| *** heyongli has quit IRC | 00:02 | |
| *** heyongli has joined #openstack-ironic | 00:04 | |
| openstackgerrit | VIncent S. Cojot proposed openstack/ironic: Hi everyone, There are a few more problems with this patch. I am working on an improved version. There are some issues I hadn't initially detected because I was simply using a VboxManage Wrapper. As in: sudo perl -pi -e "s@VBoxManage'@VBoxManageWrapper'@ https://review.openstack.org/317780 | 00:07 |
|---|---|---|
| *** ptrlv has quit IRC | 00:21 | |
| *** piet has quit IRC | 00:27 | |
| *** hoangcx has joined #openstack-ironic | 00:30 | |
| *** piet has joined #openstack-ironic | 00:32 | |
| openstackgerrit | Merged openstack/ironic: Updated from global requirements https://review.openstack.org/317376 | 00:41 |
| *** rloo has quit IRC | 00:45 | |
| *** baoli has joined #openstack-ironic | 00:46 | |
| *** rloo has joined #openstack-ironic | 00:47 | |
| *** YiboCai has joined #openstack-ironic | 00:51 | |
| *** piet has quit IRC | 00:55 | |
| *** hoangcx has quit IRC | 00:56 | |
| *** hoangcx has joined #openstack-ironic | 00:57 | |
| *** mtanino has quit IRC | 01:00 | |
| *** priteau has joined #openstack-ironic | 01:08 | |
| *** sacharya has quit IRC | 01:09 | |
| *** sabeen has quit IRC | 01:16 | |
| *** ijw has quit IRC | 01:19 | |
| *** r-daneel has quit IRC | 01:42 | |
| *** sabeen has joined #openstack-ironic | 01:43 | |
| -openstackstatus- NOTICE: Gerrit is about to be restarted to help with page timeouts | 01:46 | |
| *** rloo has quit IRC | 02:00 | |
| *** sdake has quit IRC | 02:06 | |
| zhenguo_ | jlvillal: are you still around? | 02:08 |
| *** sacharya has joined #openstack-ironic | 02:09 | |
| *** sdake has joined #openstack-ironic | 02:12 | |
| openstackgerrit | Andre keedy proposed openstack/ironic-inspector: Replace node data via uuid/data/unprocessed API https://review.openstack.org/308379 | 02:14 |
| *** sacharya has quit IRC | 02:14 | |
| *** yuanying has joined #openstack-ironic | 02:17 | |
| zhenguo_ | or other nova liaisons can you help to confirm a very obvious nova ironic driver bug https://bugs.launchpad.net/nova/+bug/1582680 ? | 02:17 |
| openstack | Launchpad bug 1582680 in OpenStack Compute (nova) "Ironic: wrong check for ready to deploy" [Undecided,In progress] - Assigned to Zhenguo Niu (niu-zglinux) | 02:17 |
| *** sdake has quit IRC | 02:22 | |
| *** spandhe has quit IRC | 02:48 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic-specs: Support rolling upgrades https://review.openstack.org/299245 | 02:49 |
| *** ChrisAusten has joined #openstack-ironic | 02:53 | |
| *** priteau has quit IRC | 02:55 | |
| *** baoli has quit IRC | 02:57 | |
| *** priteau has joined #openstack-ironic | 02:59 | |
| *** ChrisAusten has quit IRC | 03:00 | |
| jlvillal | zhenguo_, Here for the moment. | 03:01 |
| jlvillal | About to leave... | 03:01 |
| *** vishwanathj has quit IRC | 03:05 | |
| *** sdake has joined #openstack-ironic | 03:10 | |
| *** priteau has quit IRC | 03:11 | |
| *** links has joined #openstack-ironic | 03:26 | |
| *** sdake_ has joined #openstack-ironic | 03:37 | |
| *** sdake has quit IRC | 03:40 | |
| *** ijw has joined #openstack-ironic | 03:53 | |
| *** piet has joined #openstack-ironic | 03:53 | |
| *** sacharya has joined #openstack-ironic | 03:53 | |
| *** sdake_ has quit IRC | 03:58 | |
| *** sacharya has quit IRC | 03:58 | |
| *** sdake has joined #openstack-ironic | 04:09 | |
| *** sabeen has quit IRC | 04:19 | |
| *** amotoki has quit IRC | 04:26 | |
| *** amotoki has joined #openstack-ironic | 04:26 | |
| *** ChrisAusten has joined #openstack-ironic | 04:28 | |
| *** hamzy has joined #openstack-ironic | 04:28 | |
| *** zhihais has joined #openstack-ironic | 04:29 | |
| *** zhihaisong has quit IRC | 04:32 | |
| *** caoshufeng_ has joined #openstack-ironic | 04:43 | |
| *** piet has quit IRC | 04:44 | |
| *** jcoufal has joined #openstack-ironic | 04:45 | |
| caoshufeng_ | Hi, I use agent_ssh as ironic deply driver. It's ok when I boot the first vm. But after the first vm is deployed, all three's nodes "memory_mb" becomes zero.. | 04:46 |
| caoshufeng_ | I use "nova hypervisor-show" to get the memory_mb value. | 04:47 |
| *** zhihais has quit IRC | 04:51 | |
| *** zhihais has joined #openstack-ironic | 04:51 | |
| *** sacharya has joined #openstack-ironic | 04:52 | |
| *** piet has joined #openstack-ironic | 04:59 | |
| *** ijw has quit IRC | 05:01 | |
| *** zhihaisong has joined #openstack-ironic | 05:05 | |
| *** sacharya has quit IRC | 05:06 | |
| *** mikal_ has joined #openstack-ironic | 05:08 | |
| *** ptx0_ has joined #openstack-ironic | 05:08 | |
| *** davidlenwell has quit IRC | 05:09 | |
| *** sambetts has joined #openstack-ironic | 05:11 | |
| *** Shalmazar has joined #openstack-ironic | 05:12 | |
| *** sambetts|afk has quit IRC | 05:12 | |
| *** piet has quit IRC | 05:12 | |
| *** mikal has quit IRC | 05:12 | |
| *** mjturek1 has quit IRC | 05:12 | |
| *** yuriyz has quit IRC | 05:12 | |
| *** links has quit IRC | 05:12 | |
| *** sirushti has quit IRC | 05:12 | |
| *** Madasi has quit IRC | 05:12 | |
| *** zhihais has quit IRC | 05:12 | |
| *** ptx0 has quit IRC | 05:12 | |
| *** vsaienko1 has quit IRC | 05:12 | |
| *** Shalmazar is now known as Madasi | 05:12 | |
| *** sirushti has joined #openstack-ironic | 05:13 | |
| *** yuriyz has joined #openstack-ironic | 05:13 | |
| *** links has joined #openstack-ironic | 05:13 | |
| *** mjturek1 has joined #openstack-ironic | 05:14 | |
| *** vsaienko has joined #openstack-ironic | 05:16 | |
| *** davidlenwell has joined #openstack-ironic | 05:17 | |
| *** ChubYann has quit IRC | 05:24 | |
| *** sdake has quit IRC | 05:26 | |
| *** jaybeale has joined #openstack-ironic | 05:30 | |
| *** ChrisAusten has quit IRC | 05:32 | |
| *** rcernin has joined #openstack-ironic | 05:47 | |
| *** fragatin_ has joined #openstack-ironic | 05:52 | |
| *** fragatin_ has quit IRC | 05:52 | |
| *** _vsaienko has joined #openstack-ironic | 05:52 | |
| *** fragatin_ has joined #openstack-ironic | 05:53 | |
| *** fragatin_ has quit IRC | 05:53 | |
| *** _vs has joined #openstack-ironic | 05:53 | |
| *** fragatin_ has joined #openstack-ironic | 05:54 | |
| *** fragatina has quit IRC | 05:55 | |
| *** jaybeale has quit IRC | 05:57 | |
| *** fragatin_ has quit IRC | 05:59 | |
| *** _vsaienko has quit IRC | 05:59 | |
| *** _vs has quit IRC | 06:00 | |
| *** _vs has joined #openstack-ironic | 06:00 | |
| *** _vsaienko has joined #openstack-ironic | 06:00 | |
| *** mjura has joined #openstack-ironic | 06:04 | |
| *** sacharya has joined #openstack-ironic | 06:05 | |
| *** david-lyle has quit IRC | 06:05 | |
| *** _vs has quit IRC | 06:06 | |
| *** _vsaienko has quit IRC | 06:06 | |
| *** jrist has quit IRC | 06:07 | |
| *** fragatina has joined #openstack-ironic | 06:07 | |
| *** _vsaienko has joined #openstack-ironic | 06:08 | |
| *** _vs has joined #openstack-ironic | 06:08 | |
| *** sacharya has quit IRC | 06:10 | |
| *** fragatina has quit IRC | 06:12 | |
| *** moshele has joined #openstack-ironic | 06:13 | |
| *** e0ne has joined #openstack-ironic | 06:13 | |
| *** gcb has joined #openstack-ironic | 06:14 | |
| *** chopmann has joined #openstack-ironic | 06:18 | |
| *** jrist has joined #openstack-ironic | 06:22 | |
| *** jrist has quit IRC | 06:22 | |
| *** jrist has joined #openstack-ironic | 06:22 | |
| *** _vs has quit IRC | 06:24 | |
| *** _vsaienko has quit IRC | 06:24 | |
| *** fragatina has joined #openstack-ironic | 06:24 | |
| *** fragatina has quit IRC | 06:35 | |
| *** fragatina has joined #openstack-ironic | 06:36 | |
| *** jtomasek has joined #openstack-ironic | 06:45 | |
| *** rcernin has quit IRC | 06:47 | |
| *** _vsaienko has joined #openstack-ironic | 06:50 | |
| *** _vs has joined #openstack-ironic | 06:50 | |
| *** rcernin has joined #openstack-ironic | 06:51 | |
| *** e0ne has quit IRC | 06:54 | |
| *** _vsaienko has quit IRC | 07:04 | |
| *** _vs has quit IRC | 07:06 | |
| *** _vsaienko has joined #openstack-ironic | 07:07 | |
| *** _vs has joined #openstack-ironic | 07:07 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: WIP: grenade checking https://review.openstack.org/311109 | 07:12 |
| vdrok | morning ironic! | 07:24 |
| chopmann | morning | 07:25 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: DO NOT REVIEW https://review.openstack.org/296432 | 07:25 |
| *** gcb has quit IRC | 07:27 | |
| *** dmk0202 has joined #openstack-ironic | 07:35 | |
| *** _vs has quit IRC | 07:35 | |
| *** _vsaienko has quit IRC | 07:35 | |
| milan | morning vdrok chopmann | 07:35 |
| vdrok | morning milan and chopmann | 07:36 |
| vdrok | caoshufeng_: here is the fix - https://review.openstack.org/#/c/316031/2 | 07:36 |
| vdrok | it's not driver-specific issue | 07:37 |
| *** _vsaienko has joined #openstack-ironic | 07:37 | |
| *** _vs has joined #openstack-ironic | 07:37 | |
| *** gcb has joined #openstack-ironic | 07:41 | |
| *** _vsaienko has quit IRC | 07:50 | |
| *** _vs has quit IRC | 07:51 | |
| *** _vsaienko has joined #openstack-ironic | 07:53 | |
| openstackgerrit | Milan Kováčik proposed openstack/ironic-inspector: [WIP] Tempest: add basic test https://review.openstack.org/293926 | 07:53 |
| openstackgerrit | Naohiro Tamura proposed openstack/ironic-specs: Enhance Power Interface for Soft Power Off and Inject NMI https://review.openstack.org/186700 | 07:53 |
| *** _vs has joined #openstack-ironic | 07:54 | |
| *** gcb has quit IRC | 07:54 | |
| *** jistr has joined #openstack-ironic | 07:54 | |
| *** daemontool has joined #openstack-ironic | 07:55 | |
| *** david-lyle has joined #openstack-ironic | 07:57 | |
| *** _vsaienko has quit IRC | 07:57 | |
| *** _vs has quit IRC | 07:57 | |
| *** zzzeek has quit IRC | 08:00 | |
| caoshufeng_ | vdrok thanks!!! | 08:00 |
| *** zzzeek has joined #openstack-ironic | 08:01 | |
| *** aarefiev has joined #openstack-ironic | 08:02 | |
| *** _vsaienko has joined #openstack-ironic | 08:02 | |
| *** _vs has joined #openstack-ironic | 08:02 | |
| *** sacharya has joined #openstack-ironic | 08:06 | |
| *** sacharya has quit IRC | 08:10 | |
| *** gcb has joined #openstack-ironic | 08:11 | |
| *** ohamada has joined #openstack-ironic | 08:15 | |
| *** _vsaienko has quit IRC | 08:17 | |
| *** _vs has quit IRC | 08:17 | |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: WIP: grenade checking https://review.openstack.org/311109 | 08:24 |
| *** dtantsur|afk is now known as dtantsur | 08:29 | |
| dtantsur | Morning Ironic | 08:30 |
| *** lucas-dinner is now known as lucasagomes | 08:32 | |
| lucasagomes | morning all | 08:32 |
| *** mbound has joined #openstack-ironic | 08:36 | |
| dtantsur | just got a bug report that introspection does not work on VirtualBox under Mac OS X >_< | 08:38 |
| * dtantsur blames Mac OS X ofc | 08:38 | |
| *** _milan_ has joined #openstack-ironic | 08:40 | |
| *** yuikotak_ has joined #openstack-ironic | 08:41 | |
| *** derekh has joined #openstack-ironic | 08:42 | |
| *** wendar_ has joined #openstack-ironic | 08:43 | |
| *** kozhukalov_ has joined #openstack-ironic | 08:43 | |
| *** eglute_s has joined #openstack-ironic | 08:44 | |
| *** bapalm_ has joined #openstack-ironic | 08:44 | |
| *** lucasagomes_ has joined #openstack-ironic | 08:45 | |
| *** yuikotakadamori has quit IRC | 08:45 | |
| *** thiagop has quit IRC | 08:45 | |
| *** milan has quit IRC | 08:45 | |
| *** bapalm has quit IRC | 08:45 | |
| *** cloudnull has quit IRC | 08:45 | |
| *** kozhukalov has quit IRC | 08:45 | |
| *** odyssey4me has quit IRC | 08:45 | |
| *** yarkot1 has quit IRC | 08:45 | |
| *** rbradfor has quit IRC | 08:45 | |
| *** eglute has quit IRC | 08:45 | |
| *** mgagne has quit IRC | 08:45 | |
| *** aarefiev has quit IRC | 08:45 | |
| *** lucasagomes has quit IRC | 08:45 | |
| *** Nakato has quit IRC | 08:45 | |
| *** wendar has quit IRC | 08:45 | |
| *** Nakato has joined #openstack-ironic | 08:45 | |
| *** mgagne has joined #openstack-ironic | 08:45 | |
| *** kromanenko_ has joined #openstack-ironic | 08:46 | |
| *** Romanenko_K has joined #openstack-ironic | 08:46 | |
| *** yarkot1 has joined #openstack-ironic | 08:46 | |
| *** mgagne is now known as Guest92292 | 08:46 | |
| *** thiagop has joined #openstack-ironic | 08:46 | |
| *** odyssey4me has joined #openstack-ironic | 08:46 | |
| *** kromanenko__ has joined #openstack-ironic | 08:46 | |
| vdrok | morning dtantsur and lucasagomes ! | 08:46 |
| *** rbradfor has joined #openstack-ironic | 08:46 | |
| *** aarefiev has joined #openstack-ironic | 08:46 | |
| lucasagomes_ | dtantsur, WONTFIX heh | 08:47 |
| lucasagomes_ | jk | 08:47 |
| *** serverascode has quit IRC | 08:47 | |
| *** kozhukalov_ is now known as kozhukalov | 08:48 | |
| *** cloudkiller has joined #openstack-ironic | 08:48 | |
| dtantsur | lucasagomes_, no jk, I really closed it as WONTFIX | 08:48 |
| * lucasagomes_ firealarm of his building went of, wonder if it's a test | 08:48 | |
| dtantsur | and the reason is that it does not support memory in dmidecode | 08:48 |
| dtantsur | (it = vbox ofc) | 08:48 |
| lucasagomes_ | oh | 08:48 |
| dtantsur | also morning vdrok | 08:48 |
| *** lucasagomes_ is now known as lucasagomes | 08:49 | |
| *** e0ne has joined #openstack-ironic | 08:49 | |
| *** francoblanco has quit IRC | 08:49 | |
| *** sylwesterB has quit IRC | 08:49 | |
| *** fyxim has quit IRC | 08:49 | |
| *** sergek has quit IRC | 08:50 | |
| vsaienko | morning Ironic'ers | 08:50 |
| dtantsur | morning vsaienko | 08:51 |
| *** rm_work has quit IRC | 08:53 | |
| *** rm_work has joined #openstack-ironic | 08:54 | |
| *** sergek has joined #openstack-ironic | 08:59 | |
| *** francoblanco has joined #openstack-ironic | 08:59 | |
| *** serverascode has joined #openstack-ironic | 08:59 | |
| *** daemontool_ has joined #openstack-ironic | 09:00 | |
| *** daemontool_ has quit IRC | 09:00 | |
| zhenguo_ | morning dtantsur, lucasagomes, vdrok, vsaienko | 09:01 |
| vdrok | hi zhenguo_ | 09:01 |
| openstackgerrit | Merged openstack/ironic: add mitaka configuration reference link to the index page https://review.openstack.org/317525 | 09:01 |
| *** sylwesterB has joined #openstack-ironic | 09:02 | |
| lucasagomes | zhenguo_, hi there! | 09:02 |
| *** daemontool has quit IRC | 09:03 | |
| *** fyxim has joined #openstack-ironic | 09:03 | |
| *** mgould has joined #openstack-ironic | 09:05 | |
| *** agordeev has quit IRC | 09:07 | |
| *** agordeev has joined #openstack-ironic | 09:09 | |
| vsaienko | green multitenancy job again http://logs.openstack.org/32/296432/36/experimental/gate-tempest-dsvm-ironic-multitenant-network-nv/e3c6307/ :) | 09:11 |
| sergek | o/ | 09:12 |
| *** jtomasek_ has joined #openstack-ironic | 09:12 | |
| dtantsur | awesome! | 09:14 |
| lucasagomes | vsaienko, w00t! | 09:17 |
| *** daemontool has joined #openstack-ironic | 09:18 | |
| sambetts | Morning all | 09:20 |
| dtantsur | morning sambetts | 09:21 |
| sambetts | o/ dtantsur | 09:21 |
| dtantsur | sambetts, I spent some time discussing the local_gb issue on the internal tech list; seems like there is no way to know when hardware is ready | 09:22 |
| * dtantsur is lost | 09:22 | |
| vsaienko | morning sambetts | 09:23 |
| sambetts | dtantsur: that is a nightmare... so save doing a sleep 120 we really have no other way? and I guess even then the machine might be slow enough not to be ready :/ | 09:23 |
| dtantsur | yes, so we need the sleep to be configurable >_< I can't believe I'm seriously saying that.. | 09:24 |
| sambetts | sigh... | 09:24 |
| mgould | morning ironic | 09:27 |
| dtantsur | morning mgould | 09:27 |
| mgould | morning dtantsur sambetts vsaienko sergek | 09:28 |
| vsaienko | morning mgould | 09:28 |
| vdrok | morning sambetts and mgould | 09:31 |
| sambetts | dtantsur: https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/5/html/Online_Storage_Reconfiguration_Guide/scanning-storage-interconnects.html ?? | 09:32 |
| *** jtomasek_ has quit IRC | 09:32 | |
| sambetts | dtantsur: I wonder if we can watch the /sys/class/scsi_host directory to see if something is there or not :/ | 09:32 |
| _milan_ | morning sambetts dtantsur vsaienko lucasagomes :) | 09:33 |
| sambetts | _milan_: you've gained another _ | 09:33 |
| dtantsur | lol | 09:33 |
| dtantsur | morning _milan_ | 09:33 |
| dtantsur | I suggest the next nick be =^_^=milan=^_^= | 09:33 |
| * _milan_ getting more and more private | 09:33 | |
| _milan_ | attribute of this free node channel ;) | 09:34 |
| dtantsur | sambetts, the problem with waiting for one disk is that we can have more than 1 disk.. of which we don't necessary need the 1st | 09:34 |
| sambetts | dtantsur: I'm wondering if we can scan /sys/class to get a picture of what devices are plugged in and then if we notice a raid thats not ready then poll for it or something :/ | 09:35 |
| dtantsur | sambetts, I wonder if /sys/class is always populated quickly | 09:35 |
| *** KennethWilke has quit IRC | 09:36 | |
| mgould | morning vdrok _milan_ | 09:36 |
| _milan_ | morning mgould | 09:37 |
| sambetts | dtantsur: it should be populated at kernel run time | 09:37 |
| dtantsur | sambetts, welll.. while the driver is loading? ;) | 09:37 |
| dtantsur | I wish I had direct access to the offending machine, then I would experiment | 09:37 |
| sambetts | I would hope it populates it first because it has to know which device it is to load the driver | 09:38 |
| *** e0ne has quit IRC | 09:40 | |
| *** Khayam has joined #openstack-ironic | 09:41 | |
| lucasagomes | _milan_, hi there :D | 09:41 |
| *** e0ne has joined #openstack-ironic | 09:42 | |
| _milan_ | lucasagomes, o/ :D | 09:44 |
| *** stendulker has joined #openstack-ironic | 09:46 | |
| *** cloudkiller is now known as cloudnull | 09:46 | |
| mgould | morning lucasagomes | 09:48 |
| lucasagomes | mgould, hi there, good morning | 09:48 |
| sambetts | hey mgould | 09:48 |
| sambetts | and anyone else I've missed while I've been down the systemd rabbithole | 09:48 |
| *** electrofelix has joined #openstack-ironic | 09:49 | |
| Khayam | hi. | 09:49 |
| Khayam | i am new to ironic. I have read its basic documentation and workflow | 09:49 |
| Khayam | Now I am interested in contributing. Whats the easy way to get there | 09:49 |
| Khayam | ? | 09:49 |
| sambetts | Hi Khayam, welcome to Ironic! Have you contributed to OpenStack before? | 09:50 |
| Khayam | yes. | 09:51 |
| Khayam | i am contributor to Tricircle project | 09:52 |
| Khayam | tricircle is in start now much contributors. I fixed 2 bugs there | 09:52 |
| Khayam | *tricircle is in start.. not much contributors. I fixed 2 bugs there | 09:52 |
| openstackgerrit | Tan Lin proposed openstack/ironic: Add command to force a deploy failure https://review.openstack.org/311273 | 09:53 |
| sambetts | Khayam: Oh cool! So you know the processes required to commit into an OpenStack project etc? | 09:53 |
| sambetts | how to use gerrit etc | 09:53 |
| *** KennethWilke has joined #openstack-ironic | 09:53 | |
| Khayam | yes. i am aware of that. | 09:55 |
| Khayam | Actullay I want to get a highlevel overview of ironic source code. | 09:56 |
| sambetts | dtantsur: I wonder if we can add some super generic udev rules that match all block devices of something that will trigger some target we can watch | 09:57 |
| *** YiboCai has quit IRC | 09:57 | |
| dtantsur | dunno.. from what people told me, it's not possible | 09:57 |
| lucasagomes | Khayam, welcome! I would suggest you run Ironic in devstack and play a bit with it, http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html#deploying-ironic-with-devstack | 09:57 |
| lucasagomes | look at our bug list for low-hanging-fruit bugs | 09:58 |
| sambetts | For an overview of the system arch etc http://docs.openstack.org/developer/ironic/#developer-s-guide | 09:58 |
| *** PixieBoots has joined #openstack-ironic | 09:59 | |
| Khayam | ok. but 1 question. as ironic uses pxe and ipmi , so will it work in vms? | 09:59 |
| lucasagomes | Khayam, for ipmi it uses a project called VirtualBMC to translate IPMI commands to libvirt commands | 09:59 |
| lucasagomes | if you set the IRONIC_DRIVER in devstack to agent_ipmitool or pxe_ipmitool it will set up virtualbmc for you | 10:00 |
| lucasagomes | Khayam, there's also the *_ssh drivers which mock power commands by ssh'ing into a host and issuing virsh (or other hypervisors) commands | 10:00 |
| lucasagomes | Khayam, so the answer is, there are ways of doing that yes :-) | 10:01 |
| lucasagomes | #pixiesay hello dtantsur | 10:01 |
| PixieBoots | ʕ•͡ᴥ•ʔ: hello dtantsur | 10:01 |
| * lucasagomes made that yesterday | 10:01 | |
| * lucasagomes night | 10:01 | |
| Khayam | great | 10:01 |
| Khayam | thanks | 10:01 |
| dtantsur | LOL fantastic! | 10:03 |
| lucasagomes | heh | 10:03 |
| lucasagomes | it does useful things too | 10:03 |
| dtantsur | #pixiesay Ironic rules | 10:03 |
| PixieBoots | ʕ•͡ᴥ•ʔ: Ironic rules | 10:03 |
| lucasagomes | #failgraph gate-tempest-dsvm-ironic-agent_ssh gate-tempest-dsvm-ironic-pxe_ssh | 10:03 |
| PixieBoots | ʕ•͡ᴥ•ʔ: http://tinyurl.com/hq6mvgd | 10:03 |
| dtantsur | oh, this is good one | 10:03 |
| lucasagomes | yeah | 10:03 |
| lucasagomes | well I will opensource it once the code is a bit better heh | 10:03 |
| lucasagomes | so we can enhance it | 10:04 |
| * lucasagomes should just put it in github | 10:04 | |
| dtantsur | yeah, just put it there | 10:05 |
| dtantsur | even my bug dashboard is open source, and it looks terrible :D | 10:06 |
| lucasagomes | hah yup | 10:06 |
| lucasagomes | will do | 10:07 |
| *** sacharya has joined #openstack-ironic | 10:07 | |
| *** zhihaisong has quit IRC | 10:09 | |
| *** sacharya has quit IRC | 10:13 | |
| openstackgerrit | Merged openstack/python-ironic-inspector-client: Updated from global requirements https://review.openstack.org/317499 | 10:17 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-python-agent: Cache node object inside hw manager instances https://review.openstack.org/317538 | 10:18 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-python-agent: Get root device hints from the node object https://review.openstack.org/296539 | 10:18 |
| *** _vsaienko has joined #openstack-ironic | 10:19 | |
| *** _vs has joined #openstack-ironic | 10:19 | |
| *** _vsaienko has quit IRC | 10:19 | |
| *** _vs has quit IRC | 10:19 | |
| *** priteau has joined #openstack-ironic | 10:27 | |
| _milan_ | #pixiesay lucasagomes, nice one :D | 10:32 |
| PixieBoots | ʕ•͡ᴥ•ʔ: lucasagomes, nice one :D | 10:32 |
| lucasagomes | :D | 10:32 |
| lucasagomes | #pixiesay cheers _milan_ | 10:32 |
| PixieBoots | ʕ•͡ᴥ•ʔ: cheers _milan_ | 10:32 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-python-agent: [WIP] Support long hardware initialization https://review.openstack.org/317969 | 10:32 |
| dtantsur | sambetts, this is what I ended up with ^^^ | 10:32 |
| _milan_ | :D | 10:32 |
| sambetts | dtantsur: pretty neat | 10:37 |
| *** yuikotak_ has quit IRC | 10:37 | |
| sambetts | dtantsur: unless we can polling /sys/devices or something like that I think thats the only way forward | 10:37 |
| dtantsur | sambetts, I hope we can figure out polling $something and put it to initialize_hardware later | 10:38 |
| dtantsur | but that will require a reproducer environment, which I don't personally have :( I can't ping a customer every time I need to test a new image... | 10:38 |
| jroll | morning y'all | 10:39 |
| jroll | lucasagomes: nice bot :D | 10:39 |
| lucasagomes | #pixiesay morning jroll :-) | 10:40 |
| PixieBoots | ʕ•͡ᴥ•ʔ: morning jroll :-) | 10:40 |
| lucasagomes | jroll, yeah it has a failgraph command only for now | 10:40 |
| lucasagomes | we can start adding useful stuff to it, like searching for bugs etc | 10:40 |
| lucasagomes | #failgraph gate-tempest-dsvm-ironic-pxe_ssh | 10:40 |
| PixieBoots | ʕ•͡ᴥ•ʔ: http://tinyurl.com/zn748at | 10:40 |
| *** Khayam has quit IRC | 10:40 | |
| jroll | yeah, saw that | 10:40 |
| *** _vsaienko has joined #openstack-ironic | 10:41 | |
| *** _vs has joined #openstack-ironic | 10:41 | |
| dtantsur | morning jroll | 10:41 |
| *** thrash|g0ne is now known as thrash | 10:41 | |
| jroll | dtantsur: woo https://review.openstack.org/#/c/310288/1 | 10:41 |
| sambetts | o/ jroll | 10:41 |
| _milan_ | morning jroll | 10:42 |
| dtantsur | yeah, that's good | 10:42 |
| *** _vs has quit IRC | 10:42 | |
| *** _vsaienko has quit IRC | 10:42 | |
| dtantsur | now we need to claim the upgrade tag, and we're finally responsible adults :D | 10:42 |
| *** _vs has joined #openstack-ironic | 10:42 | |
| *** _vsaienko has joined #openstack-ironic | 10:42 | |
| dtantsur | #pixiesay I am a responsible adult, and I maintain my stable branches carefully | 10:42 |
| PixieBoots | ʕ•͡ᴥ•ʔ: I am a responsible adult, and I maintain my stable branches carefully | 10:42 |
| jroll | lol | 10:42 |
| *** ptx0_ has quit IRC | 10:44 | |
| *** ptx0_ has joined #openstack-ironic | 10:44 | |
| *** ptx0_ is now known as ptx0 | 10:44 | |
| vdrok | morning jroll ! | 10:44 |
| vdrok | lucasagomes: wow, cool :) | 10:44 |
| jroll | hey vdrok :) | 10:44 |
| lucasagomes | heh | 10:44 |
| lucasagomes | there's an easter-egg in the pixiesay, just saying... | 10:44 |
| lucasagomes | tip: something that was said about the project that we didn't like | 10:45 |
| vdrok | is it in portuguese? :) | 10:45 |
| lucasagomes | nop | 10:45 |
| lucasagomes | english | 10:45 |
| * dtantsur is intrigued | 10:45 | |
| jroll | #pixiesay shuttleworth | 10:45 |
| PixieBoots | ʕ•͡ᴥ•ʔ: shuttleworth | 10:45 |
| lucasagomes | so close | 10:46 |
| jroll | #pixiesay nsaaas | 10:46 |
| PixieBoots | ʕ•͡ᴥ•ʔ: nsaaas | 10:46 |
| jroll | #pixiesay nsa | 10:46 |
| PixieBoots | ʕ•͡ᴥ•ʔ: nsa | 10:46 |
| * jroll stops spamming | 10:46 | |
| lucasagomes | #pixiesay NSA-as-a-Service | 10:46 |
| PixieBoots | ╭∩╮ʕ•͡ᴥ•ʔ╭∩╮: Mark Shuttleworth is an intergalactic ass****. See: http://www.markshuttleworth.com/archives/1493 | 10:46 |
| dtantsur | ROFL :D | 10:46 |
| _milan_ | :D | 10:46 |
| jroll | lol | 10:46 |
| jroll | we... probably shouldn't have that here | 10:46 |
| lucasagomes | yeah... I should remove it | 10:46 |
| sambetts | ROFLCOPTER | 10:46 |
| lucasagomes | was funny when I was coding tho | 10:46 |
| dtantsur | yeah, it's enough that we had good laughter right now | 10:46 |
| jroll | heh | 10:47 |
| dtantsur | lucasagomes, we need a special version for our internal IRC. | 10:47 |
| lucasagomes | dtantsur, yeah | 10:47 |
| *** _vsaienko has quit IRC | 10:48 | |
| *** _vs has quit IRC | 10:48 | |
| dtantsur | jroll, should we push someone to review https://review.openstack.org/316655? I'm sick on unstable DIB jobs... | 10:48 |
| jroll | dtantsur: sure | 10:49 |
| * jroll does it | 10:49 | |
| dtantsur | thnx | 10:49 |
| jroll | I'll push again during US workday | 10:49 |
| dtantsur | fair | 10:49 |
| _milan_ | aarefiev, wdyt about the timeouts here: https://review.openstack.org/#/c/293926/12/ironic_inspector/test/inspector_tempest_plugin/tests/test_basic.py | 10:50 |
| vsaienko | green grenade job http://logs.openstack.org/09/311109/22/experimental/gate-grenade-dsvm-ironic/bd11b12/ | 10:50 |
| vsaienko | :P | 10:50 |
| jroll | vsaienko: \o/ | 10:50 |
| dtantsur | OMG YES!! | 10:51 |
| vsaienko | need to summarize all that we have in "WIP and DO NOT REVIEW" patches | 10:51 |
| jroll | vsaienko: well, 2016-05-18 10:49:37.500 | {0} ironic_tempest_plugin.tests.scenario.test_baremetal_basic_ops.BaremetalBasicOps.test_baremetal_server_ops ... SKIPPED: Skipped because the baremetal service is not available | 10:51 |
| jroll | but this is progress! | 10:51 |
| dtantsur | heh | 10:51 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/virtualbmc: Set IPMI response timeout for the virtual BMCs https://review.openstack.org/316022 | 10:52 |
| aarefiev | _milan_: I'll take a look in 1h, btw thanks for updating, which timeout are you facing, probably we need to tune some | 10:53 |
| *** moshele has quit IRC | 10:53 | |
| vsaienko | jroll need to check why Ironic is disabled in tempest.conf http://logs.openstack.org/09/311109/22/experimental/gate-grenade-dsvm-ironic/bd11b12/logs/new/tempest_conf.txt.gz | 10:53 |
| _milan_ | aarefiev, I mean I've hard-coded 360s there | 10:53 |
| * _milan_ not really sure what to do about that | 10:54 | |
| jroll | vsaienko: yeah, likely just missing something | 10:54 |
| vdrok | yeah, driver_enabled is true | 10:54 |
| _milan_ | aarefiev, np ;) | 10:54 |
| aarefiev | _milan_: I'll take a look closer later, it should be config var https://review.openstack.org/#/c/293926/12/ironic_inspector/test/inspector_tempest_plugin/config.py | 10:55 |
| *** moshele has joined #openstack-ironic | 10:56 | |
| lucasagomes | #pixiebot my code is here https://github.com/umago/pixiebot | 10:56 |
| lucasagomes | #pixiesay my code is here https://github.com/umago/pixiebot | 10:57 |
| PixieBoots | ʕ•͡ᴥ•ʔ: my code is here https://github.com/umago/pixiebot | 10:57 |
| * lucasagomes #fail | 10:57 | |
| sambetts | #pixiesay my creator has lost control of me | 10:57 |
| PixieBoots | ʕ•͡ᴥ•ʔ: my creator has lost control of me | 10:57 |
| jroll | vsaienko: vdrok: probably related to https://github.com/openstack-dev/grenade/blob/master/upgrade-tempest#L49 | 10:57 |
| jroll | not sure how to populate that | 10:57 |
| jroll | (from a plugin) | 10:57 |
| *** caoshufeng_ has quit IRC | 10:58 | |
| vdrok | jroll: yep, seems like it | 10:58 |
| lucasagomes | sambetts, heh well let's contribute to it | 10:59 |
| sambetts | ;) | 10:59 |
| *** dtantsur is now known as dtantsur|brb | 10:59 | |
| lucasagomes | the idea is make it useful | 10:59 |
| *** _vs has joined #openstack-ironic | 10:59 | |
| *** _vsaienko has joined #openstack-ironic | 10:59 | |
| sambetts | lucasagomes: we should totally have it pick up if the gate is borked | 11:00 |
| lucasagomes | sambetts, exactly things like that | 11:00 |
| lucasagomes | search for bugs, etc | 11:01 |
| lucasagomes | approved specs and all that | 11:01 |
| sambetts | jroll, dtantsur|brb: I put together a spec for the interface attach detach API | 11:04 |
| *** stendulker has quit IRC | 11:08 | |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: WIP: grenade checking https://review.openstack.org/311109 | 11:16 |
| jroll | vsaienko: vdrok: that adds https://review.openstack.org/#/c/317993/1 | 11:16 |
| *** Goneri has joined #openstack-ironic | 11:20 | |
| *** _vsaienko has quit IRC | 11:24 | |
| *** _vs has quit IRC | 11:24 | |
| vsaienko | jroll I don't think that adding Depends-On in the https://review.openstack.org/#/c/311109/ will work. Since patch already contain depends-on to grenade project Depends-On: I9a1cf8aab959b926f7f853208a4f54658e24e07e | 11:28 |
| openstackgerrit | Kurt Taylor proposed openstack/ironic-specs: Revise third party continuous integration spec for new drivers https://review.openstack.org/314768 | 11:28 |
| vsaienko | I suggest to put both patches in the chain | 11:28 |
| vsaienko | https://review.openstack.org/#/c/315268/ and https://review.openstack.org/#/c/317993/ | 11:29 |
| *** chopmann has quit IRC | 11:29 | |
| *** hoangcx has quit IRC | 11:31 | |
| jroll | vsaienko: it should work fine as long as there's no merge conflict | 11:32 |
| *** anteaya has joined #openstack-ironic | 11:32 | |
| *** nico-afk is now known as nicodemos | 11:34 | |
| vsaienko | jroll: zuul-cloner uses checkout https://github.com/openstack-infra/zuul/blob/33f080a8801dabac5450892b164ca510a8296405/zuul/lib/cloner.py#L180-L181 | 11:35 |
| jroll | vsaienko: oh :( | 11:36 |
| jroll | vsaienko: lemme fix | 11:36 |
| vsaienko | jroll: sure | 11:36 |
| openstackgerrit | Jim Rollenhagen proposed openstack/ironic: WIP: grenade checking https://review.openstack.org/311109 | 11:38 |
| jroll | that should do it | 11:38 |
| *** lucasagomes is now known as lucas-hungry | 11:38 | |
| *** priteau has quit IRC | 11:41 | |
| *** baoli has joined #openstack-ironic | 11:45 | |
| *** baoli_ has joined #openstack-ironic | 11:46 | |
| *** baoli has quit IRC | 11:50 | |
| vsaienko | jroll do we have an etherpad with grenade status? I would like to add all patchese/fixes that need to be merged there | 11:51 |
| jroll | vsaienko: https://etherpad.openstack.org/p/ironic-newton-grenade-whiteboard | 11:52 |
| *** dtantsur|brb is now known as dtantsur | 11:52 | |
| *** priteau has joined #openstack-ironic | 11:52 | |
| anteaya | jroll: hello | 11:54 |
| anteaya | jroll: did you get any follow up from anyone in the scientific working group? | 11:54 |
| *** priteau has quit IRC | 11:57 | |
| jroll | anteaya: yes, there's a post on the ops list, I have yet to respond | 11:57 |
| anteaya | jroll: I must have missed that post, I'll look again | 11:57 |
| anteaya | jroll: thanks | 11:58 |
| jroll | anteaya: [Openstack-operators] [scientific] Ironic Summit recap - ops experiences | 11:58 |
| dtantsur | TIL we have scientific working group | 11:58 |
| * jroll steps away for breakfast | 11:58 | |
| anteaya | jroll: thanks | 11:58 |
| jroll | you're welcome :) | 11:58 |
| anteaya | dtantsur: we do indeed | 11:58 |
| anteaya | :) | 11:58 |
| anteaya | enjoy breakfast | 11:58 |
| *** priteau has joined #openstack-ironic | 12:02 | |
| anteaya | what is BMC? | 12:02 |
| anteaya | bare metal c<something> | 12:02 |
| sambetts | anteaya: https://en.wikipedia.org/wiki/Intelligent_Platform_Management_Interface#Baseboard_management_controller | 12:03 |
| anteaya | thanks | 12:03 |
| *** dprince has joined #openstack-ironic | 12:05 | |
| *** dhellmann has joined #openstack-ironic | 12:06 | |
| *** sacharya has joined #openstack-ironic | 12:09 | |
| *** fellypefca has joined #openstack-ironic | 12:10 | |
| *** trown|outtypewww is now known as trown | 12:12 | |
| *** sacharya has quit IRC | 12:13 | |
| vsaienko | jroll found another issue http://logs.openstack.org/09/311109/22/experimental/gate-grenade-dsvm-ironic/bd11b12/logs/new/screen-n-cpu.txt.gz | 12:13 |
| vsaienko | n-cpu didn't started because nova_compute_driver from Mitaka doesn't work on nova from Newton | 12:14 |
| vsaienko | there was a bug in Newton that fixed issue https://bugs.launchpad.net/ironic/+bug/1574990 | 12:14 |
| openstack | Launchpad bug 1574990 in Ironic "nova-cpu can't start, Unable to load the virtualization driver" [Critical,Fix released] | 12:14 |
| vsaienko | jroll: so probably we have to change compute driver in Mitaka to ironic.IronicDriver | 12:15 |
| *** yonglihe has quit IRC | 12:18 | |
| sambetts | :/ that sounds like something we should be able to configure in the devstack scripts right/ | 12:20 |
| sambetts | ? | 12:20 |
| sambetts | in the local.conf files | 12:20 |
| thiagop | good morning Ironic | 12:23 |
| mgould | morning thiagop | 12:24 |
| *** priteau has quit IRC | 12:25 | |
| *** Goneri has quit IRC | 12:25 | |
| *** priteau has joined #openstack-ironic | 12:26 | |
| *** links has quit IRC | 12:31 | |
| *** lucas-hungry is now known as lucasagomes | 12:32 | |
| *** links has joined #openstack-ironic | 12:35 | |
| *** Goneri has joined #openstack-ironic | 12:37 | |
| openstackgerrit | Jarrod Johnson proposed openstack/pyghmi: Disable futile attempts to correct uncorrectable time https://review.openstack.org/316218 | 12:42 |
| openstackgerrit | Merged openstack/pyghmi: Remove support for py33/py26 https://review.openstack.org/264570 | 12:42 |
| *** itamarl has joined #openstack-ironic | 12:48 | |
| *** joprovost has joined #openstack-ironic | 12:50 | |
| *** daemontool has quit IRC | 12:52 | |
| *** _vsaienko has joined #openstack-ironic | 12:53 | |
| *** _vs has joined #openstack-ironic | 12:53 | |
| jroll | vsaienko: oh, fun :) | 12:54 |
| jroll | vsaienko: so I could backport that quickly, or we could do the from-mitaka thing | 12:54 |
| vdrok | jroll: in theory, we should not change it in mitaka, right? like users who have this setting will be broken after upgrade :( | 12:55 |
| vsaienko | jroll: I agree with vdrok, changing default value in Mitaka may harm uses, that upgrade production clouds | 12:56 |
| vdrok | but changing that in grenade is also bad, we should not fiddle with config during upgrade | 12:56 |
| jroll | well, it's more like, that nova change potentially broke users | 12:57 |
| * jroll trying to find it | 12:57 | |
| vsaienko | I think we may change default value in Mitaka, and add a release note to Newton that during upgrade user have to change nova_compute driver | 12:57 |
| vsaienko | jroll here is the change https://review.openstack.org/309504/ | 12:57 |
| xavierr | gooood morning 1r0n1c | 12:57 |
| jroll | right, so | 12:58 |
| jroll | there's kind of a release note for this: http://docs.openstack.org/releasenotes/nova/unreleased.html#other-notes | 12:58 |
| sambetts | did that go through a deprecation? ... | 12:59 |
| jroll | bah, our mitaka install guide has the bad one http://docs.openstack.org/developer/ironic/mitaka/deploy/install-guide.html#configure-compute-to-use-the-bare-metal-service | 12:59 |
| jroll | sambetts: no clue, it was during summit | 13:00 |
| *** joprovost1 has joined #openstack-ironic | 13:00 | |
| *** rbudden has joined #openstack-ironic | 13:00 | |
| vsaienko | jroll that is why I propose to change default value in Mitaka docs/devstack and add a release note in Newton | 13:00 |
| jroll | vsaienko: well, it would need to be a nova release note | 13:01 |
| *** joprovost has quit IRC | 13:01 | |
| *** joprovost1 is now known as joprovost | 13:01 | |
| jroll | I would do a release note for Nova newton release, and a from-mitaka in nova's grenade bits | 13:01 |
| *** athomas has joined #openstack-ironic | 13:02 | |
| * jroll chats with nova | 13:02 | |
| sambetts | thats still a breaking change right? would stable/mitaka load ironic.IronicDriver? | 13:03 |
| vsaienko | sambetts yes, it would local | 13:03 |
| vsaienko | according to devstack change nova supports long and short driver names starting from 2013 | 13:04 |
| vsaienko | sambetts here is the change made in 2013 that works with short names https://github.com/openstack-dev/devstack/commit/8c032d1635320ad9b5162136a8876cc48e7fa8bd#diff-32175de45f20afcd48b5a65d96cf7aaeR110 | 13:05 |
| jroll | vsaienko: nah, it was before that even: https://github.com/openstack-dev/devstack/commit/8c032d1635320ad9b5162136a8876cc48e7fa8bd#diff-4f76c30de6fd72bd49643dbcf1007a61L1048 | 13:07 |
| sambetts | :/ hmmm ok... so basically there was a deprecation period it just wasn't documented as one | 13:07 |
| jroll | I think it was somewhat accidental | 13:08 |
| *** rloo has joined #openstack-ironic | 13:09 | |
| sambetts | they accidentaly supported long names? | 13:10 |
| vsaienko | so we have 3 options: 1 - revert nova change and deprecate it in Ocata, 2 - change Mitaka value and add a release note, 3 - Change nova.conf during upgrade and add a release note anyway | 13:10 |
| jroll | sambetts: well, the intent was to not allow loading drivers outside of nova.virt | 13:11 |
| jroll | AIUI | 13:11 |
| jroll | however that also breaks loading drivers from inside nova.virt by the full module path | 13:11 |
| *** lindycoder has joined #openstack-ironic | 13:11 | |
| jroll | vsaienko: yeah, I'm bugging the nova people, I like 3 because it reflects the true state of the world | 13:12 |
| sambetts | yeah :/ tbh not sure why they want to prevent people loading compute drivers from out of tree, seems very GPL to me | 13:12 |
| jroll | they're moving to a more... "defined behavior" model | 13:13 |
| *** _vsaienko has quit IRC | 13:13 | |
| *** _vs has quit IRC | 13:13 | |
| sambetts | :/ | 13:14 |
| jroll | sambetts: they've also deprecated out of tree network drivers, iirc | 13:15 |
| *** wajdi has quit IRC | 13:16 | |
| *** links has quit IRC | 13:16 | |
| *** _vsaienko has joined #openstack-ironic | 13:17 | |
| *** _vs has joined #openstack-ironic | 13:17 | |
| *** mbound has quit IRC | 13:17 | |
| sambetts | jroll: this makes it even more important to me that the attach / detach API is landed, plus tbh I'd like to do the same to all other logic in the nova driver, because there'll be no more oh I'll just monkey patch that thing with an out of tree version of the Ironic driver any more ... | 13:17 |
| jlvillal | sambetts, It went wonky again: https://review.openstack.org/#/c/311109/ | 13:18 |
| jroll | sambetts: yep, I hear you | 13:18 |
| sambetts | jlvillal: gah?! Why?! | 13:18 |
| jlvillal | sambetts, Uh, I'm kinda looking at you for that answer ;) | 13:19 |
| sambetts | jlvillal: pinging people who might know more internally to find out why | 13:19 |
| jroll | sambetts: I wonder if there's something about depends-on that makes your ci freak out | 13:19 |
| jroll | I feel like that's the common thing | 13:19 |
| sambetts | ooooh I wonder! its like its detecting the recheck message in the Failure message we're commenting with?! I wonder if thats whats happening! | 13:22 |
| sambetts | so its basiclly rechecking itself or something | 13:22 |
| *** sabeen has joined #openstack-ironic | 13:22 | |
| jlvillal | vsaienko, Very nice to wake up and see that there was a green grenade job :) Thanks! | 13:22 |
| * sambetts does some digging | 13:22 | |
| jroll | hah | 13:23 |
| sambetts | only a hypothosis right now, but thats my theory | 13:23 |
| vsaienko | jlvillal, we still need to fix a lot of thing here | 13:24 |
| jlvillal | vsaienko, Yep. But we have had HUGE progress this week. | 13:25 |
| vsaienko | jroll: looks like https://review.openstack.org/#/c/317993/1/upgrade-tempest didn't help | 13:25 |
| jlvillal | vsaienko, I think we will need this merged: https://review.openstack.org/317766 | 13:26 |
| *** nathan-h has joined #openstack-ironic | 13:26 | |
| jlvillal | At least I ran into an issue with that value set in my local testing. | 13:26 |
| jroll | vsaienko: oh? :/ | 13:26 |
| NobodyCam | Good morning Ironicers... | 13:27 |
| vsaienko | jroll: tempest doesn't know about ir-cond or ir-api https://github.com/openstack-dev/devstack/blob/59ed73a3d46402f01021794ecc565aa8945104c1/lib/tempest#L500 | 13:27 |
| NobodyCam | :) | 13:27 |
| *** _vs has quit IRC | 13:27 | |
| dtantsur | morning NobodyCam | 13:27 |
| vsaienko | we have to add ironic to enabled_services I think | 13:27 |
| *** _vsaienko has quit IRC | 13:28 | |
| jroll | vsaienko: oh right, we need to add that here https://github.com/openstack/ironic/blob/master/devstack/upgrade/settings | 13:28 |
| vsaienko | jroll right | 13:28 |
| NobodyCam | morning dtantsur, jroll, vsaienko :) | 13:28 |
| jlvillal | vsaienko, jroll: That patches I thought we needed. But mine doesn't pass yet. https://review.openstack.org/#/c/317751/ | 13:29 |
| vsaienko | Morning NobodyCam | 13:29 |
| NobodyCam | and jlvillal too :) | 13:29 |
| *** _vsaienko has joined #openstack-ironic | 13:29 | |
| *** _vs has joined #openstack-ironic | 13:29 | |
| jlvillal | Good morning vsaienko NobodyCam jroll dtantsur sambetts lucasagomes rloo vdrok BadCub and everyone else :) | 13:30 |
| *** mjturek1 has quit IRC | 13:30 | |
| dtantsur | morning jlvillal | 13:30 |
| jroll | jlvillal: seems there is more :) | 13:30 |
| jroll | jlvillal: see the list at https://review.openstack.org/#/c/311109/ | 13:30 |
| jroll | I know I've seen that apache issue | 13:30 |
| jlvillal | jroll, thanks | 13:30 |
| jroll | jlvillal: thanks vsaienko :) | 13:30 |
| jroll | thank* | 13:30 |
| jlvillal | :) | 13:31 |
| lucasagomes | #pixiesay morning NobodyCam jroll :-) | 13:31 |
| PixieBoots | ʕ•͡ᴥ•ʔ: morning NobodyCam jroll :-) | 13:31 |
| lucasagomes | jlvillal, * | 13:31 |
| * lucasagomes #fail | 13:31 | |
| vsaienko | jlvillal you have missed https://review.openstack.org/#/c/317139 | 13:31 |
| jlvillal | vsaienko, vdrok I wouldn't be opposed to a new patch to replace 311109. Since it takes so long to load in my browser :) | 13:32 |
| jroll | heh ++ | 13:32 |
| NobodyCam | lucasagomes: omg | 13:32 |
| jlvillal | vsaienko, I have that one | 13:32 |
| NobodyCam | PixieBoots: help | 13:33 |
| vsaienko | jlvillal, it should be in the chain with https://review.openstack.org/#/c/317751/ | 13:33 |
| NobodyCam | #PixieBoots help | 13:33 |
| NobodyCam | "p | 13:33 |
| lucasagomes | NobodyCam, it does not have it yet | 13:33 |
| rloo | morning jlvillal, NobodyCam, jroll, PixieBoots, vsaienko, lucasagomes, dtantsur, sambetts and all you ironic'ers out there that have been quietly working the past few minutes :) | 13:34 |
| jlvillal | vsaienko, I see https://review.openstack.org/#/c/317139 brought in by https://review.openstack.org/#/c/317751/ | 13:34 |
| lucasagomes | NobodyCam, it only has 2 commands #pixiesay and #failgraph <test name [test name] ...> | 13:34 |
| vsaienko | Morning rloo | 13:34 |
| NobodyCam | morning rloo, and lucasagomes :) | 13:34 |
| jlvillal | vsaienko, Maybe you aren't scrolling the commit message? | 13:34 |
| lucasagomes | rloo, morning | 13:35 |
| dtantsur | morning rloo | 13:35 |
| NobodyCam | lucasagomes: very cool! | 13:35 |
| *** _vsaienko has quit IRC | 13:35 | |
| vsaienko | jlvillal both patches are to the ironic project, Depends-On will not work I think | 13:35 |
| lucasagomes | NobodyCam, I did it yesterday night, so... it's very frsh | 13:35 |
| vsaienko | jlvillal they should be in the same chain | 13:36 |
| *** priteau has quit IRC | 13:36 | |
| *** _vs has quit IRC | 13:36 | |
| vsaienko | jroll who is gonna to update https://github.com/openstack/ironic/blob/master/devstack/upgrade/settings#L16 ? | 13:36 |
| jroll | vsaienko: I can if you are not | 13:37 |
| jroll | vsaienko: which patch should I stack it on :) | 13:38 |
| jlvillal | vsaienko, Zuul seems to understand http://status.openstack.org/zuul/ 317751 | 13:38 |
| vsaienko | jroll, I will do it | 13:38 |
| jroll | vsaienko: ok, thanks | 13:38 |
| *** athomas has quit IRC | 13:38 | |
| *** athomas has joined #openstack-ironic | 13:39 | |
| jlvillal | jroll, your vote on the grenade patch would be good: https://review.openstack.org/#/c/317076/ | 13:39 |
| *** _vsaienko has joined #openstack-ironic | 13:39 | |
| *** _vs has joined #openstack-ironic | 13:39 | |
| jlvillal | vsaienko, and I are co-authors, so we shouldn't vote | 13:39 |
| vdrok | Morning NobodyCam jlvillal and rloo :) | 13:39 |
| NobodyCam | morning vdrok :) | 13:39 |
| rloo | hi vdrok! | 13:39 |
| *** ChrisAusten has joined #openstack-ironic | 13:40 | |
| *** cdearborn has joined #openstack-ironic | 13:41 | |
| openstackgerrit | Ruby Loo proposed openstack/ironic: Remove link to Liberty configs https://review.openstack.org/318092 | 13:41 |
| jroll | jlvillal: I've voted, probably not the way you hoped :) | 13:41 |
| jlvillal | Argh! ;) | 13:41 |
| *** mtanino has joined #openstack-ironic | 13:43 | |
| mat128 | jroll, jlvillal I suggested a different wording that avoids the $1 altogether :) | 13:44 |
| jlvillal | jroll, I responded. I was copying style I had seen in Grenade. | 13:44 |
| vsaienko | jroll, I would like to add a patch to upgrade.sh that will update nova config with new compute_driver | 13:44 |
| *** mgoddard_ has joined #openstack-ironic | 13:45 | |
| jlvillal | mat128, There are multiple phases though. pre_create, create, verify, verify_noapi, etc... | 13:45 |
| jroll | vsaienko: that would go in the from-mitaka thing, right? | 13:46 |
| vsaienko | right | 13:47 |
| vsaienko | jrill after the stop_nova https://github.com/openstack/ironic/blob/master/devstack/upgrade/upgrade.sh#L79 | 13:47 |
| vsaienko | sorry, jroll | 13:47 |
| *** mgoddard has quit IRC | 13:48 | |
| *** ametts has joined #openstack-ironic | 13:49 | |
| jroll | vsaienko: and I think that should go in the nova from-mitaka script as that is a breakage in nova | 13:49 |
| sambetts | jlvillal: It seems I am hitting this https://storyboard.openstack.org/#!/story/2000445 | 13:49 |
| jlvillal | jroll, Updated https://review.openstack.org/#/c/317076/ | 13:50 |
| jlvillal | sambetts, Ah cool. You know why. | 13:50 |
| sambetts | jlvillal: Not really sure, something to do with when all my jenkins slaves go dead or something :/ I need to update to get the fix for it I think | 13:51 |
| vsaienko | jroll what is nova from-mitaka script? | 13:53 |
| jroll | vsaienko: something like https://github.com/openstack-dev/grenade/blob/master/projects/60_nova/from-liberty/upgrade-nova | 13:53 |
| jroll | but for mitaka | 13:53 |
| sambetts | jlvillal: oh man... the fix for that issue is merged 5 months ago but their last release was 11 months ago ... | 13:54 |
| openstackgerrit | Andre keedy proposed openstack/ironic-inspector: Replace node data via uuid/data/unprocessed API https://review.openstack.org/308379 | 13:54 |
| jroll | vsaienko: see also the bottom of http://docs.openstack.org/developer/grenade/readme.html#theory-of-upgrade where it says "Grenade supports per release specific upgrade scripts (from-juno, from-kilo)." | 13:54 |
| jlvillal | sambetts, :( | 13:54 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: WIP: grenade checking https://review.openstack.org/311109 | 13:55 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Add ironic to enabled_services https://review.openstack.org/318112 | 13:55 |
| *** mjturek1 has joined #openstack-ironic | 13:57 | |
| *** absubram has quit IRC | 13:59 | |
| anteaya | so I'm reading http://jroll.ghost.io/newton-summit-recap/ as I have to give a presentation tonight to a user group on the austin summit | 14:00 |
| anteaya | what is a pizza box hardware type? | 14:00 |
| mat128 | generic brandless machine | 14:00 |
| mat128 | rackmount | 14:00 |
| *** mgoddard_ has quit IRC | 14:01 | |
| *** mgoddard has joined #openstack-ironic | 14:01 | |
| anteaya | mat128: thank you | 14:01 |
| anteaya | I'm personally very excited about the thing resource | 14:01 |
| sambetts | anteaya: pizza box: http://i131.twenga.com/computers/rackmount-server-case/generic-rack-mountable-server-tp_5377784842778054130f.jpg | 14:01 |
| sambetts | :D | 14:01 |
| anteaya | sambetts: ah thank you | 14:01 |
| jroll | anteaya: to be clear, I was using "pizza box" as a strawman, it isn't something we plan to have :) | 14:02 |
| anteaya | very good, thank you | 14:02 |
| *** vishwanathj has joined #openstack-ironic | 14:03 | |
| anteaya | I'm not sure how in depth I'll get, first time addressing this crowd and the title I was given was overview of the Austin Summit | 14:03 |
| anteaya | I have a lot of ground to cover | 14:03 |
| anteaya | think I'll see what the crowd is interested in and pay attention to those topics if that is feasible | 14:04 |
| rloo | good luck anteaya; I'm sure you'll wow them and convince them all to go to Barcelona :) | 14:04 |
| anteaya | rloo: thanks ruby, I appreciate your vote of confidence | 14:05 |
| anteaya | I'll certainly be telling them about Barcelona | 14:05 |
| anteaya | but to be honest my biggest takeaway personally from Austin was there were far too many people | 14:06 |
| anteaya | I wasn't sick and still had to sleep for 2 weeks after returning home | 14:06 |
| rloo | anteaya: interesting. I don't think that I found it any bigger than Tokyo's. | 14:06 |
| anteaya | oh okay, good to know | 14:07 |
| rloo | anteaya: but then, I'm not a mover/shaker/social-creature like you are :D I stuck mostly with the same developers. | 14:07 |
| anteaya | I found there were many conversations I never got to have | 14:07 |
| anteaya | as I kept getting interrupted | 14:07 |
| rloo | anteaya: yes, that's cuz you are popular. | 14:07 |
| anteaya | rloo: that has its benefits, staying with the same folks | 14:07 |
| anteaya | rloo: not sure I would go that far :) | 14:08 |
| anteaya | but thank you :) | 14:08 |
| rloo | anteaya: ha ha. You are well known anyway! | 14:08 |
| anteaya | that I will accept | 14:08 |
| anteaya | :) | 14:08 |
| *** sacharya has joined #openstack-ironic | 14:09 | |
| vsaienko | jroll once you have a time please have look https://review.openstack.org/#/c/318119/ | 14:09 |
| *** links has joined #openstack-ironic | 14:10 | |
| jroll | vsaienko: I don't remember co-authoring that :) | 14:11 |
| *** rloo has quit IRC | 14:11 | |
| jroll | vsaienko: +1 though | 14:11 |
| *** rloo has joined #openstack-ironic | 14:11 | |
| *** sacharya has quit IRC | 14:14 | |
| *** sdake has joined #openstack-ironic | 14:14 | |
| *** mbound has joined #openstack-ironic | 14:17 | |
| openstackgerrit | Merged openstack/ironic: Fix shutdown.sh & upgrade.sh for grenade https://review.openstack.org/317139 | 14:19 |
| sambetts | jlvillal: seems like the guy who manages my jenkins server did a reboot to update, and because of that bug it went into an infinite loop ... | 14:19 |
| *** mbound has quit IRC | 14:23 | |
| *** wajdi has joined #openstack-ironic | 14:23 | |
| sambetts | jlvillal: we'll likely have to bring down zuul first then jenkins then restart jenkins and then zuul to get around the problem I think :/ | 14:24 |
| *** Guest92292 is now known as mgagne | 14:25 | |
| *** mgagne has quit IRC | 14:25 | |
| *** mgagne has joined #openstack-ironic | 14:25 | |
| sambetts | jlvillal: but it shouldn't happen unless the jenkins/gearman is taken away | 14:25 |
| jlvillal | sambetts, Glad you figured it out :) | 14:26 |
| *** sdake has quit IRC | 14:26 | |
| jlvillal | sambetts, Thanks for investigating! | 14:26 |
| *** sdake has joined #openstack-ironic | 14:27 | |
| wajdi | morning all | 14:28 |
| *** daemontool has joined #openstack-ironic | 14:28 | |
| *** mjturek1 has quit IRC | 14:31 | |
| *** _vsaienko has quit IRC | 14:31 | |
| *** _vs has quit IRC | 14:31 | |
| NobodyCam | morning wajdi | 14:32 |
| *** fellypefca has quit IRC | 14:32 | |
| *** marios is now known as marios|call | 14:35 | |
| jlvillal | vsaienko, +2 on https://review.openstack.org/#/c/318112 Because I didn't feel like I co-authored it :) | 14:37 |
| *** links has quit IRC | 14:39 | |
| *** _vs has joined #openstack-ironic | 14:41 | |
| *** _vsaienko has joined #openstack-ironic | 14:41 | |
| *** sdake_ has joined #openstack-ironic | 14:42 | |
| *** sdake has quit IRC | 14:43 | |
| *** _vs has quit IRC | 14:43 | |
| *** _vsaienko has quit IRC | 14:43 | |
| *** _vsaienko has joined #openstack-ironic | 14:44 | |
| *** _vs has joined #openstack-ironic | 14:44 | |
| * rloo shakes fist at jroll for changing ironic to the semvered releases. makes rolling upgrades much harder. | 14:45 | |
| openstackgerrit | Andre keedy proposed openstack/ironic-inspector: Replace node data via uuid/data/unprocessed API https://review.openstack.org/308379 | 14:45 |
| *** mjura has quit IRC | 14:47 | |
| *** _vsaienko has quit IRC | 14:47 | |
| *** _vs has quit IRC | 14:47 | |
| *** mjura has joined #openstack-ironic | 14:47 | |
| *** ChrisAusten has quit IRC | 14:50 | |
| *** ChrisAusten has joined #openstack-ironic | 14:51 | |
| *** _vsaienko has joined #openstack-ironic | 14:52 | |
| *** _vs has joined #openstack-ironic | 14:52 | |
| *** xavierr is now known as xavierr_afk | 14:58 | |
| mat128 | rloo: are you pointing to sha1s directly or tags? | 14:59 |
| mat128 | rloo: trying to understand how semver impacts you | 14:59 |
| rloo | mat128: ? oh. | 14:59 |
| rloo | so... well, take a look at tan's comment. sec. | 14:59 |
| rloo | mat128: https://review.openstack.org/#/c/299245/7/specs/approved/support-rolling-upgrade.rst | 14:59 |
| *** sacharya has joined #openstack-ironic | 14:59 | |
| rloo | mat128: line 75 | 15:00 |
| rloo | mat128: the question is how to handle changes like tan mentioned. | 15:00 |
| rloo | mat128: if we only support rolling upgrades between named releases, it'd be much easier. | 15:00 |
| mat128 | Ahh | 15:00 |
| rloo | mat128: or if we had some definition/policy wrt sem-versioning and when/what the versioning might be when something is deleted. | 15:01 |
| rloo | mat128: maybe. i'm not totally sure. | 15:01 |
| jroll | rloo: well, if we only deprecate/remove things across cycle boundaries, then the non-cycle-boundary is fine, right? | 15:01 |
| mat128 | if we follow semver to the letter, major changes when something is gone (was deprecated for X time, now gone) | 15:01 |
| mat128 | jroll: should be | 15:02 |
| rloo | jroll: yes, the non-cycle-boundary is fine. but how do you word that wrt semversion'd releases. major.minor | 15:02 |
| *** jistr is now known as jistr|mtg | 15:02 | |
| mat128 | rloo: it probably means every named release is a major | 15:02 |
| jroll | rloo: "we support upgrades between two named releases and/or anything in between two named releases" | 15:03 |
| mat128 | but there can be major in-between | 15:03 |
| jroll | we'll never deprecate *and* remove something without a named release between | 15:03 |
| *** _vsaienko has quit IRC | 15:03 | |
| *** _vs has quit IRC | 15:03 | |
| rloo | jroll: did you see tan's example? | 15:03 |
| mat128 | rloo: Tan's case only makes for more cases to handle | 15:04 |
| jroll | rloo: yes, we will not drop something without a major version bump, right? | 15:04 |
| mat128 | 5.0 -> 5.1 has to work, but also 5.0 -> 6.0 | 15:04 |
| rloo | jroll: so with tan's example, what if 5.1 is during mitaka, add node.fake; 5.2 is during newton, delete node.extra? | 15:04 |
| mat128 | in the end, commit1 -> commit2 should work too | 15:04 |
| rloo | jroll: OH, can we cay that. we won't drop anything w/o a major version bump? that would help. | 15:05 |
| mat128 | 5.2 turns into 6.0 if you delete node.extra | 15:05 |
| *** Mr_T has left #openstack-ironic | 15:05 | |
| jroll | rloo: yeah, exactly | 15:05 |
| mat128 | every non backwards compatible change bumps the major | 15:05 |
| jroll | ^^^ | 15:05 |
| mat128 | like it or not (there are 2 schools here) :) | 15:05 |
| mat128 | major means non-backwards compatible vs major means "new features" | 15:05 |
| jroll | rloo: shall I comment with that? | 15:05 |
| mat128 | I'm from the former, we introduce new features all the time | 15:05 |
| rloo | jroll: yes please. comment. | 15:06 |
| jroll | k, on it | 15:06 |
| *** Sukhdev has joined #openstack-ironic | 15:06 | |
| sambetts | mat128: I also prefer the former approach | 15:06 |
| rloo | jroll: wish we had some documentation page that describes the semver but it will be documented in this spec anyway :) | 15:06 |
| mat128 | rloo: http://semver.org/ | 15:06 |
| mat128 | then you have to define what is *your* API | 15:06 |
| lucasagomes | can I get some eyes at https://review.openstack.org/#/c/317538/ ? | 15:06 |
| mat128 | I think we wrote that somewhere | 15:07 |
| mat128 | IIRC, driver APi was part of it | 15:07 |
| mat128 | not only http api | 15:07 |
| jroll | what mat128 said :) | 15:07 |
| rloo | mat128: i know we talked about it but don't know that it is documented anywhere. | 15:07 |
| rloo | lucasagomes: is it urgent? my head is rolling in upgrades now. | 15:08 |
| mat128 | we should be documenting what *our* API is, because this will help identifying commits that might have impacts | 15:08 |
| lucasagomes | rloo, nop not urgent | 15:08 |
| mat128 | then it's a matter of tagging them in the commit message and the tools do the rest ;) | 15:08 |
| mat128 | "API-impact" or something along those lines | 15:08 |
| *** jaybeale has joined #openstack-ironic | 15:09 | |
| mat128 | lucasagomes: what was the previous way, before caching? | 15:09 |
| mat128 | lucasagomes: method call? | 15:09 |
| jroll | rloo: posted that comment | 15:10 |
| rloo | jroll: thx! | 15:10 |
| *** e0ne has quit IRC | 15:10 | |
| lucasagomes | mat128, before caching (merged) it gets it from the kernel cmdline. That's because of the bash ramdisk, that didn't have an API so that was the way we passed parameters from Ironic to the ramdisk | 15:11 |
| mat128 | we passed all node information by cmdline? | 15:11 |
| mat128 | or anything required via cmdline, is that it? | 15:11 |
| *** e0ne has joined #openstack-ironic | 15:11 | |
| mat128 | there was no way to get node details from a hardware manager? | 15:11 |
| lucasagomes | mat128, anything required via the cmdline (e.g root device hints) | 15:12 |
| mat128 | lucasagomes: ouch | 15:12 |
| mat128 | lucasagomes: I only have a question w/r/t the upgrade nodes then | 15:13 |
| lucasagomes | mat128, ok | 15:14 |
| lucasagomes | one thing is the bash ramdisk is now gone, we do not support it anymore | 15:14 |
| mat128 | lucasagomes: do you want to mean "as long as you have heartbeat going on" ? | 15:14 |
| mat128 | thus why in "standalone mode" you won't have node information available? | 15:14 |
| *** athomas has quit IRC | 15:16 | |
| *** nathan-h has quit IRC | 15:17 | |
| lucasagomes | mat128, standalone mode basically means that IPA will not talk to the Ironic API | 15:17 |
| lucasagomes | so basically there's no lookup() in standalone | 15:17 |
| lucasagomes | (therefore cache is not invoked when in standalone) | 15:17 |
| *** athomas has joined #openstack-ironic | 15:18 | |
| mat128 | lucasagomes: got it :) | 15:19 |
| *** marios|call is now known as marios | 15:19 | |
| mat128 | lucasagomes: Didn't think of any use case for standalone IPA, but I guess it's always possible | 15:19 |
| *** jistr|mtg is now known as jistr | 15:19 | |
| lucasagomes | yeah, I think jroll JayF uses it!? | 15:19 |
| mat128 | lucasagomes: with your other patch, it means you cannot set root device except when using Ironic | 15:19 |
| lucasagomes | for the always running nodes I believe | 15:19 |
| jroll | lucasagomes: we use it for testing, not in prod | 15:19 |
| mat128 | lucasagomes: Thought it was always used with Ironic | 15:19 |
| jroll | or rather, it was put there for testing | 15:20 |
| lucasagomes | mat128, yeah... that's a plumbing patch. Since we want to have a more complex root device hints structure (to support operators such as >= == etc...) | 15:20 |
| jroll | I'm commenting on this btw, I thought of a couple questions but now forgot the second :P | 15:20 |
| lucasagomes | we can't use the kernel cmdline anymore, but since the hints are always in the node we can just get it from there | 15:20 |
| mat128 | jroll: we have integration testing around IPA with a simply mocked "lookup" and "heartbeat" responses | 15:20 |
| mat128 | using a simple web server :) | 15:20 |
| jroll | mat128: neat | 15:20 |
| lucasagomes | jroll, gotcha | 15:21 |
| jroll | this was mainly for local testing | 15:21 |
| *** mjura has quit IRC | 15:21 | |
| *** athomas has quit IRC | 15:23 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-python-agent: Cache node object inside hw manager instances https://review.openstack.org/317538 | 15:23 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic-python-agent: Get root device hints from the node object https://review.openstack.org/296539 | 15:23 |
| * lucasagomes fixes nit spotted by sam | 15:23 | |
| *** dprince has quit IRC | 15:24 | |
| JayF | lucasagomes: yeah, what jroll said; we basically use standalone mode + a script to run clean steps to 'emulate' cleaning in our the lab we have with shiny new secret hardware | 15:24 |
| *** rcernin has quit IRC | 15:24 | |
| jroll | JayF: yeah so that's my concern if things start depending on self.node | 15:25 |
| JayF | I mean, that safely fell back | 15:25 |
| lucasagomes | JayF, jroll gotcha... wouldn't be a bad idea to have some tempest against IPA's api in gate either | 15:25 |
| JayF | if you remember our downstream clean step running script already has to pass in a faked node | 15:25 |
| JayF | because node is passed to clean steps | 15:25 |
| JayF | the patch I +2'd last night from lucasagomes appeared to handle the "where is my node?" case very reasonably | 15:26 |
| jroll | right, but it doesn't work out here if things start depending on it | 15:26 |
| jroll | not worried about that one I guess | 15:26 |
| JayF | jroll: your comment was on the wrong patchset. you looked at 4 as well? | 15:26 |
| jroll | JayF: 4 just changes a release note | 15:27 |
| JayF | oh; that's not the patch I thought it was | 15:27 |
| JayF | https://review.openstack.org/#/c/296539 was what I thought context was | 15:27 |
| lucasagomes | jroll, yeah... we don't necessarily needs to rely on the node object to be there, the next patch I actually test see if it's None | 15:28 |
| jroll | lucasagomes: right, I guess I'm just worried about the future when someone decides to clean up the node argument to all these things | 15:28 |
| lucasagomes | not sure what would be the best way to handle that tho, because the way it's now, if I extend a method to pass the node as parameter | 15:28 |
| lucasagomes | I may break out of tree hardware managers | 15:28 |
| jroll | right | 15:29 |
| jroll | this case is fine | 15:29 |
| JayF | I mean, isn't this the same as any other interface? It's OK to do new things as long as you don't break the API | 15:29 |
| JayF | this doesn't break the API, only makes it ... easier to do the wrong thing in the future | 15:29 |
| lucasagomes | jroll, yeah if someone cleans it up, it may break out of tree stuff indeed | 15:29 |
| JayF | but that doesn't make it wrong on the face, it just means we have to be vigilant | 15:29 |
| JayF | right? | 15:29 |
| mat128 | JayF: Yes, someone might depend on self.node being there and not handle the standalone case | 15:29 |
| jroll | sure | 15:30 |
| mat128 | Maybe we need to test it in standalone in the gate? | 15:30 |
| JayF | mat128: if we miss it in review, that's right | 15:30 |
| jroll | just makes it easier to shoot yourself in the foot | 15:30 |
| JayF | I care less about testing standalone | 15:30 |
| jroll | but I think in this case the gun is an airsoft gun, not a shotgun | 15:30 |
| JayF | and more about changing function signatures in HardwareManager | 15:30 |
| JayF | or GenericHardwareManager | 15:30 |
| jroll | right, that's besides the point | 15:30 |
| mat128 | so lucasagomes' approach is right | 15:30 |
| jroll | I'm not saying this is bad or wrong or -1'ing it | 15:30 |
| jroll | I'm just saying some words we need to keep in mind here | 15:31 |
| lucasagomes | jroll, agreed. Yeah, there are flaws in that approach, we should make sure that methods accessing self.node should consider the case where it's not loaded | 15:31 |
| mat128 | crazy thoughts: what if we had a different base class having .node, instead of checking for it? | 15:32 |
| mat128 | sure means more stuff to manage, but explicitely clearer that you have .node | 15:33 |
| JayF | mat128: that's going to be a bear for transition, but I like the idea for the future | 15:33 |
| mat128 | oh crap, same classes in both standalone/non-standalone | 15:33 |
| *** Sukhdev has quit IRC | 15:34 | |
| JayF | I mean, I'm just saying *today* there are places agent API expects a node object | 15:34 |
| *** mbound has joined #openstack-ironic | 15:34 | |
| JayF | and breaks badly if it doesn't get it | 15:34 |
| JayF | I should check to make sure all of those aren't in my downstream hwm real quick, lol | 15:34 |
| mat128 | haha | 15:34 |
| JayF | nope | 15:34 |
| JayF | https://github.com/openstack/ironic-python-agent/blob/master/ironic_python_agent/hardware.py#L243 | 15:34 |
| JayF | clean steps totally get node passed in, I thuoght so | 15:34 |
| JayF | if you don't provide a node via the API to call the clean step, it fails | 15:35 |
| JayF | and those methods aren't "missing-a-node safe" either | 15:35 |
| * JayF either doesn't get it or the cat is already out of the bag | 15:35 | |
| mat128 | so erase_devices cannot be called in standalone | 15:36 |
| jroll | JayF: right, so with standalone you can pass a node object in the run clean step api call | 15:36 |
| mat128 | you pass it a node as an argument, so it might work | 15:36 |
| mat128 | it's not relying on self.node | 15:36 |
| jroll | if you replaced those with self.node, you would have to mock a heartbeat response instead of running standalone | 15:36 |
| JayF | ooooh | 15:36 |
| JayF | so passing it in = it can be provided in the api and mocked in standalone | 15:36 |
| JayF | pulled in via lookup/hb = doesn't exist at all in standalone | 15:37 |
| *** dmk0202 has quit IRC | 15:37 | |
| mat128 | https://github.com/openstack/ironic-python-agent/blob/6da6ace3840d56c7145ddf528bbdcbb813fc6ce2/ironic_python_agent/cmd/agent.py#L100 | 15:37 |
| mat128 | JayF: that's correct | 15:37 |
| *** dmk0202 has joined #openstack-ironic | 15:37 | |
| mat128 | standalone clearly states for debugging only | 15:37 |
| JayF | Yeah, honestly doesn't hurt my feelings that much even if it broke standalone, because it's in a hwm | 15:37 |
| JayF | if you really needed that to work in standalone, you could load an alternative hwm | 15:38 |
| mat128 | would it be better to have an empty node instead of checking for it? | 15:38 |
| JayF | hmmm | 15:38 |
| JayF | I don't think so | 15:38 |
| mat128 | then node is officially part of HardwareManager, just unset if you happen to have never reached Ironic | 15:39 |
| mat128 | (oh, new fun cases: not running standalone but ir-api not reachable / not been reached yet) | 15:39 |
| mat128 | just thinking about those race conditions Dimtry had with the networking/inspector | 15:40 |
| JayF | oh interesting | 15:40 |
| JayF | you can't use node in evaluate_hardware_support() Then | 15:40 |
| openstackgerrit | Ruby Loo proposed openstack/ironic-specs: follows cycle-with-intermediary release model https://review.openstack.org/318188 | 15:41 |
| *** thrash is now known as thrash|f00dz | 15:43 | |
| mat128 | JayF: hmm | 15:43 |
| jroll | posted some more comments | 15:43 |
| JayF | mat128: can cache_node_object be taken off the HardwareManager class? | 15:43 |
| jroll | JayF: one thing to think of: long-running agents, this will get stale | 15:43 |
| mat128 | jroll: updated with every heartbeat the way lucasagomes designed it | 15:44 |
| jroll | oh is it? | 15:44 |
| mat128 | are you not heartbeating long-running agents? | 15:44 |
| * lucasagomes reads | 15:44 | |
| jroll | no it isn't | 15:44 |
| jroll | it's only updated on lookup | 15:44 |
| lucasagomes | yeah it's only set once | 15:44 |
| jroll | and yes, we are heartbeating, that's why it's called heartbeat :P | 15:44 |
| mat128 | jroll: you are absolutely right, I missed that | 15:45 |
| mat128 | it will get stale | 15:45 |
| jroll | which isn't a problem today, upstream | 15:45 |
| jroll | but, ya know. | 15:45 |
| JayF | I mean honestly, even thoguht we don't "support" long running agents | 15:46 |
| JayF | we do support it | 15:46 |
| JayF | power on a node in available when using static, external dhcp | 15:46 |
| JayF | and ironic will happily later deploy to that already running agent | 15:46 |
| mat128 | yup | 15:46 |
| jroll | yeah | 15:46 |
| mat128 | except if your driver issues a REBOOT :) | 15:46 |
| mat128 | https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/agent.py#L289 | 15:47 |
| mat128 | ;) | 15:47 |
| jroll | hehe | 15:48 |
| jroll | I think all drivers do that | 15:48 |
| jroll | https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/iscsi_deploy.py#L602 | 15:48 |
| mat128 | yeah | 15:48 |
| jlvillal | vsaienko: vdrok: jroll: Is https://review.openstack.org/#/c/315422/ (the tempest smoke patch) ready to be proposed? Should we clean up the commit message (assumes yes, if ready to be proposed) | 15:49 |
| mat128 | back to the stale cache problem | 15:50 |
| mat128 | maybe heartbeat could return node details? | 15:50 |
| *** dmk0202 has quit IRC | 15:51 | |
| vsaienko | jlvillal, at the moment the jobs are running, I will update it when they finished | 15:51 |
| jroll | mat128: then you end up dispatching to managers every heartbeat, mehhhhh | 15:51 |
| jroll | (which is fine but sigh) | 15:51 |
| mat128 | maybe managers could call something to get node information | 15:51 |
| mat128 | so you wont have to dispatch to them, only make the node information cached somewhere else | 15:52 |
| jlvillal | vsaienko: Thanks | 15:52 |
| jroll | if we just do the auth thing properly they could get it on demand from the api :) | 15:52 |
| mat128 | one more thing to let through :S | 15:52 |
| JayF | jroll: mat128: I actually commented that I don't want it being dispatched to managers | 15:52 |
| lucasagomes | jroll, yeah that would facilitate indeed | 15:52 |
| JayF | jroll: mat128: That even if it has to be on the HWM class, I don't think it should be dispatched | 15:52 |
| *** moshele has quit IRC | 15:53 | |
| *** itamarl has quit IRC | 15:54 | |
| *** penick has joined #openstack-ironic | 15:55 | |
| *** Sukhdev has joined #openstack-ironic | 15:55 | |
| *** MattMan has quit IRC | 15:55 | |
| *** MattMan has joined #openstack-ironic | 15:56 | |
| *** mjturek1 has joined #openstack-ironic | 15:57 | |
| *** mjturek1 has left #openstack-ironic | 15:58 | |
| *** mjturek1 has joined #openstack-ironic | 15:58 | |
| *** mbound has quit IRC | 16:00 | |
| *** jistr has quit IRC | 16:00 | |
| JayF | lucasagomes: mat128: Thought harder, and the code in gerrit is broken as it sits with multiple hwms | 16:00 |
| JayF | lucasagomes: mat128: and other alternatives are kinda gross; I put the details on the review | 16:01 |
| *** mgoddard_ has joined #openstack-ironic | 16:01 | |
| lucasagomes | JayF, thanks a lot, will take a look asap | 16:03 |
| jlvillal | vsaienko: Do you have a different version of: https://review.openstack.org/#/c/317082/ ? | 16:03 |
| jlvillal | vsaienko: It would be good to update that patch with a better commit message and explain things in the code. I will admit I do not understand why that is needed. | 16:04 |
| jlvillal | I know it is needed. But I don't understand exactly what it does and why we couldn't use the pre-existing network. | 16:04 |
| *** mgoddard has quit IRC | 16:04 | |
| jlvillal | Possibly others maybe as clueless as me :) | 16:05 |
| *** ifarkas has quit IRC | 16:05 | |
| *** kromanenko_ has quit IRC | 16:08 | |
| *** Romanenko_K has quit IRC | 16:08 | |
| *** kromanenko__ has quit IRC | 16:08 | |
| *** mbound has joined #openstack-ironic | 16:11 | |
| *** mbound has quit IRC | 16:13 | |
| *** mbound has joined #openstack-ironic | 16:13 | |
| *** lindycoder has quit IRC | 16:13 | |
| *** ijw has joined #openstack-ironic | 16:16 | |
| *** anteaya has quit IRC | 16:21 | |
| lucasagomes | JayF, mat128 thanks for the discussion... I will have to think how to implement that differently then | 16:22 |
| *** afaranha_ has quit IRC | 16:23 | |
| jlvillal | FYI: QA meeting with discussion on Grenade in 35 minutes (1700 UTC) over on #openstack-meeting. No audio bridge this time, unless people really want it. | 16:25 |
| *** thrash|f00dz is now known as thrash | 16:30 | |
| openstackgerrit | Merged openstack/ironic: Add ironic to enabled_services https://review.openstack.org/318112 | 16:30 |
| openstackgerrit | VIncent S. Cojot proposed openstack/ironic: Makes the vbox driver work on headless VirtualBox machines https://review.openstack.org/311278 | 16:30 |
| *** ChubYann has joined #openstack-ironic | 16:30 | |
| *** sdake_ has quit IRC | 16:31 | |
| *** sdake has joined #openstack-ironic | 16:31 | |
| *** swatson1 has joined #openstack-ironic | 16:40 | |
| *** swatson1 is now known as swatson | 16:41 | |
| *** moshele has joined #openstack-ironic | 16:42 | |
| lucasagomes | JayF, mat128 jroll the more I think about how to cache it the worst it gets heh... I'm afraid that the only way we can guarantee that the node object is updated is by passing it as a parameter (but that may break people) | 16:46 |
| JayF | not may | 16:47 |
| JayF | I can attest it will break me | 16:47 |
| lucasagomes | yeah | 16:47 |
| openstackgerrit | Merged openstack/ironic: Add require_exclusive_lock decorators to conductor methods https://review.openstack.org/295734 | 16:48 |
| JayF | lucasagomes: this is all for get_os_install_device, right? | 16:48 |
| lucasagomes | JayF, another way would be call cache_node_object() before calling get_os_install_device | 16:49 |
| lucasagomes | JayF, initially yes | 16:49 |
| lucasagomes | potentially could be used for other things | 16:49 |
| JayF | lucasagomes: there is another API we could modify, that we control both ends of ... the IPA deploy api. You could make root device hints actually a part of the info sent to IPA for deploy | 16:49 |
| JayF | lucasagomes: that won't work, because we can't fetch the node on demand. It's pushed to us on lookup, lookup only happens on first boot. | 16:49 |
| lucasagomes | JayF, yeah, that's a good idea | 16:49 |
| *** dprince has joined #openstack-ironic | 16:50 | |
| * lucasagomes checks | 16:50 | |
| *** _vsaienko has joined #openstack-ironic | 16:51 | |
| *** _vs has joined #openstack-ironic | 16:51 | |
| *** afaranha has joined #openstack-ironic | 16:53 | |
| *** dtantsur is now known as dtantsur|afk | 16:54 | |
| jlvillal | FYI: Ironic QA meeting with discussion on Grenade in 5 minutes (1700 UTC) over on #openstack-meeting. No audio bridge this time. | 16:55 |
| *** mgould has quit IRC | 16:55 | |
| *** Sukhdev has quit IRC | 16:57 | |
| *** derekh has quit IRC | 17:02 | |
| lucasagomes | JayF, yeah no :-/ prepare_image(), install_bootloader() etc... already have a copy of the node object passed via ironic, the problem is that they use get_os_install_device() to find the device and that does not accept any parameter | 17:02 |
| * lucasagomes is thinking of a get_os_install_device2() | 17:02 | |
| JayF | that's still breaky in the same way | 17:02 |
| lucasagomes | it can check for NotImplemented | 17:03 |
| JayF | if you start calling get_os_install_device_new() that's a worse breakage | 17:03 |
| JayF | it will be implemented | 17:03 |
| lucasagomes | and call get_os_install_device() if that's the case | 17:03 |
| JayF | most external hwm I've seen | 17:03 |
| JayF | subclass GenericHWM | 17:03 |
| JayF | so it would have that method on it | 17:03 |
| lucasagomes | right, and that method will raise NotImplemented | 17:03 |
| JayF | no it won't | 17:03 |
| JayF | because it will be implemented by the parent class | 17:03 |
| JayF | and inherited to the more specific manager | 17:04 |
| lucasagomes | exactly, and not overwritten | 17:04 |
| lucasagomes | so it will raise the exception | 17:04 |
| lucasagomes | unless they overwrite it, so means that they actually implemented | 17:04 |
| JayF | well, I wish you were right but you aren't | 17:04 |
| JayF | so there are 2 ways you can do a custom hardware manager | 17:04 |
| JayF | subclass HardwareManager | 17:04 |
| JayF | subclass GenericHardwareManager | 17:04 |
| JayF | your example works for HWM, not for GenericHWM | 17:05 |
| JayF | because you'll have to put get_os_install_device2() into GenericHWM | 17:05 |
| lucasagomes | oh I get what you mean ... | 17:05 |
| lucasagomes | damn | 17:05 |
| lucasagomes | yeah | 17:05 |
| lucasagomes | the design seems very flawed :-/ | 17:05 |
| mat128 | wait, if there's nothing in either GenericHWM or 3rdpartyHWM | 17:05 |
| lucasagomes | ossified | 17:05 |
| JayF | this is shit from like, back in teeth-agent days | 17:05 |
| *** sabeen has quit IRC | 17:05 | |
| lucasagomes | yeah | 17:05 |
| mat128 | it will call HWM | 17:05 |
| JayF | I don't think "design" was much there as "worky worky fast fast" | 17:06 |
| *** lindycoder has joined #openstack-ironic | 17:06 | |
| JayF | lucasagomes: if you did node as a hwarg that defaulted to None, you still can't chagne that method sig without breaking folks? | 17:06 |
| lucasagomes | JayF, another way, is to do similar to what is done now... but before the methods dispatch get_os_install_device, they can dispatch a cache function | 17:06 |
| lucasagomes | 1- cache 2- get_os | 17:07 |
| JayF | lucasagomes: oh, it wouldn't break *callers* it'd be invalid because of subclass | 17:07 |
| JayF | lucasagomes: ah, so you'd cache a node right at deploy time | 17:07 |
| *** mgoddard_ has quit IRC | 17:07 | |
| JayF | lucasagomes: cache the node sent to you to deploy with | 17:07 |
| *** mgoddard has joined #openstack-ironic | 17:07 | |
| lucasagomes | JayF, even if it's none, some callers may want to pass it and it may not be presented on the 3rd party hw managers | 17:07 |
| JayF | oh; of course | 17:08 |
| lucasagomes | JayF, exactly, cache that node | 17:08 |
| *** nicodemos has quit IRC | 17:08 | |
| lucasagomes | the ones that are passed via the prepare_image and so on | 17:08 |
| JayF | So we still have a problem though | 17:08 |
| JayF | unless you 'cache the node' in a global or something | 17:08 |
| *** mgould has joined #openstack-ironic | 17:08 | |
| JayF | the comment I put on the review is still valid; you would have to cache a copy in every hardware manager | 17:08 |
| lucasagomes | JayF, yeah, it's possible to cache it in the hardware module | 17:08 |
| lucasagomes | so we avoid caching it on each hw manager | 17:09 |
| JayF | if you cache it in the hardware module rather than the hwm, and keep the cache_a_node method off the hwm | 17:09 |
| JayF | I'm very +1 to your suggestion | 17:09 |
| lucasagomes | JayF, exactly | 17:09 |
| JayF | and would basically say we need an update_node() method | 17:09 |
| JayF | that anytime Ironic sends us a new node, we update it | 17:09 |
| lucasagomes | yup | 17:09 |
| lucasagomes | so that keeps updated as we go | 17:09 |
| JayF | including, most likely, for clean steps | 17:09 |
| lucasagomes | yeah | 17:09 |
| lucasagomes | I have to take a look at the code, but I find that a way of doing it avoind breakages | 17:10 |
| lucasagomes | and duplication | 17:10 |
| JayF | I just kinda want to ask the question... should we break it anyway | 17:10 |
| lucasagomes | lol | 17:10 |
| lucasagomes | yeah... I wish | 17:10 |
| JayF | this is a place where inconsitency abounds | 17:10 |
| JayF | we pass nodes directly into clean steps | 17:10 |
| JayF | but not get_os_install_device? | 17:10 |
| *** absubram has joined #openstack-ironic | 17:11 | |
| JayF | honestly a cleanup item, at this point probably add to the list for P, Q, or R, would be a v2, formalized, agent api | 17:11 |
| lucasagomes | JayF, let's try not break, if we see that the interfaces are too ossified we probably send a ML warning | 17:11 |
| lucasagomes | and break it | 17:11 |
| *** mgould has quit IRC | 17:12 | |
| JayF | cool, glad we worked something out | 17:12 |
| * lucasagomes feels like developing in kernel mode with loads of duplicated syscalls... dup(), dup2() and mechanisms dnotify, inotify, fanotify | 17:13 | |
| lucasagomes | JayF, thanks for the help with ideas, appreciated! | 17:13 |
| mat128 | never break old stuff :) | 17:13 |
| *** nicodemos has joined #openstack-ironic | 17:13 | |
| lucasagomes | mat128, yup | 17:14 |
| JayF | lucasagomes: another idea/ | 17:14 |
| lucasagomes | JayF, go on :-) | 17:15 |
| *** mgould has joined #openstack-ironic | 17:15 | |
| *** daemontool has quit IRC | 17:15 | |
| JayF | lucasagomes: get_os_install_device_wrapper() could take a node as an argument, then do a dispatch_to_managers() for get_os_install_device -- if it exists, it gets called sans node. If it doesn't exist, it proceeds to run the "right new code" with the passed in node | 17:15 |
| *** mgould has quit IRC | 17:16 | |
| JayF | lucasagomes: we would pull get_os_install_device from GenericHWM though, which I guess is an API break in a sense as well | 17:16 |
| *** mgould has joined #openstack-ironic | 17:17 | |
| mat128 | JayF, lucasagomes: why not create a new one called "get_os_install_device_with_root_hints_from_node" and fall back to get_os_install_device (essentially your suggestion Jay, but it's the core that falls back) | 17:18 |
| JayF | mat128: the problem is for the "fall back" to make sense, it can't be in generichardwaremanager, and removing it would be an api break | 17:18 |
| mat128 | what happens when you dispatch_all and only some of them raise something? | 17:18 |
| mat128 | I would put in where you issue the dispatch call | 17:18 |
| lucasagomes | JayF, hmm yeah removing it def will break something | 17:18 |
| mat128 | https://github.com/openstack/ironic-python-agent/blob/master/ironic_python_agent/extensions/standby.py#L530 | 17:20 |
| mat128 | try catch :S | 17:20 |
| lucasagomes | yeah I'm not sure, will try the global thing first thing in the morning tomorrow | 17:23 |
| lucasagomes | I think that way we can avoid breakages | 17:23 |
| devananda | sambetts: chatting here for a sec while meeting continues -- your config for devstack flat network, have you tried applying that to the gate to see if it passes? | 17:29 |
| sambetts | devananda: I've not tried it with the VM BMs, I'm using it for my third party CI though | 17:30 |
| sambetts | devananda: I need to setup a devstack and see if it works | 17:30 |
| devananda | sambetts: if you have time, I'm really interested in that | 17:31 |
| *** fragatina has quit IRC | 17:31 | |
| devananda | sambetts: if it means we could replace/remove the network fudgery we're doing in devstack right now, and in the future have a "dsvm-ironig-flat-network" test along side a "dsvm-ironic-multitenant-network" test, that would be ideal | 17:31 |
| devananda | *and have those jobs merely be config differences | 17:32 |
| sambetts | devananda: yeah I think that would be really good to have | 17:32 |
| *** ohamada has quit IRC | 17:32 | |
| *** mgould is now known as mgould|afk | 17:34 | |
| *** athomas has joined #openstack-ironic | 17:36 | |
| *** trown is now known as trown|lunch | 17:36 | |
| *** electrofelix has quit IRC | 17:37 | |
| NobodyCam | good morning devananda | 17:40 |
| *** sabeen has joined #openstack-ironic | 17:40 | |
| openstackgerrit | Andre keedy proposed openstack/ironic-inspector: Replace node data via uuid/data/unprocessed API https://review.openstack.org/308379 | 17:44 |
| *** lucasagomes is now known as lucas-dinner | 17:44 | |
| sambetts | jlvillal, mgould|afk, vdrok: btw I updated this https://review.openstack.org/#/c/293954/ so that we might be able to remove that stuff to do with the tempest configuration out of our devstack plugin again, we'll have to wait and see responses | 17:44 |
| vdrok | sambetts: thanks, I will watch it | 17:45 |
| jlvillal | sambetts: Thanks I will take a look at that. Have a conflict right now... | 17:45 |
| sambetts | :) thanks | 17:45 |
| vdrok | good night everyone! | 17:48 |
| *** rama_y has joined #openstack-ironic | 17:50 | |
| *** piet has joined #openstack-ironic | 17:51 | |
| NobodyCam | night vdrok | 17:56 |
| jroll | totally missed the qa meeting :( | 18:02 |
| sambetts | krtaylor: https://review.openstack.org/#/c/318268/ | 18:03 |
| *** moshele has quit IRC | 18:04 | |
| sambetts | jroll: we talked about how devstack does networking and how we'll need to test in the future upgrades of flat network ironic -> flat network ironic and multitenant -> multitenet ironic | 18:04 |
| *** wendar_ is now known as wendar | 18:04 | |
| jroll | sambetts: yeah, I read back | 18:05 |
| *** fragatina has joined #openstack-ironic | 18:05 | |
| jroll | sambetts: what I'd rather like to think about is "how do we deprecate the flat network mode" | 18:06 |
| jroll | or rather, is that possible? | 18:06 |
| mat128 | jroll: is this something we want? | 18:07 |
| jroll | mat128: well, it's more like | 18:07 |
| mat128 | jroll: as an operator, we have always had multi-tenant networking downstream, but that might not be the case of private deployments | 18:07 |
| jroll | if you use the multitenant code proposed, where tenant net == cleaning net == provisioning net == ???, it works just like flat network ya? | 18:08 |
| mat128 | jroll: good point ! | 18:08 |
| jroll | :) | 18:08 |
| mat128 | jroll: except if you don't have an ml2 plugin for the switches you use | 18:09 |
| mat128 | guess you can always use the noop driver | 18:09 |
| mat128 | yeah, works | 18:09 |
| mat128 | so your question makes sense | 18:09 |
| jroll | right, exactly | 18:09 |
| *** Sukhdev has joined #openstack-ironic | 18:13 | |
| *** jcoufal has quit IRC | 18:15 | |
| sambetts | jroll: the only difference being that as we've seen in the grenade testing they expect the tenant network to be deleteable and recreatable which doesn't fit with our flat model where the network always exists :/ | 18:22 |
| jroll | sambetts: well, our CI frameworks are orthogonal to this I think | 18:24 |
| jroll | like, in that case, fix grenade | 18:24 |
| jroll | idk | 18:24 |
| *** amit213 has quit IRC | 18:24 | |
| jroll | or rather... don't test "use multitenant code to do flat networking" with grenade | 18:24 |
| *** fragatin_ has joined #openstack-ironic | 18:26 | |
| *** amit213 has joined #openstack-ironic | 18:27 | |
| sambetts | thats what I'm trying to workout I'm currently running a test to see if our devstack code works with Q_USE_PROVIDER_NETWORKING turned on where devstack will only configure a single flat network like we would in a real world deployment, and like I do in my third party CI to talk to real hardware | 18:27 |
| sambetts | I know devstack will pass successfully but I don't know if we'll plumb the VMs in right | 18:28 |
| sambetts | I'm expecting not | 18:29 |
| *** PixieBoots_ has joined #openstack-ironic | 18:29 | |
| *** fragatina has quit IRC | 18:29 | |
| *** fragatina has joined #openstack-ironic | 18:29 | |
| *** fragatin_ has quit IRC | 18:29 | |
| *** PixieBoots_ has quit IRC | 18:31 | |
| *** PixieBoots_ has joined #openstack-ironic | 18:31 | |
| *** sdake_ has joined #openstack-ironic | 18:31 | |
| *** PixieBoots_ has quit IRC | 18:33 | |
| *** sdake has quit IRC | 18:33 | |
| *** ijw has quit IRC | 18:40 | |
| *** marksei has joined #openstack-ironic | 18:43 | |
| *** hoonetorg has joined #openstack-ironic | 18:43 | |
| *** PixieBoots has quit IRC | 18:43 | |
| *** sacharya_ has joined #openstack-ironic | 18:43 | |
| *** trown|lunch is now known as trown | 18:44 | |
| *** PixieBoots has joined #openstack-ironic | 18:44 | |
| lucas-dinner | #findspec raid | 18:45 |
| PixieBoots | 3 spec(s) found: "inband raid configuration": http://specs.openstack.org/openstack/ironic-specs/specs/approved/inband-raid-configuration.html | "ironic generic raid interface": http://specs.openstack.org/openstack/ironic-specs/specs/approved/ironic-generic-raid-interface.html | "cisco ucs oob raid configuration": http://specs.openstack.org/openstack/ironic-specs/specs/approved/cisco-ucs-oob-raid-configuration.html | 18:45 |
| lucas-dinner | #pixiesay w00t --m rnr | 18:45 |
| PixieBoots | \m/ʕ>ᴥ<ʔ\m/: w00t | 18:45 |
| * lucas-dinner goes grab dinner now :D | 18:45 | |
| jroll | \o/ | 18:45 |
| *** sacharya has quit IRC | 18:47 | |
| *** moshele has joined #openstack-ironic | 18:47 | |
| sambetts | \o/ | 18:47 |
| *** anteaya has joined #openstack-ironic | 18:47 | |
| * sambetts -> heads off | 18:47 | |
| sambetts | night all | 18:47 |
| *** sambetts is now known as sambetts|afk | 18:47 | |
| *** e0ne has quit IRC | 18:48 | |
| *** jaybeale has quit IRC | 18:53 | |
| *** moshele has quit IRC | 18:55 | |
| NobodyCam | night sambetts|afk | 18:56 |
| *** amotoki has quit IRC | 18:56 | |
| *** anteaya has quit IRC | 18:57 | |
| *** jaypipes has quit IRC | 18:58 | |
| jroll | vsaienko: nova folks want to run with https://review.openstack.org/#/c/318027/ can you unwip that? | 18:59 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: WIP: grenade checking https://review.openstack.org/311109 | 19:03 |
| openstackgerrit | Vasyl Saienko proposed openstack/ironic: Run smoke tests after upgrade https://review.openstack.org/318288 | 19:03 |
| jroll | devananda: dtantsur|afk: when you have a minute, https://review.openstack.org/#/c/318045/1 | 19:05 |
| *** piet has quit IRC | 19:05 | |
| *** piet has joined #openstack-ironic | 19:06 | |
| *** sacharya has joined #openstack-ironic | 19:12 | |
| *** sacharya_ has quit IRC | 19:16 | |
| *** Sukhdev has quit IRC | 19:16 | |
| *** PixieBoots_ has joined #openstack-ironic | 19:17 | |
| *** PixieBoots_ has quit IRC | 19:18 | |
| *** piet has quit IRC | 19:21 | |
| *** sdake_ has quit IRC | 19:23 | |
| *** KennethWilke has quit IRC | 19:26 | |
| *** piet has joined #openstack-ironic | 19:27 | |
| *** jjohnson2 has joined #openstack-ironic | 19:28 | |
| *** KennethWilke has joined #openstack-ironic | 19:30 | |
| *** rama_y has quit IRC | 19:34 | |
| *** baoli_ has quit IRC | 19:34 | |
| *** KennethWilke has quit IRC | 19:35 | |
| *** KennethWilke has joined #openstack-ironic | 19:37 | |
| *** fragatina has quit IRC | 19:38 | |
| *** fragatina has joined #openstack-ironic | 19:38 | |
| openstackgerrit | VIncent S. Cojot proposed openstack/ironic: Makes the vbox driver work on headless VirtualBox machines https://review.openstack.org/311278 | 19:39 |
| *** rloo has quit IRC | 19:39 | |
| *** dprince has quit IRC | 19:44 | |
| *** e0ne has joined #openstack-ironic | 19:47 | |
| *** sdake has joined #openstack-ironic | 19:49 | |
| *** rloo has joined #openstack-ironic | 19:55 | |
| *** e0ne has quit IRC | 19:58 | |
| *** marksei has quit IRC | 19:58 | |
| *** ElCoyote_ has joined #openstack-ironic | 20:00 | |
| ElCoyote_ | Hi | 20:02 |
| *** bodepd has joined #openstack-ironic | 20:03 | |
| *** e0ne has joined #openstack-ironic | 20:04 | |
| *** ijw has joined #openstack-ironic | 20:05 | |
| ElCoyote_ | vsaienko: About https://review.openstack.org/#/c/311278/3, could you elaborate a little bit? I cannot commit to changing a functionnality that's not even yet present in ironic (letting the user pick a behaviour for a specific hypervisor based on some field in driver_info). | 20:05 |
| JayF | I think you're timezone'd trying to talk to vsaienko right now | 20:08 |
| JayF | but generally it's very OK to reply directly in the review to the person if you disagree or think they are wrong | 20:08 |
| ElCoyote_ | JayF: Thank you.. I'm very new to this. | 20:09 |
| JayF | no problem | 20:10 |
| *** ijw has quit IRC | 20:10 | |
| JayF | ElCoyote_: fwiw driver_info is a driver specific field | 20:10 |
| JayF | ElCoyote_: so if your vbox driver is a separate thing, it's not a bad thing to have something that would only make sense for vbox and not for other hardware/drivers | 20:10 |
| ElCoyote_ | Actually.. it's a small bugfix that I'm using to make the vbox driver useable in ironic for things like tripleo. | 20:11 |
| ElCoyote_ | I'm actually using my patch to do tripleo setups on OSP7(kilo) and OSP8(liberty) | 20:12 |
| JayF | OSP? | 20:12 |
| ElCoyote_ | Without that patch, ironic cannot drive the virtual baremetal machines.. | 20:12 |
| ElCoyote_ | OSP: RedHat OpenStack. | 20:12 |
| JayF | So why does your vbox have controlvn --type headless and others don't? | 20:13 |
| JayF | that seems to be the main issue on comments i nyour patch | 20:13 |
| *** ametts has quit IRC | 20:13 | |
| jroll | I don't think this needs to be configurable via driver_info, fwiw | 20:14 |
| jroll | well, maybe I see the use case, hm | 20:15 |
| JayF | jroll: yeah, I avoided saying anything solidly one way or another because I think it could go either way | 20:15 |
| JayF | jroll: with leaning towards driver_info just b/c --type headless seems to be a parameter not supported in all versions or something based on reviews | 20:16 |
| jroll | yeah | 20:16 |
| jroll | JayF: well, I think only startvm has it, is the thing, the rest is likely assumed by whether the vm started headless | 20:16 |
| *** thrash is now known as thrash|biab | 20:16 | |
| jroll | JayF: the reason I like configurability is so that I can get me a console if it isn't headless :) | 20:17 |
| JayF | which is why it should probably be configured per node | 20:17 |
| JayF | ElCoyote_: ^ I think jroll has a pretty good case why doing it in driver_info is not a bad idea | 20:17 |
| jroll | now that said, the node isn't passed in there | 20:17 |
| jroll | so it's going to be annoying | 20:17 |
| JayF | full disclosure I don't think either of us run this driver | 20:17 |
| JayF | lol | 20:17 |
| jroll | heck no :) | 20:18 |
| ElCoyote_ | Is there anyone using that driver these days? I had the feeling (from the comments I got) that none of the reviewers were using that driver. | 20:18 |
| JayF | I don't know; but it's probably going to be evicted out of the tree at the end of the cycle for not having CI | 20:19 |
| jroll | ElCoyote_: well, most of us that do run this driver use the virsh version | 20:19 |
| jroll | (usually via devstack) | 20:19 |
| JayF | oh this is all _ssh driver? | 20:19 |
| jroll | but um | 20:19 |
| JayF | just ssh driver can do vbox too? | 20:19 |
| jroll | ElCoyote_: let me comment on this patch and hopefully it will help you | 20:19 |
| jroll | JayF: yeah | 20:19 |
| ElCoyote_ | JayF: Yes, that's what I am using. | 20:19 |
| jroll | JayF: the other vbox driver is via some http api rather than ssh, primarily for windows | 20:20 |
| ElCoyote_ | There are several problems with the vbox driver in the current implementation of ssh.py: 1) several options do not work if no-one is loggued on the graphical X11 display, the VBoxManage just fails. | 20:21 |
| ElCoyote_ | 1) can be solved by some (documented and undocumented) passing of '--type headless' . | 20:21 |
| *** liliars has joined #openstack-ironic | 20:21 | |
| ElCoyote_ | and 2) you cannot set the boot device if someone has open the VM in a window or if it's already powered on. | 20:22 |
| ElCoyote_ | 2) is solved by my patch because it forces the vm to be off if the driver is vbox before issuing a set_boot_device. | 20:22 |
| jroll | sure | 20:23 |
| jroll | so (1) is fixed by this patch, right? | 20:23 |
| ElCoyote_ | jroll. Yes | 20:23 |
| jroll | oh, and there's another patch for (2), awesome | 20:24 |
| ElCoyote_ | jroll: (1) and (2) are fixed by my patch. Note that I am only using VBox 5.x (has been out for almost two years).. it's the latest and it's free so I doubt people would stay on the 4.x branch | 20:24 |
| jroll | ElCoyote_: oh, same patch, I see that now, cool | 20:25 |
| JayF | ElCoyote_: one thing you could do, is if your changes require vbox 5.x is add a release note to your patch saying that | 20:25 |
| ElCoyote_ | jroll: I couldn't do KVM for my OSP setups at home so I went vbox and I figured that perhaps I could help improve the code to make other's life easier a bit.. | 20:25 |
| jroll | ElCoyote_: ya, fair enough, I appreciate it :) | 20:25 |
| ElCoyote_ | JayF: What would be a good way to add a note..? A comment in the code? Gerrit? | 20:25 |
| JayF | ElCoyote_: we have release notes, they're created using a program called "reno" | 20:26 |
| jroll | http://docs.openstack.org/developer/reno/usage.html | 20:26 |
| ElCoyote_ | I'd think that these changes work in VBox 4.x too but I'm not testing them.. | 20:26 |
| ElCoyote_ | JayF: Oh great.. I didn't know about 'reno'. | 20:26 |
| JayF | ElCoyote_: ah, yeah I have vbox 5.x as well or else I'd maybe give it a shot | 20:27 |
| jroll | yeah, I think it would be hard to find 4.x in the wild | 20:28 |
| ElCoyote_ | The nice thing (IMHO) about using a different hypervisor than the default native one is that you can compare your results with those who do the right thing (tm). | 20:28 |
| JayF | the right hypervisor is no hypervisor | 20:29 |
| JayF | lol | 20:29 |
| ElCoyote_ | Even genymotion (the free Android emulator) moved to a co-bundled VBox 5.x quite some time ago already.. | 20:29 |
| *** ametts has joined #openstack-ironic | 20:30 | |
| ElCoyote_ | JayF: How would I go about adding a release note for a specific driver (vbox) to openstack-ironic? | 20:30 |
| JayF | ElCoyote_: ~/dev/ironic # reno new some-short-description | 20:30 |
| jroll | see http://docs.openstack.org/developer/reno/usage.html | 20:31 |
| JayF | then follow the docs jroll linked on how to write the release note | 20:31 |
| JayF | other patches will have examples | 20:31 |
| jroll | all the notes are grouped together, just specify what you mean :) | 20:31 |
| *** rama_y has joined #openstack-ironic | 20:31 | |
| ElCoyote_ | Ok.. great. many thanks for your help. | 20:31 |
| JayF | then when you push up a new patchset make sure the file reno made is included in your patch | 20:31 |
| *** sacharya_ has joined #openstack-ironic | 20:31 | |
| ElCoyote_ | Ohh.. I get it. does reno require something like python3 or such? | 20:32 |
| *** e0ne has quit IRC | 20:32 | |
| jroll | no, it works with both 2 and 3 | 20:32 |
| ElCoyote_ | Ok, let me try.. :) | 20:32 |
| jroll | recommended use is actually: tox -evenv -- reno new some-short-description | 20:32 |
| jroll | because it will install reno (and stuff) | 20:32 |
| *** e0ne has joined #openstack-ironic | 20:34 | |
| JayF | til | 20:34 |
| *** sacharya has quit IRC | 20:34 | |
| *** e0ne has quit IRC | 20:35 | |
| jroll | (I just use a non-tox virtualenv so I don't need libmysql and such to do it | 20:35 |
| jroll | ) | 20:35 |
| ElCoyote_ | jroll: I like the change you suggested.. | 20:36 |
| ElCoyote_ | Let me try to code it up. | 20:36 |
| jroll | awesome :) | 20:37 |
| *** ChrisAusten has quit IRC | 20:37 | |
| JayF | jroll: any reason I shouldn't land that 3rd party ci spec edit? | 20:37 |
| ElCoyote_ | How do would I go about adding a 'headless' toggle to driver_info for a node? I've never done that. | 20:37 |
| JayF | jroll: noticed you didn't so I was curious if there was a reason | 20:38 |
| JayF | ElCoyote_: ironic node-update UUID add driver_info/headless=true # or something like this? add may be the wrong verb | 20:38 |
| jroll | JayF: just giving others a chance to check it out | 20:38 |
| JayF | jroll: given only me/you/kurt/ruby have voted on that spec ever | 20:38 |
| JayF | I'm going to land it | 20:38 |
| ElCoyote_ | ok, standard ironic not-update | 20:38 |
| jroll | ElCoyote_: what JayF said is right, driver_info is a json field in the db, no need to specify it in a schema or anything | 20:38 |
| ElCoyote_ | Are there other such toggles ? ones I could use as an example..? | 20:39 |
| JayF | there def are, I can't remember any right off the top of my head | 20:40 |
| jroll | ElCoyote_: that _parse_driver_info method loads a bunch of things from driver_info | 20:40 |
| jroll | it's just a dictionary at that point | 20:40 |
| jroll | so something like use_headless = driver_info.get('headless', False) | 20:40 |
| ElCoyote_ | Ok, gotcha.. I've seen that code at the end of the ssh.py file.. | 20:40 |
| *** nicodemos is now known as nico-afk | 20:41 | |
| ElCoyote_ | if it's specific to vbox, shouldn't it be called use_vbox_headless ? | 20:41 |
| ElCoyote_ | unless we want to port this kind of stuff to some of the other drivers.. | 20:42 |
| ElCoyote_ | the thing is all of the other drivers are 'headless' by default.. except for parallels | 20:42 |
| jroll | ElCoyote_: yeah, maybe | 20:43 |
| jroll | naming things is the hardest part | 20:43 |
| openstackgerrit | Merged openstack/ironic-specs: Revise third party continuous integration spec for new drivers https://review.openstack.org/314768 | 20:43 |
| openstackgerrit | Ruby Loo proposed openstack/ironic-specs: Support rolling upgrades https://review.openstack.org/299245 | 20:43 |
| JayF | I think namespacing it is a solid idea | 20:43 |
| NobodyCam | rloo: ohh | 20:43 |
| rloo | NobodyCam: huh? | 20:43 |
| NobodyCam | the spec | 20:44 |
| NobodyCam | :) | 20:44 |
| *** _vs has quit IRC | 20:44 | |
| *** _vsaienko has quit IRC | 20:44 | |
| rloo | NobodyCam: what about it? Did I do something bad? | 20:44 |
| NobodyCam | no no... I like the title | 20:44 |
| JayF | I love it when a fresh spec comes out of the oven right as I was about to review it | 20:44 |
| rloo | NobodyCam: :) | 20:44 |
| ElCoyote_ | jroll: I think I'll stick to use_headless (in case some other driver wants to offer the same functionnality) but only the vbox driver will make use of it. | 20:44 |
| JayF | ahhh the smell of fresh baked design in the afternoon | 20:44 |
| NobodyCam | JayF: :) | 20:45 |
| rloo | JayF: if you mean that spec, I feel like it is still a work in progress. Cuz i don't understand it all yet. But yeah, coment. | 20:45 |
| jroll | ElCoyote_: yeah, I'm okay with that but wouldn't be surprised if other people want it namespaced :) | 20:45 |
| ElCoyote_ | so you think I should got with use_vbox_headless? | 20:45 |
| jroll | ElCoyote_: oh, you should add it here as well: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/ssh.py#L78 | 20:46 |
| ElCoyote_ | s/got/go/ | 20:46 |
| jroll | ElCoyote_: maybe vbox_use_headless? | 20:46 |
| *** absubram has quit IRC | 20:46 | |
| ElCoyote_ | jroll: deal! :) | 20:46 |
| jroll | :) | 20:46 |
| *** hemna has joined #openstack-ironic | 20:49 | |
| *** ijw has joined #openstack-ironic | 20:50 | |
| ElCoyote_ | vbox_use_headless = info.get('vbox_use_headless', false) | 20:57 |
| ElCoyote_ | in parse_driver, would that be ok? | 20:57 |
| jroll | lgtm | 20:59 |
| *** ametts has quit IRC | 21:07 | |
| *** trown is now known as trown|outtypewww | 21:08 | |
| *** piet has quit IRC | 21:11 | |
| *** Sukhdev has joined #openstack-ironic | 21:12 | |
| *** thrash|biab is now known as thrash | 21:14 | |
| *** ayoung has quit IRC | 21:18 | |
| *** ayoung has joined #openstack-ironic | 21:18 | |
| *** rbudden has quit IRC | 21:18 | |
| jroll | JayF: are there any artifacts I can link people for "make bmc management better" | 21:20 |
| JayF | jroll: I'm going to translate that to "get your split maintenance spec up asap even if it isn't perfect" | 21:21 |
| JayF | jroll: right? | 21:21 |
| jroll | JayF: you have that -ops list email in your inbox I assume | 21:21 |
| jroll | I'm responding to that with "here's some ways we're working on this" | 21:22 |
| jroll | JayF: "no" is a fine answer, but if you have things written, I'd love to link to them | 21:22 |
| JayF | one of a handful of things on a list that would be done already if I wasn't sick this week :( | 21:22 |
| JayF | I'll try to get something up soon | 21:22 |
| jroll | fair enough | 21:23 |
| *** jjohnson2 has quit IRC | 21:23 | |
| jroll | I assume no rfe exists? | 21:23 |
| JayF | I can do that now if it's useful | 21:23 |
| JayF | just didn't know it would be useful without the spec to go with it :) | 21:23 |
| jroll | meh, I can link to my blog | 21:23 |
| JayF | wfm | 21:24 |
| *** piet has joined #openstack-ironic | 21:32 | |
| ElCoyote_ | jroll: Getting ERROR oslo_messaging.rpc.dispatcher NameError: global name 'vbox_use_headless' is not defined | 21:34 |
| jroll | ElCoyote_: got code? | 21:34 |
| ElCoyote_ | in ironic-conductor.log.. should I patch some place else too.? | 21:35 |
| ElCoyote_ | jroll: Let me commit the latest patchset | 21:35 |
| JayF | ElCoyote_: first, I'd push up the broken code you're testing so jroll and I can look, secondly, I usually try to get a unit test in verifying behavior as it can often help me suss out these kinds of problems | 21:35 |
| ElCoyote_ | ok, gimme a sec.. :) | 21:37 |
| ElCoyote_ | Returning exception global name 'vbox_use_headless' is not defined to caller | 21:41 |
| ElCoyote_ | I must be missing an addition to a more global dict.. | 21:41 |
| jroll | I suspect you're missing quotes in something like driver_info('vbox_use_headless') | 21:42 |
| ElCoyote_ | in _parse_driver, I have this: vbox_use_headless = info.get('vbox_use_headless', 'False') | 21:45 |
| JayF | False shouldn't be quoted there fwiw | 21:45 |
| JayF | but is doubtful i'ts the problem | 21:45 |
| *** ijw has quit IRC | 21:45 | |
| ElCoyote_ | JayF: Yup.. I agree.. | 21:45 |
| *** ijw has joined #openstack-ironic | 21:47 | |
| *** joprovost has quit IRC | 21:52 | |
| *** piet has quit IRC | 21:57 | |
| *** ijw has quit IRC | 22:01 | |
| *** rloo has quit IRC | 22:01 | |
| *** thiagop has quit IRC | 22:01 | |
| *** rloo has joined #openstack-ironic | 22:02 | |
| *** sacharya has joined #openstack-ironic | 22:03 | |
| *** wajdi_ has joined #openstack-ironic | 22:03 | |
| mrda | Morning Ironic | 22:04 |
| *** ijw has joined #openstack-ironic | 22:06 | |
| NobodyCam | good morning mrda | 22:06 |
| JayF | gm | 22:06 |
| *** sacharya_ has quit IRC | 22:07 | |
| *** sabeen has quit IRC | 22:07 | |
| devananda | morning! | 22:07 |
| *** wajdi has quit IRC | 22:07 | |
| *** hemna has quit IRC | 22:10 | |
| mrda | o/ | 22:11 |
| *** anteaya has joined #openstack-ironic | 22:14 | |
| *** lindycoder has quit IRC | 22:15 | |
| *** wajdi_ has quit IRC | 22:16 | |
| *** cdearborn has quit IRC | 22:19 | |
| *** harlowja has quit IRC | 22:22 | |
| *** athomas has quit IRC | 22:22 | |
| ElCoyote_ | JayF: It doesn't work.. cause _get_command_sets is called from _parse_driver_info and _get_command_sets is the place where vbox_use_headless would be used. | 22:30 |
| ElCoyote_ | JayF: I updated the gerrit comments for this.. | 22:30 |
| *** rloo has quit IRC | 22:37 | |
| *** rloo has joined #openstack-ironic | 22:45 | |
| *** mbound has quit IRC | 22:54 | |
| *** jaybeale has joined #openstack-ironic | 22:55 | |
| *** ijw has quit IRC | 22:57 | |
| *** ijw has joined #openstack-ironic | 22:59 | |
| *** harlowja has joined #openstack-ironic | 23:00 | |
| *** anteaya has quit IRC | 23:01 | |
| *** ijw has quit IRC | 23:03 | |
| *** rloo has quit IRC | 23:11 | |
| *** hemna has joined #openstack-ironic | 23:23 | |
| openstackgerrit | VIncent S. Cojot proposed openstack/ironic: Makes the vbox driver work on headless VirtualBox machines https://review.openstack.org/311278 | 23:23 |
| *** hemna has quit IRC | 23:23 | |
| *** jaybeale has quit IRC | 23:29 | |
| *** PixieBoots has quit IRC | 23:36 | |
| JayF | ElCoyote_: one comment on that patch | 23:36 |
| JayF | ElCoyote_: tl;dr: startvm --type headless is a thing; but controlvm --type headless is not | 23:36 |
| ElCoyote_ | JayF: Yes..? | 23:36 |
| JayF | ElCoyote_: unless --help on VBoxManage is lying :) | 23:36 |
| ElCoyote_ | JayF: Yes and yes | 23:36 |
| ElCoyote_ | That's why I have posted output in the previous comments | 23:37 |
| ElCoyote_ | On all the 5.0.x I tested it works. | 23:37 |
| JayF | Are you sure --type isn't just noop on the controlvm setting/ | 23:37 |
| JayF | i.e. does no harm but does no good either? | 23:38 |
| ElCoyote_ | [raistlin@thorbardin ~]$ VBoxManage controlvm osp-baremetal-1 reset --type headless | 23:38 |
| ElCoyote_ | [raistlin@thorbardin ~]$ | 23:38 |
| ElCoyote_ | [raistlin@thorbardin ~]$ echo $? | 23:38 |
| JayF | and that command fails without --type headless? | 23:38 |
| ElCoyote_ | Without --type headless ironic got an ssh failure due to a VboxManage failure | 23:39 |
| JayF | you're probably going to have to pastebin/gist this into the review :) | 23:39 |
| JayF | but was that with the code as is | 23:39 |
| ElCoyote_ | even on controlvm | 23:39 |
| JayF | because if so, it means you did startvm initially without --type headless | 23:39 |
| JayF | so if you run that command, the one you just ran, without --type headless it fails? Can you gist that? | 23:39 |
| JayF | if so we should flog whoever is NOT writing the vbox documentation, lol | 23:39 |
| ElCoyote_ | JayF: That's a good idea.. I hadn't thought that perhaps it was because I had started the VMs with --type headless that it worked at a later time. | 23:40 |
| JayF | yeah exactly | 23:40 |
| JayF | like it's a flag to be set on the vm at launch vs a flag to the cli | 23:40 |
| JayF | that's how I read the docs | 23:40 |
| ElCoyote_ | How do I put a pastebin in my gerrit review? | 23:40 |
| JayF | when I said that I meant more | 23:41 |
| JayF | if you're adding support for something not documented | 23:41 |
| JayF | people will want proof it doesn't work | 23:41 |
| ElCoyote_ | Yes the docs lie. Only the headless addition to startvm is documented but I'm using it in two other places | 23:41 |
| JayF | so pastebin some results of the modifyvm --type headless failing (on a VM started with --type headless) | 23:41 |
| ElCoyote_ | I agree. This needs a pastebin | 23:41 |
| JayF | and paste the link to it into a comment in the review | 23:41 |
| JayF | otherwise you're never going to get a core to approve landing code that isn't backed by the documentation in vbox | 23:42 |
| *** Sukhdev has quit IRC | 23:42 | |
| ElCoyote_ | Ok, will do it that way. What pastebin is good practive with openstack? | 23:42 |
| JayF | I think there's actually an openstack one? I tend to just use gist.github.com | 23:42 |
| JayF | even a txt file on a webserver is fine :) | 23:42 |
| JayF | just avoiding not pasting in lots of lines of code to IRC or the review ++ | 23:42 |
| ElCoyote_ | Ok.. sorry about that and thanks for your patience and kind words of advice. | 23:43 |
| JayF | hey it's not a problem, we're all new at some point | 23:44 |
| JayF | some of us for longer than others :P | 23:44 |
| ElCoyote_ | controlvm.. type headless works in both cases (vm started with or whitout --type headless). Gonna paste that. | 23:46 |
| JayF | So I think you accidentally proved me right | 23:46 |
| JayF | the test I'd suggest doing is: | 23:46 |
| JayF | 1) start two vms, one using createvm --type headless and one not | 23:47 |
| ElCoyote_ | createvm doesn't have a type headless. | 23:47 |
| JayF | 2) Try to manage both of them using the 'wrong' modifyvm setting (--type headless on the not headless one, and vice versa) | 23:47 |
| JayF | oh, I meant startvm | 23:47 |
| JayF | damn that's confusing | 23:47 |
| ElCoyote_ | yeah. Will try that and paste the output in https://paste.fedoraproject.org/ | 23:47 |
| ElCoyote_ | Yes, it is | 23:47 |
| JayF | I guess it makes sense that headless would be determined on start though | 23:47 |
| JayF | because it has to decide to hookup fake vga or not at start time | 23:48 |
| *** Sukhdev has joined #openstack-ironic | 23:48 | |
| JayF | still same thought applies there :) does --type headless do anything on any command other than startvm | 23:48 |
| JayF | my hypothesis is that it does not | 23:48 |
| *** jaybeale has joined #openstack-ironic | 23:51 | |
| *** ijw has joined #openstack-ironic | 23:51 | |
| *** mbound has joined #openstack-ironic | 23:55 | |
| *** rloo has joined #openstack-ironic | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
