshe Levi proposed openstack/ironic-python-agent: tinyipa: Fix git clone tags when 1.6.5 <= git < 1.7.10 https://review.openstack.org/327825 | 16:41 |
| *** e0ne has quit IRC | 16:41 | |
| *** spandhe has quit IRC | 16:44 | |
| openstackgerrit | Mario Villaplana proposed openstack/ironic: Add power state change notifications https://review.openstack.org/321865 | 16:45 |
| openstackgerrit | Mario Villaplana proposed openstack/ironic: Add notification base classes and docs https://review.openstack.org/298461 | 16:45 |
| *** cdearborn has quit IRC | 16:48 | |
| *** sdake has joined #openstack-ironic | 16:50 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic-python-agent: Add basic troubleshooting doc https://review.openstack.org/326743 | 16:51 |
| dtantsur | g'night and have a great weekend | 16:52 |
| *** dtantsur is now known as dtantsur|afk | 16:52 | |
| NobodyCam | night dtantsur|afk have a great weekend | 16:53 |
| JayF | nn | 16:53 |
| dtantsur|afk | jroll, I suspect 4x +2 might be enough for https://review.openstack.org/#/c/327083/ ;) I'm afraid we merge something conflicting.. | 16:53 |
| * dtantsur|afk is now afk for real | 16:53 | |
| *** amotoki has quit IRC | 16:56 | |
| JayF | mariojv: the oslo config stuff has changed | 16:56 |
| JayF | mariojv: I think you need an entrypoint or something like that for the configs to get pulled in | 16:56 |
| JayF | mariojv: look in alineb's ironic-lib patch for an example | 16:57 |
| *** PollyZ has joined #openstack-ironic | 16:59 | |
| mariojv | oh, totally works, thanks JayF | 16:59 |
| mariojv | had to specify an addition in the default opts list in ironic/conf/opts.py | 17:00 |
| jlvillal | mariojv: Are you adding to an existing section. Thinks 95% chance the answer is yes. | 17:01 |
| jlvillal | mariojv: https://review.openstack.org/295946 This is sort of the future. | 17:01 |
| mariojv | jlvillal: yeah, i'm adding an option for min notification priority to be sent | 17:02 |
| jlvillal | mariojv: Okay :) I was thinking if there was a small chance you were creating a new section, you could see what that patch is doing. | 17:02 |
| mariojv | i registered it in common/rpc.py but it's used in 1 other place too | 17:02 |
| openstackgerrit | Mario Villaplana proposed openstack/ironic: Add power state change notifications https://review.openstack.org/321865 | 17:05 |
| openstackgerrit | Mario Villaplana proposed openstack/ironic: Add notification base classes and docs https://review.openstack.org/298461 | 17:05 |
| JayF | mariojv: glad to help :D | 17:05 |
| *** ijw has joined #openstack-ironic | 17:05 | |
| *** ijw has quit IRC | 17:05 | |
| *** ijw has joined #openstack-ironic | 17:05 | |
| *** jaybeale has quit IRC | 17:06 | |
| *** milan has quit IRC | 17:09 | |
| vdrok | have a good weekend everyone! | 17:13 |
| *** derekh has quit IRC | 17:14 | |
| *** ayoung has joined #openstack-ironic | 17:14 | |
| jlvillal | vdrok: Dos ve dos! | 17:15 |
| *** ohamada has quit IRC | 17:16 | |
| Jabadia | I have a very basic method. execute iscsistart -f , if no err, execute iscsi -b . | 17:18 |
| Jabadia | Trying write test as follow - "mocked_execute.assert_called_with('iscsistart', '-f')" | 17:18 |
| Jabadia | running the test, and getting : | 17:19 |
| Jabadia | AssertionError: Expected call: execute('iscsistart', '-f') , Actual call: execute('iscsistart', '-b') | 17:19 |
| *** cdearborn has joined #openstack-ironic | 17:19 | |
| *** narasimha_SV has quit IRC | 17:21 | |
| JayF | It's telling you exactly what's happening right there | 17:21 |
| JayF | either your test expects the wrong thing, or the code is doing the wrong thing | 17:21 |
| *** sdake has quit IRC | 17:22 | |
| Jabadia | http://pastebin.com/6HKimCDZ | 17:23 |
| Jabadia | I want to test - -f first , -b second | 17:23 |
| Jabadia | ( see code in pastebin .. ) | 17:24 |
| jlvillal | Jabadia: There is a way to check that you are called multiple times with specific arguments. | 17:26 |
| jlvillal | Should be in the mock documentation. | 17:26 |
| Jabadia | hmmm , checking. tnx. | 17:27 |
| jlvillal | Jabadia: assert_has_calls | 17:29 |
| jlvillal | Jabadia: https://docs.python.org/3/library/unittest.mock.html | 17:29 |
| *** e0ne has joined #openstack-ironic | 17:30 | |
| Jabadia | awesome , checking. | 17:30 |
| jlvillal | Jabadia: Probably something like. blah.assert_has_calls([mock.call('iscsistart', '-f'), mock.call('iscsistart', '-b')]) | 17:30 |
| Jabadia | Running now , you guys are great | 17:31 |
| jlvillal | Jabadia: Thanks and good luck :) | 17:31 |
| openstackgerrit | Chris Krelle proposed openstack/ironic-python-agent: Add basic troubleshooting doc https://review.openstack.org/326743 | 17:32 |
| *** dprince has quit IRC | 17:34 | |
| *** spandhe has joined #openstack-ironic | 17:34 | |
| JayF | aha | 17:34 |
| * JayF misunderstood the question | 17:34 | |
| *** dprince has joined #openstack-ironic | 17:35 | |
| JayF | NobodyCam: going to -1 that doc again, /me feels bad | 17:37 |
| JayF | NobodyCam: one other question: ``sudo apt-get install docker docker.io`` | 17:37 |
| NobodyCam | JayF: not at all | 17:37 |
| JayF | NobodyCam: ^ that seems like it should be one or the other | 17:37 |
| JayF | NobodyCam: you could maybe just point them to https://github.com/openstack/ironic-python-agent/blob/master/imagebuild/coreos/full_trusty_build.sh ? | 17:38 |
| JayF | NobodyCam: I guess that's fairly gate specific, but would be a good template | 17:38 |
| JayF | NobodyCam: actually, you could just make that line "install docker container engine" and leave the command as an exercise to the reader | 17:38 |
| JayF | to prevent having to if-statement it across all distros | 17:39 |
| NobodyCam | ahh that was reused from https://github.com/openstack/bifrost/blob/master/troubleshooting.rst#building-an-ipa-image | 17:39 |
| JayF | ah | 17:39 |
| NobodyCam | happy to update :) | 17:39 |
| JayF | yeah, the way it sits is def wrong | 17:39 |
| cinerama | great artists steal :) | 17:39 |
| Jabadia | Works like a charm. thanks jlvillal&JayF ! | 17:39 |
| JayF | but if I were you I'd totally just punt on the "what packages to install" question | 17:39 |
| NobodyCam | :) | 17:40 |
| NobodyCam | let me grab a fresh cup of joe and I'll rework that section | 17:41 |
| JayF | I always feel bad -1'ing docs like that | 17:41 |
| JayF | when clearly it's much better than what exists | 17:41 |
| NobodyCam | not at all | 17:41 |
| jlvillal | Jabadia: :) | 17:42 |
| NobodyCam | doing it because if what we hit earlier in the week | 17:42 |
| NobodyCam | and it should be correct :p | 17:42 |
| jroll | dtantsur|afk: +A'd \o/ | 17:42 |
| *** mtanino has quit IRC | 17:44 | |
| openstackgerrit | Yossi Ovadia proposed openstack/ironic-python-agent: Handle diskless hardware connected to remote iscsi. https://review.openstack.org/327807 | 17:46 |
| *** klindgren has quit IRC | 17:46 | |
| *** e0ne has quit IRC | 17:50 | |
| *** vishwanathj has quit IRC | 17:51 | |
| *** sabeen has joined #openstack-ironic | 17:51 | |
| *** klindgren has joined #openstack-ironic | 17:51 | |
| *** Jabadia has quit IRC | 17:54 | |
| *** e0ne has joined #openstack-ironic | 17:54 | |
| *** Jabadia has joined #openstack-ironic | 17:55 | |
| *** mbound has quit IRC | 17:55 | |
| *** Sukhdev has joined #openstack-ironic | 17:56 | |
| *** catintheroof has joined #openstack-ironic | 17:59 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic-python-agent: Add basic troubleshooting doc https://review.openstack.org/326743 | 17:59 |
| NobodyCam | JayF: added a ref to the docker install doc page :) | 18:00 |
| JayF | perfect | 18:00 |
| JayF | +2 | 18:03 |
| NobodyCam | :) | 18:03 |
| JayF | seriously, ty for that | 18:03 |
| JayF | something I should've written a long time ago | 18:03 |
| JayF | sambetts: https://review.openstack.org/326743 when this lands, would love to have some info about tinyipa added to it as well | 18:03 |
| * JayF wonders if anyone is actually using tinyipa ramdisks in production, or if it's just used in the gate | 18:03 | |
| JayF | So question for the community: | 18:04 |
| TheJulia | JayF: better question, do you want people to use it | 18:04 |
| JayF | TheJulia: sure, why not? | 18:04 |
| JayF | TheJulia: it's gated. We know it works. It's their choice how it integrates into the environment | 18:04 |
| JayF | k back to my question: I was going to put together a small handful of "example" hardware managers | 18:05 |
| TheJulia | drivers, so as long as they can identify if it works for them or not, I don't see why not | 18:05 |
| JayF | just structured in a few different ways to show how you could use multiple hardware managers for cool things | 18:05 |
| TheJulia | That would be totally awesome | 18:05 |
| JayF | I was going to just, at least short term, drop them in a github repo | 18:05 |
| TheJulia | Your using past tense, what are you thining now? | 18:06 |
| JayF | but is it significant enough to go in openstack/ ? should there be a dir added in ipa, etc | 18:06 |
| JayF | I'm just trying to shortcut the "this is great but why did you put it in github!?!" conversation | 18:06 |
| JayF | lol | 18:06 |
| TheJulia | I kind of like the idea of adding a directory to IPA | 18:06 |
| cinerama | i personally think that'd be useful to have in ipa itself | 18:06 |
| jroll | JayF: it would be cool to write useful ones that we could just have in tree :) | 18:06 |
| *** mtanino has joined #openstack-ironic | 18:06 | |
| TheJulia | jroll: ++++++++ * 10000 | 18:06 |
| JayF | jroll: well, I'm thinking more skeleton ones | 18:06 |
| JayF | jroll: Taking a bit more of a bite sized approach, lol | 18:07 |
| jroll | heh, yeah | 18:07 |
| JayF | like an example for "adding support for a single piece of hardware", "adding support for verification of hardware attributes", "subclassing GHM and writing a monolithic manager" | 18:07 |
| JayF | stuff like that | 18:07 |
| jroll | JayF: one easy example is one that makes sure actual ram matches node[properties][memory_mb] | 18:07 |
| JayF | that's probably what I'm going to spike on for the rest of the day/weekend, and just was curious where to put it | 18:07 |
| JayF | jroll: exactly | 18:07 |
| jroll | but I'd rather have it on and working | 18:08 |
| jroll | rather than as an example | 18:08 |
| JayF | Sure; I'd rather have that too, but that's a significantly larger undertaking | 18:08 |
| TheJulia | I say make an example, I'm sure we can adapt one of the examples for something fairly quickly once we see the options | 18:08 |
| JayF | than a documented example | 18:08 |
| TheJulia | well, some examples | 18:09 |
| JayF | I almost feel like a different repo would be a good place to put it though | 18:09 |
| JayF | simply because someone could fork, edit, deploy | 18:10 |
| jroll | yeah, starting with examples is fine | 18:10 |
| jroll | that's also a good point | 18:10 |
| JayF | I guess I'll just do it in github to avoid the 'paperwork' for now, and later we can import it in under ironic governance if folks find it useful | 18:11 |
| JayF | I'll still PR everything in and link the PRs here so someone other than me has seen it :) | 18:11 |
| *** sacharya has quit IRC | 18:13 | |
| *** adreznec has quit IRC | 18:15 | |
| jroll | sounds good | 18:15 |
| jroll | moving a project is easy anyway | 18:15 |
| jlvillal | JayF: Did you ever figure out the logstash search command for the 'apache restart' failure? I think I remember something about it working, then not working. I had never seemed to get it to work. | 18:15 |
| JayF | jlvillal: no | 18:15 |
| JayF | jlvillal: I shifted my attention to other things | 18:15 |
| jlvillal | JayF: Understandable. Thanks | 18:16 |
| *** adreznec has joined #openstack-ironic | 18:17 | |
| *** ChubYann has joined #openstack-ironic | 18:19 | |
| *** e0ne has quit IRC | 18:22 | |
| *** garthb has joined #openstack-ironic | 18:22 | |
| *** garthb_ has joined #openstack-ironic | 18:23 | |
| *** e0ne has joined #openstack-ironic | 18:23 | |
| *** garthb_ has quit IRC | 18:23 | |
| *** garthb has quit IRC | 18:23 | |
| *** garthb has joined #openstack-ironic | 18:24 | |
| *** garthb_ has joined #openstack-ironic | 18:25 | |
| *** sacharya has joined #openstack-ironic | 18:31 | |
| JayF | Chris's troubleshooting doc (https://review.openstack.org/326743) is a pretty quick review if someone else core on IPA wants to have a look and land it | 18:33 |
| openstackgerrit | Thiago Paiva Brito proposed openstack/ironic: [WIP] Using Oneview drivers w/ IRONIC_IS_HARDWARE https://review.openstack.org/328451 | 18:36 |
| jroll | JayF: TIL ssh -l O_o | 18:36 |
| jroll | does that mean I'm young? | 18:36 |
| *** praneshp has joined #openstack-ironic | 18:36 | |
| *** garthb_ has quit IRC | 18:36 | |
| jroll | also this doc makes me feel like we should have per-node append_params | 18:38 |
| jroll | (also the thread kris started that talks about different console devices per node) | 18:39 |
| JayF | jroll: ++ absolutely | 18:39 |
| JayF | jroll: honestly it's a psuedo-requirement for onmetal to change to dynamic dhcp | 18:39 |
| jroll | is it? | 18:39 |
| jroll | why's that? | 18:40 |
| JayF | We mix hardware in a given cell. | 18:40 |
| jroll | oh, that | 18:40 |
| *** praneshp_ has joined #openstack-ironic | 18:40 | |
| *** praneshp has quit IRC | 18:41 | |
| *** praneshp_ is now known as praneshp | 18:41 | |
| *** jcoufal has quit IRC | 18:46 | |
| *** jcoufal has joined #openstack-ironic | 18:48 | |
| *** vishwanathj has joined #openstack-ironic | 18:48 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic-python-agent: Add basic troubleshooting doc https://review.openstack.org/326743 | 18:51 |
| jroll | fixed the tiny nits and approved | 18:51 |
| jroll | NobodyCam: a couple comments here, if you want to do a follow up patch, if not it's cool too :) https://review.openstack.org/#/c/326743/7/doc/source/troubleshooting.rst | 18:52 |
| *** vishwanathj has quit IRC | 18:53 | |
| *** yolanda has quit IRC | 18:53 | |
| *** mbound has joined #openstack-ironic | 18:56 | |
| *** e0ne has quit IRC | 18:59 | |
| *** fragatina has quit IRC | 18:59 | |
| jroll | dtantsur|afk: I see you | 18:59 |
| jroll | ಠ_ಠ| 18:59 |
| jroll | it's friday night, go have fun | 18:59 |
| *** spandhe has quit IRC | 18:59 | |
| *** fragatina has joined #openstack-ironic | 18:59 | |
| *** mbound has quit IRC | 19:01 | |
| *** yolanda has joined #openstack-ironic | 19:03 | |
| *** fragatina has quit IRC | 19:04 | |
| *** Jabadia has quit IRC | 19:05 | |
| *** athomas has joined #openstack-ironic | 19:06 | |
| * NobodyCam takes a look | 19:08 | |
| NobodyCam | I can toss up a quick rev for that ;) | 19:10 |
| jroll | awesome, thanks! | 19:11 |
| sambetts | Night all, JayF: I'll check out that troubleshooting doc stuff on Monday :D | 19:13 |
| *** sambetts is now known as sambetts|afk | 19:13 | |
| NobodyCam | night sambetts|afk | 19:19 |
| *** rama_y has quit IRC | 19:19 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic-python-agent: Add basic troubleshooting doc https://review.openstack.org/326743 | 19:26 |
| NobodyCam | JayF: jroll ^^^ | 19:26 |
| jroll | nice, thanks | 19:26 |
| NobodyCam | :) | 19:28 |
| * NobodyCam makes a quick run to subway for some lunch stuff | 19:28 | |
| *** PollyZ has quit IRC | 19:28 | |
| * jroll +2 | 19:28 | |
| *** PollyZ has joined #openstack-ironic | 19:30 | |
| *** MattMan has quit IRC | 19:31 | |
| *** MattMan has joined #openstack-ironic | 19:31 | |
| *** mbound has joined #openstack-ironic | 19:34 | |
| *** PollyZ has quit IRC | 19:41 | |
| *** Sukhdev has quit IRC | 19:45 | |
| *** vishwanathj has joined #openstack-ironic | 19:49 | |
| *** sdake has joined #openstack-ironic | 19:52 | |
| *** jcoufal has quit IRC | 19:54 | |
| *** vishwanathj has quit IRC | 19:55 | |
| *** PollyZ has joined #openstack-ironic | 19:55 | |
| openstackgerrit | Merged openstack/ironic-python-agent: Add basic troubleshooting doc https://review.openstack.org/326743 | 19:56 |
| *** ametts has quit IRC | 20:00 | |
| openstackgerrit | Mathieu Mitchell proposed openstack/ironic: Support for APC AP7922 https://review.openstack.org/328476 | 20:04 |
| *** jaybeale has joined #openstack-ironic | 20:05 | |
| *** baoli_ has quit IRC | 20:06 | |
| JayF | mat128: ^ you're aware that driver is likely going to be pulled after this cycle due to lack of CI (see: https://wiki.openstack.org/wiki/Ironic/Drivers) | 20:06 |
| mat128 | hmm | 20:07 |
| mat128 | it's our fallback when machien dont have IPMI | 20:07 |
| mat128 | do a virtual PDU works? :) | 20:07 |
| mat128 | we already have that, I guess I could split that into it's own project and wire it up like vBMC | 20:07 |
| JayF | mat128: yeah; or you could just move it over to ironic-staging-drivers and use it | 20:08 |
| mat128 | given we already have everything needed (we actually test it downstream), I guess I can just make that into a small project | 20:08 |
| JayF | mat128: if you were able to get a virtual pdu running, I presume we could run it in the gate, but there might need to be a conversation about ci resources vs popularity of the driver | 20:08 |
| mat128 | JayF: I guess that means any unpopular combo is excluded from being in the official tree? | 20:09 |
| mat128 | JayF: CI resource (or lack of) will always be an argument | 20:09 |
| JayF | I mean, I don't know. | 20:09 |
| jroll | I don't think it's a waste of resources to run one job for that driver | 20:09 |
| JayF | sounds good to me :) | 20:10 |
| mat128 | ok good | 20:10 |
| mat128 | I'll create vPDU and hook it to devstack so we can run it | 20:10 |
| mat128 | and avoid that driver from getting moved into -staging :P | 20:10 |
| mat128 | plus that will be one less thing to worry about downstream | 20:10 |
| jroll | ++ | 20:10 |
| JayF | So I have what feels like a super dumb question | 20:12 |
| JayF | in the examples of new hardware managers I've seen downstream | 20:13 |
| jroll | I have what feels like a super dumb answer :D | 20:13 |
| JayF | some of them implement it in name_of_manager/__init__.py | 20:13 |
| JayF | some of them have a blank init, and put the actual code in a named file | 20:13 |
| JayF | is there a convention or tradeoff to going one way or another? | 20:13 |
| jroll | so, for small modules it's fine to just put everything in __init__ | 20:13 |
| jroll | for larger split it out | 20:13 |
| jroll | the practical difference is: | 20:13 |
| jroll | import name_of_manager, vs from name_of_manager import foo | 20:14 |
| openstackgerrit | Julia Kreger proposed openstack/ironic-specs: Boot from Volume - Reference Drivers https://review.openstack.org/294995 | 20:14 |
| JayF | for these small sample ones then, likely just doing everything in __init__ is sensible then | 20:14 |
| * JayF didn't want to document some "wrong way" to do things | 20:14 | |
| * jlvillal prefers it to have much stuff in __init__.py, because he usually thinks it will be empty. Which is not a good reason :) | 20:14 | |
| jroll | JayF: actually, now that I think about it, maybe convention is if nothing imports it then put it in __init__ | 20:14 |
| * jlvillal prefers it to NOT have much stuff in __init__.py, because he usually thinks it will be empty. Which is not a good reason :) | 20:14 | |
| JayF | in this case, it wouldn't really be imported | 20:15 |
| *** ametts has joined #openstack-ironic | 20:15 | |
| JayF | it'd be given an entrypoint in setup.cfg | 20:15 |
| TheJulia | jlvillal: your first statement had me slightly confused ;) | 20:15 |
| devananda | JayF: just checking - any progress on policy docs / specs? | 20:15 |
| JayF | which means also no advantage to having it in __init__ either | 20:15 |
| mat128 | I'm with jlvillal, always surprised when I see code in __init__ (esp. if theres already code in files around) | 20:15 |
| jlvillal | TheJulia: me too | 20:15 |
| *** thrash is now known as thrash|wknd | 20:15 | |
| openstackgerrit | Yossi Ovadia proposed openstack/ironic-python-agent: Handle diskless hardware connected to remote iscsi. https://review.openstack.org/327807 | 20:15 |
| JayF | devananda: I have a very draft spec up right now, trying to work through some other low hanging fruit on my todo list today | 20:16 |
| devananda | jroll: openstack follows the convention of not importing modules directly, so we should always be doing "from foo import bar", even in IPA | 20:16 |
| jroll | devananda: no, openstack follows the convention of not importing symbols directly from modules | 20:16 |
| jroll | foo.py is a module, convention is import foo | 20:16 |
| devananda | jroll: that is better words | 20:16 |
| jroll | so __init__ is okay, afaict | 20:17 |
| JayF | either way, sounds like in this case it's a no brainer to go not in __init__ | 20:17 |
| JayF | given it wouldn't be imported at all anyway, would only be an entrypoint | 20:17 |
| JayF | right? | 20:17 |
| jroll | sure :) | 20:18 |
| *** Jabadia has joined #openstack-ironic | 20:18 | |
| Jabadia | OSPd RH question- I can extract the agent.kernel using gzip -dc .... | cpio -id. | 20:25 |
| Jabadia | I need to do the same for the overcloud image | 20:25 |
| Jabadia | I tried to openstack-save overcloud-full-vmlinuz , but trying to extract it gives me 'overcloud-full-vmlinuz: not in gzip format' | 20:26 |
| TheJulia | no clue about OSP, but the kernel is not like a ramdisk file | 20:27 |
| * jroll thinks most of the red hatters are gone for the weekend | 20:27 | |
| jroll | #tripleo might have some people that know, though | 20:27 |
| Jabadia | tnx, tryng there. | 20:28 |
| *** ametts has quit IRC | 20:30 | |
| *** dprince has quit IRC | 20:30 | |
| *** dprince has joined #openstack-ironic | 20:31 | |
| klindgren | when ironic creates the port in neutron is that done as a single action or a series of actions? | 20:32 |
| klindgren | IE create port, add dhcp_opts? | 20:33 |
| jroll | the dhcp opts are a port-update | 20:33 |
| jroll | nova actually creates the port | 20:33 |
| jroll | iroinc adds the dhcp stuff | 20:33 |
| klindgren | hrm... have an issue where the dhcp-opts never get configured in the dnsmasq config files | 20:35 |
| klindgren | after we upgraded to liberty - trying to look throug hthe neutron change logs re: dhcp to see wtf | 20:35 |
| mat128 | Jabadia: I think there's a review in flight for working with the IPA image | 20:35 |
| klindgren | if I remove/re-add the dhcp agent to the network, it gets created correctly | 20:36 |
| Jabadia | tnx mat128 | 20:36 |
| mat128 | Jabadia: trying to find it | 20:36 |
| jroll | klindgren: right, saw that earlier, still sounds like a neutron bug to me | 20:36 |
| mat128 | Jabadia: it landed, here's the rendered docs: http://docs.openstack.org/developer/ironic/deploy/troubleshooting.html#patching-the-deploy-ramdisk | 20:38 |
| klindgren | 100% agree. Just trying to track down the entry path to neutron to narrow down whats going on. Added some debug statements around dhcp writing to the file and it seems like when it processes the port.update.end it doesn't see that it has dhcp_opts | 20:38 |
| Jabadia | mat128: that's awesome, thanks! | 20:39 |
| *** fragatina has joined #openstack-ironic | 20:39 | |
| Jabadia | actually, that's only for the introspective image. that works for me. I need to do the same for the overcloud image | 20:40 |
| Jabadia | :( | 20:40 |
| mat128 | Jabadia: glad it helped you, make sure you thank dtantsur :) | 20:40 |
| mat128 | hmm | 20:40 |
| Jabadia | Not thanking him yet :) | 20:40 |
| mat128 | Jabadia: haha | 20:40 |
| mat128 | Jabadia: what does your overcloud image look like? | 20:40 |
| mat128 | qcow? | 20:41 |
| *** Goneri has quit IRC | 20:41 | |
| Jabadia | overcloud-full-vmlinuz: Linux kernel x86 boot executable bzImage, version 3.10.0-327.13.1.el7.x86_64 (mockbuild@x86-019.build.eng.bos.red, RO-rootFS, swap_dev 0x4, Normal VGA | 20:42 |
| Jabadia | Anyway, I'll take it with RH. | 20:42 |
| mat128 | Jabadia: thats just a kernel | 20:42 |
| mat128 | Jabadia: the instructions for the inspection (or even deployment) image are the same as any ramdisk | 20:42 |
| mat128 | I suspect you have a -ramdisk right next to it | 20:43 |
| mat128 | Jabadia: http://docs.openstack.org/developer/tripleo-docs/basic_deployment/basic_deployment_cli.html | 20:43 |
| mat128 | overcloud-full.initrd | 20:43 |
| Jabadia | I also have overcloud-full-vmlinuz | 20:43 |
| mat128 | unless you want to change kernels, the only thing you can switch / edit would be the ramdisk | 20:45 |
| Jabadia | allright, tnx | 20:45 |
| mat128 | good luck :) | 20:46 |
| *** ayoung has quit IRC | 20:49 | |
| *** vishwanathj has joined #openstack-ironic | 20:50 | |
| *** dprince has quit IRC | 20:51 | |
| *** jaybeale has quit IRC | 20:52 | |
| *** vishwanathj has quit IRC | 20:56 | |
| *** Jabadia has quit IRC | 21:09 | |
| openstackgerrit | Merged openstack/ironic: Remove support for the old ramdisk (DIB deploy-ironic element) https://review.openstack.org/327083 | 21:17 |
| *** lindycoder has quit IRC | 21:19 | |
| openstackgerrit | Merged openstack/ironic: Support for APC AP7922 https://review.openstack.org/328476 | 21:19 |
| mat128 | ^thanks everyone :) | 21:19 |
| mat128 | weekend for me | 21:19 |
| mat128 | have a good one | 21:19 |
| JayF | Just put the initial commit here: https://github.com/jayofdoom/ipa-example-hardware-managers | 21:22 |
| JayF | would love for someone to have a look and provide comments | 21:22 |
| JayF | jroll: ^ | 21:22 |
| *** garthb has quit IRC | 21:28 | |
| jroll | JayF: so... we have an official init_hardware thing now, right? why not use that? | 21:33 |
| JayF | no we don't | 21:35 |
| JayF | I -1' | 21:35 |
| JayF | -1'd that patch and suggested it be put in evaluate_hardware_support() | 21:35 |
| jroll | ah I see | 21:35 |
| JayF | that whole patch was why it was decided something like this was needed | 21:35 |
| JayF | (jay knowledge sharing on hwms) | 21:35 |
| jroll | JayF: looks pretty good | 21:35 |
| JayF | because you have to init the hardware to know if you support it | 21:35 |
| jroll | in the upgrade fw thing... | 21:35 |
| *** sdake has quit IRC | 21:36 | |
| jroll | I'd do a try/except around the upgrade to show how to propogate a failure | 21:36 |
| JayF | well that's the thing, right | 21:36 |
| JayF | imo the best thing is to /not/ catch it | 21:36 |
| JayF | an exception raised would pass on through and fail cleaning | 21:37 |
| JayF | which is what you want | 21:37 |
| JayF | if anything, the try;except would be better in the evaluate_hardware_support(), where you may not want IPA to die if the hardwaer manager can't initalize | 21:37 |
| JayF | but that's not really, hwm specific, it's just python | 21:37 |
| JayF | so I skipped it | 21:37 |
| jroll | JayF: maybe try: foo except: LOG.exception(e), raise | 21:37 |
| jroll | ? | 21:37 |
| JayF | sure, that'd be fine | 21:38 |
| jroll | but yeah A+ | 21:38 |
| JayF | jroll: https://github.com/jayofdoom/ipa-example-hardware-managers/commit/4a01e803c56ce2b35a9a71c1001504079b14e5c1 | 21:39 |
| * JayF needs to change branches so he stops accidentally committing all this stuff to master and can PR it | 21:39 | |
| *** rbudden has quit IRC | 21:40 | |
| jroll | lol | 21:40 |
| JayF | that's what you wanted, right? | 21:40 |
| jroll | nit: s/pass through/bubble up/ | 21:40 |
| jroll | yeah | 21:40 |
| jroll | lgtm | 21:40 |
| JayF | woo | 21:40 |
| JayF | going to work on the other two examples: a MonolithicHardwareManager which would be subclassed off GenericHardwareManager and loaded along | 21:40 |
| JayF | and a BusinessLogicHardwareManager which would assert things about the hardware and fail if they don't pass (like maybe wear on disks or something) | 21:41 |
| *** sdake has joined #openstack-ironic | 21:41 | |
| jroll | nice | 21:41 |
| JayF | those will catch all the use cases we're doing downstream for hwms right? | 21:41 |
| JayF | monolithic = original v1 manager | 21:41 |
| JayF | the device managers we have, obviously | 21:41 |
| JayF | and the one business logic manager | 21:41 |
| jroll | yeah, it should | 21:42 |
| JayF | I should've done this a long, long time ago | 21:42 |
| JayF | do you think I'll still need to do a presentation or something on this? | 21:42 |
| JayF | or will the examples be enough | 21:42 |
| * JayF guesses that question is better targeted at wajdi / mat128 or dtantsur|afk | 21:43 | |
| *** sdake_ has joined #openstack-ironic | 21:43 | |
| *** Sukhdev has joined #openstack-ironic | 21:43 | |
| jroll | heh, yeah ask them | 21:44 |
| jroll | I assume this should be enough | 21:44 |
| JayF | So do I :) | 21:44 |
| JayF | I might use this as some of the basis for an IPA deep dive at the summit though | 21:44 |
| jroll | ++ | 21:44 |
| jroll | alright, 1745, I should bounce | 21:45 |
| jroll | have a good weekend | 21:45 |
| JayF | have a good weekend o/ | 21:45 |
| *** sdake has quit IRC | 21:45 | |
| *** spandhe has joined #openstack-ironic | 21:51 | |
| *** sabeen has quit IRC | 21:52 | |
| *** lascii is now known as alaski | 21:52 | |
| *** vishwanathj has joined #openstack-ironic | 21:52 | |
| *** sabeen has joined #openstack-ironic | 21:53 | |
| *** vishwanathj has quit IRC | 21:57 | |
| *** fragatin_ has joined #openstack-ironic | 22:03 | |
| *** catintheroof has quit IRC | 22:06 | |
| *** fragatina has quit IRC | 22:07 | |
| *** Sukhdev has quit IRC | 22:11 | |
| *** garthb has joined #openstack-ironic | 22:18 | |
| *** PollyZ has quit IRC | 22:19 | |
| *** cdearborn has quit IRC | 22:22 | |
| *** garthb_ has joined #openstack-ironic | 22:23 | |
| *** garthb has quit IRC | 22:23 | |
| *** Jabadia has joined #openstack-ironic | 22:30 | |
| Jabadia | Last one for the weekend if anyone still here. Jenkins failed on some tests. the failure occurs also without my changes. what's the process ? | 22:33 |
| Jabadia | it's not fun to get -1 from jenkins. | 22:34 |
| *** Jabadia has quit IRC | 22:40 | |
| *** Jabadia has joined #openstack-ironic | 22:40 | |
| *** mtanino__ has joined #openstack-ironic | 22:43 | |
| *** wajdi_ has joined #openstack-ironic | 22:43 | |
| *** mtanino has quit IRC | 22:45 | |
| *** Jabadia has quit IRC | 22:45 | |
| *** wajdi has quit IRC | 22:47 | |
| *** wajdi_ has quit IRC | 22:47 | |
| *** jaybeale has joined #openstack-ironic | 22:48 | |
| *** mtanino has joined #openstack-ironic | 22:54 | |
| *** rama_y has joined #openstack-ironic | 22:55 | |
| *** mtanino__ has quit IRC | 22:56 | |
| JayF | jroll: (others as well) rfr -> https://github.com/jayofdoom/ipa-example-hardware-managers/pull/1/files | 22:58 |
| *** sabeen has quit IRC | 23:01 | |
| *** garthb_ has quit IRC | 23:14 | |
| JayF | NobodyCam: we should combine http://docs.openstack.org/developer/ironic-python-agent/#customizing-the-image that section with what you just put in at some point | 23:16 |
| JayF | NobodyCam: probably could justify ripping out the 'customize a ramdisk' doc and putting it in its own file | 23:16 |
| *** vishwanathj has joined #openstack-ironic | 23:18 | |
| *** vishwanathj has quit IRC | 23:22 | |
| *** fragatin_ has quit IRC | 23:22 | |
| *** fragatina has joined #openstack-ironic | 23:23 | |
| *** sabeen has joined #openstack-ironic | 23:28 | |
| *** sacharya has quit IRC | 23:34 | |
| *** spandhe has quit IRC | 23:35 | |
| *** mbound has quit IRC | 23:36 | |
| *** spandhe has joined #openstack-ironic | 23:43 | |
| JayF | While working on these docs, I've come to the conclusion we should strongly discourage people subclassing HWMs off GenericHardwareManager | 23:48 |
| JayF | there's almost zero reason to do it that isn't better served by something else | 23:48 |
| JayF | I can't even come up with a fake convoluted example for my docs. | 23:48 |
| *** ppiela has quit IRC | 23:51 | |
| *** spandhe has quit IRC | 23:52 | |
| *** praneshp has quit IRC | 23:52 | |
| devananda | JayF: i thought that's _exactly_ what we encouraged? | 23:52 |
| JayF | Docs say you can do it subclassing HardwareManager OR GenericHardwareManager | 23:52 |
| devananda | aahhh | 23:52 |
| JayF | GenericHardwareManager being where all the "do stuff" in IPA is, and HardwareManager is just a shell | 23:53 |
| JayF | and I'm thinking we need to change that doc to discourage it | 23:53 |
| JayF | because it's all downsides and no upsides to subclassing GHM afict | 23:53 |
| devananda | what should it encourage? | 23:53 |
| JayF | have you seen what I've been working on today :D | 23:53 |
| devananda | nope | 23:53 |
| JayF | https://github.com/jayofdoom/ipa-example-hardware-managers | 23:53 |
| JayF | has one example in it | 23:53 |
| JayF | and another in a PR waiting for jroll to review | 23:53 |
| JayF | my hope is this can become officialish. It's the documentation we always needed but never really did. | 23:54 |
| JayF | the one in the PR is sorta my favorite example. We have a lot of business logic-y clean steps like that downstream | 23:54 |
| JayF | pretty much did this all afternoon in lieu of doing a presentation | 23:55 |
| devananda | JayF: good stuff | 23:59 |
| JayF | you find that helpful though? | 23:59 |
| JayF | Pretty hard to write docs for something you know that well :) I don't know what folks don't know | 23:59 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!