| harlowja | ^ something like that :-P | 00:00 |
|---|---|---|
| *** Marga_ has quit IRC | 00:01 | |
| harlowja | or u could have read-only 'node' objects on the API used FSM (and read/write on the conductor used FSM), anyway, u get the point i think | 00:02 |
| devananda | yup | 00:02 |
| devananda | or I could leave that refactoring 'till later :) | 00:02 |
| harlowja | that to | 00:02 |
| *** sdake has quit IRC | 00:16 | |
| *** smoriya has joined #openstack-ironic | 00:16 | |
| *** shakamunyi has quit IRC | 00:18 | |
| harlowja | devananda an idea, doing that kind of separation of FSM + reactions and the thing being worked on ('node' or whatever), if u then had the thing running the FSM have a lock on a 'node' in zookeeper, that tells others that node is currently having its FSM ran, then u could detect that the 'thing' has died, recreate the 'node' object on some other machine, resume the FSM from wherever it was, annnd profit... i think this is | 00:19 |
| harlowja | similar to whats already happening maybe? :-P | 00:19 |
| *** ChuckC has joined #openstack-ironic | 00:20 | |
| harlowja | * with a database | 00:20 |
| jroll | harlowja: I like this | 00:21 |
| harlowja | ya, my terminology might not match ironics, but the idea hopefully makes sorta sense :-P | 00:21 |
| harlowja | its similar in concept to taskflow ;) | 00:21 |
| harlowja | anyways, some random idea from me :-P | 00:23 |
| *** jb11211 has quit IRC | 00:24 | |
| *** naohirot has joined #openstack-ironic | 00:25 | |
| naohirot | good morning ironic! | 00:25 |
| NobodyCam | morning naohirot | 00:26 |
| *** andreykurilin_ has quit IRC | 00:26 | |
| naohirot | NobodyCam: Hi, thank you for the review :) | 00:26 |
| *** romcheg1 has quit IRC | 00:32 | |
| *** romcheg1 has joined #openstack-ironic | 00:32 | |
| NobodyCam | :) | 00:33 |
| *** romcheg1 has quit IRC | 00:33 | |
| jiangfei | morning all | 00:33 |
| NobodyCam | Good morning jiangfei | 00:35 |
| jiangfei | good morning NobodyCam | 00:36 |
| *** Masahiro has joined #openstack-ironic | 00:40 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic: Add fsm.py module https://review.openstack.org/139215 | 00:41 |
| *** Marga_ has joined #openstack-ironic | 00:43 | |
| NobodyCam | rloo: that should also address your comments ^^^^^ | 00:43 |
| NobodyCam | devananda: some initial tests ^^^^ | 00:43 |
| *** davideagnello has quit IRC | 00:44 | |
| devananda | d'oh | 00:44 |
| devananda | lemme rebase all my things :) | 00:44 |
| *** Masahiro has quit IRC | 00:45 | |
| *** Marga_ has quit IRC | 00:45 | |
| *** Marga_ has joined #openstack-ironic | 00:46 | |
| *** Marga_ has quit IRC | 00:46 | |
| *** Marga_ has joined #openstack-ironic | 00:47 | |
| *** Marga_ has quit IRC | 00:47 | |
| *** Marga_ has joined #openstack-ironic | 00:47 | |
| NobodyCam | oh pooh I'm sorry about that devananda | 00:48 |
| devananda | no no | 00:48 |
| devananda | the timing was funny -- I had just hit 'git review' right after you | 00:49 |
| NobodyCam | oh | 00:49 |
| devananda | i'll pull yours down and do a quick merge, then push back up | 00:49 |
| NobodyCam | ack :) | 00:49 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Refactor async helper methods in conductor/manager.py https://review.openstack.org/139217 | 00:53 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Begin using the state machine for node deploy/teardown https://review.openstack.org/139216 | 00:53 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Enable async callbacks from task.process_event() https://review.openstack.org/140868 | 00:53 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Convert check_deploy_timeout to use process_event https://review.openstack.org/140869 | 00:53 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Add fsm.py module https://review.openstack.org/139215 | 00:53 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Start using process_event() within drivers https://review.openstack.org/140883 | 00:53 |
| devananda | :) | 00:53 |
| *** afaranha__ has joined #openstack-ironic | 00:56 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Start using process_event() within drivers https://review.openstack.org/140883 | 00:57 |
| devananda | withthat, I think I'm done for the day | 00:57 |
| NobodyCam | ack | 00:58 |
| NobodyCam | have a good night devananda :) | 00:58 |
| harlowja | devananda http://i.imgur.com/2uZ521l.png NobodyCam | 00:59 |
| harlowja | regenereated | 00:59 |
| harlowja | just incase u guys changed stuff :-P | 00:59 |
| *** alexpilotti has quit IRC | 00:59 | |
| NobodyCam | we did | 00:59 |
| NobodyCam | :) | 00:59 |
| harlowja | k | 00:59 |
| NobodyCam | or devananda did :) and it shows up | 00:59 |
| harlowja | cols | 01:00 |
| harlowja | cools | 01:00 |
| *** alexpilotti has joined #openstack-ironic | 01:00 | |
| *** ryanpetrello has joined #openstack-ironic | 01:02 | |
| *** Marga_ has quit IRC | 01:02 | |
| *** alexpilotti has quit IRC | 01:03 | |
| devananda | hah, neat | 01:03 |
| devananda | looks like rebuild is still incorrect | 01:03 |
| harlowja | possibly | 01:04 |
| harlowja | http://i.imgur.com/n7Qalvf.png with colors | 01:04 |
| harlowja | lol | 01:04 |
| *** cuihaozhi has joined #openstack-ironic | 01:04 | |
| harlowja | devananda is rebuilt an alternative start point? | 01:04 |
| * harlowja will make that so if so | 01:04 | |
| devananda | http://paste.openstack.org/show/LbGFhaEP3jYl6FayKhAd/ | 01:04 |
| *** Marga__ has joined #openstack-ironic | 01:04 | |
| devananda | it's never used. anywhere | 01:04 |
| harlowja | kk | 01:06 |
| devananda | it's a "verb" in the new taxonomy of things | 01:06 |
| cuihaozhi | Hi,all. when use "pxe_ipmi" driver, do i need enable tftp-service with xinet.d? | 01:06 |
| devananda | but we never represent it as a state ... /me updates | 01:06 |
| devananda | cuihaozhi: yes | 01:06 |
| cuihaozhi | devananda: thanks | 01:07 |
| harlowja | updated generator of that diagram @ http://paste.ubuntu.com/9469615/ ; have fun and be merry, lol | 01:09 |
| *** ryanpetrello has quit IRC | 01:13 | |
| devananda | harlowja: where's that link to the fix for assert six.callable? | 01:13 |
| devananda | I'll incorporate that now. about to push a fix for rebuild | 01:14 |
| devananda | (I really should be afk ...) | 01:14 |
| * harlowja locating | 01:14 | |
| harlowja | https://github.com/openstack/taskflow/commit/a57c8119227feb | 01:15 |
| harlowja | search for taskflow/types/fsm.py | 01:15 |
| harlowja | in that | 01:15 |
| devananda | ah, ty | 01:15 |
| harlowja | or in https://github.com/openstack/taskflow/commit/a57c8119227feb.patch | 01:15 |
| harlowja | ^ might be easier to suck down just the chunk of that patch | 01:15 |
| devananda | yea. forgot how simple that change was | 01:16 |
| harlowja | ya, nothing special | 01:16 |
| *** Marga__ has quit IRC | 01:16 | |
| *** Marga_ has joined #openstack-ironic | 01:17 | |
| *** Marga_ has quit IRC | 01:17 | |
| *** Marga_ has joined #openstack-ironic | 01:18 | |
| *** Marga_ has quit IRC | 01:19 | |
| *** Marga_ has joined #openstack-ironic | 01:20 | |
| *** davideagnello has joined #openstack-ironic | 01:20 | |
| *** Marga_ has quit IRC | 01:20 | |
| *** Marga_ has joined #openstack-ironic | 01:21 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Refactor async helper methods in conductor/manager.py https://review.openstack.org/139217 | 01:23 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Begin using the state machine for node deploy/teardown https://review.openstack.org/139216 | 01:23 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Convert drivers to use process_event() https://review.openstack.org/140883 | 01:23 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Enable async callbacks from task.process_event() https://review.openstack.org/140868 | 01:23 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Convert check_deploy_timeout to use process_event https://review.openstack.org/140869 | 01:23 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Add base state machine https://review.openstack.org/139215 | 01:23 |
| *** davideagnello has quit IRC | 01:23 | |
| devananda | harlowja: that should no longer have a rebuild state bubble :) | 01:24 |
| harlowja | rebuilding | 01:24 |
| *** Masahiro has joined #openstack-ironic | 01:24 | |
| harlowja | ya http://i.imgur.com/0CdomcK.png | 01:25 |
| harlowja | no more hole :-P | 01:25 |
| devananda | woot | 01:25 |
| devananda | ta, i'm really afk now :) | 01:27 |
| harlowja | k, another neat idea, expose this state machine to horizon, then people can watch it move between states, ha | 01:27 |
| harlowja | using some javascript magic | 01:28 |
| harlowja | :-P | 01:28 |
| cuihaozhi | How can i get tftp-service log? | 01:28 |
| cuihaozhi | with dnsmasq option? | 01:28 |
| *** jmanko has quit IRC | 01:30 | |
| *** chenglch has joined #openstack-ironic | 01:30 | |
| *** davideagnello has joined #openstack-ironic | 01:30 | |
| rloo | NobodyCam, thx for updating the fsm patch. I'll look at it tomorrow. | 01:30 |
| *** davideagnello has quit IRC | 01:31 | |
| *** r-daneel has quit IRC | 01:33 | |
| *** marcoemorais has quit IRC | 01:35 | |
| jroll | cuihaozhi: I think it's usually in syslog | 01:36 |
| *** ChuckC has quit IRC | 01:38 | |
| *** marcoemorais has joined #openstack-ironic | 01:39 | |
| *** shakamunyi has joined #openstack-ironic | 01:43 | |
| *** Marga_ has quit IRC | 01:43 | |
| *** ChuckC has joined #openstack-ironic | 01:44 | |
| *** killer_prince has quit IRC | 01:45 | |
| *** killer_prince has joined #openstack-ironic | 01:46 | |
| *** killer_prince is now known as lazy_prince | 01:46 | |
| *** marcoemorais has quit IRC | 01:46 | |
| *** afaranha__ has quit IRC | 01:48 | |
| *** ChuckC has quit IRC | 01:48 | |
| *** ChuckC has joined #openstack-ironic | 01:49 | |
| *** r-daneel has joined #openstack-ironic | 01:49 | |
| *** lazy_prince has quit IRC | 01:52 | |
| adam_g | cuihaozhi, if you're using xinetd, add -v's to to server_args in /etc/xinet.d/tftp or wherever you've configured that | 01:54 |
| *** killer_prince has joined #openstack-ironic | 01:54 | |
| adam_g | and look in syslo | 01:54 |
| *** killer_prince is now known as lazy_prince | 01:54 | |
| *** rloo has quit IRC | 02:13 | |
| *** Marga_ has joined #openstack-ironic | 02:20 | |
| *** Marga_ has quit IRC | 02:25 | |
| *** david-lyle is now known as david-lyle_afk | 02:30 | |
| openstackgerrit | Merged openstack/ironic: Update dev quick-start for devstack https://review.openstack.org/140124 | 02:31 |
| openstackgerrit | Merged openstack/ironic: Updated from global requirements https://review.openstack.org/140779 | 02:31 |
| *** Masahiro has quit IRC | 02:42 | |
| *** ramineni has joined #openstack-ironic | 02:43 | |
| *** Marga_ has joined #openstack-ironic | 02:44 | |
| *** Masahiro has joined #openstack-ironic | 02:49 | |
| *** Haomeng has joined #openstack-ironic | 03:06 | |
| *** Marga_ has quit IRC | 03:07 | |
| *** Haomeng|2 has quit IRC | 03:07 | |
| *** Marga_ has joined #openstack-ironic | 03:07 | |
| *** Masahiro has quit IRC | 03:08 | |
| *** Marga_ has quit IRC | 03:11 | |
| *** Marga_ has joined #openstack-ironic | 03:11 | |
| *** shakamunyi has quit IRC | 03:12 | |
| *** Marga_ has quit IRC | 03:12 | |
| *** Marga_ has joined #openstack-ironic | 03:13 | |
| *** Marga_ has quit IRC | 03:13 | |
| *** Marga_ has joined #openstack-ironic | 03:13 | |
| *** Marga__ has joined #openstack-ironic | 03:15 | |
| *** Marga_ has quit IRC | 03:15 | |
| *** Masahiro has joined #openstack-ironic | 03:16 | |
| *** Marga__ has quit IRC | 03:16 | |
| *** Marga_ has joined #openstack-ironic | 03:16 | |
| *** Marga_ has quit IRC | 03:17 | |
| *** Marga_ has joined #openstack-ironic | 03:17 | |
| *** Marga_ has quit IRC | 03:18 | |
| *** Marga_ has joined #openstack-ironic | 03:19 | |
| *** dlaube has quit IRC | 03:20 | |
| *** Marga_ has quit IRC | 03:20 | |
| *** Marga_ has joined #openstack-ironic | 03:21 | |
| *** Marga_ has quit IRC | 03:23 | |
| *** Marga_ has joined #openstack-ironic | 03:23 | |
| *** Marga_ has quit IRC | 03:25 | |
| *** Marga_ has joined #openstack-ironic | 03:26 | |
| *** Marga__ has joined #openstack-ironic | 03:27 | |
| *** Marga_ has quit IRC | 03:27 | |
| *** yuanying has joined #openstack-ironic | 03:28 | |
| *** yuanying_ has quit IRC | 03:28 | |
| *** Marga__ has quit IRC | 03:28 | |
| *** Marga_ has joined #openstack-ironic | 03:28 | |
| *** Marga_ has quit IRC | 03:29 | |
| *** Marga_ has joined #openstack-ironic | 03:29 | |
| *** Marga_ has quit IRC | 03:31 | |
| *** Marga_ has joined #openstack-ironic | 03:32 | |
| *** naohirot has quit IRC | 03:33 | |
| *** Marga_ has quit IRC | 03:39 | |
| *** Marga_ has joined #openstack-ironic | 03:39 | |
| *** ryanpetrello has joined #openstack-ironic | 03:40 | |
| *** lazy_prince has quit IRC | 03:45 | |
| *** killer_prince has joined #openstack-ironic | 03:47 | |
| *** killer_prince is now known as lazy_prince | 03:47 | |
| *** ryanpetrello has quit IRC | 03:54 | |
| *** naohirot has joined #openstack-ironic | 04:03 | |
| *** Masahiro has quit IRC | 04:09 | |
| *** Marga_ has quit IRC | 04:10 | |
| *** Masahiro has joined #openstack-ironic | 04:12 | |
| *** dlaube has joined #openstack-ironic | 04:14 | |
| *** Masahiro has quit IRC | 04:25 | |
| *** Masahiro has joined #openstack-ironic | 04:33 | |
| *** pradipta_away is now known as pradipta | 04:34 | |
| *** dlaube has quit IRC | 04:46 | |
| openstackgerrit | Ramakrishnan G proposed openstack/ironic-specs: Enable boot images sharing for iLO driver(s) https://review.openstack.org/137291 | 04:58 |
| *** achanda has joined #openstack-ironic | 04:59 | |
| *** masco has joined #openstack-ironic | 05:10 | |
| *** pensu has joined #openstack-ironic | 05:16 | |
| *** achanda has quit IRC | 05:16 | |
| *** jerryz has joined #openstack-ironic | 05:20 | |
| *** achanda has joined #openstack-ironic | 05:22 | |
| *** mrda is now known as mrda-away | 05:34 | |
| *** rakesh_hs has joined #openstack-ironic | 05:35 | |
| *** harlowja is now known as harlowja_away | 05:36 | |
| Haomeng | pensu: ping | 05:45 |
| *** smoriya has quit IRC | 05:47 | |
| pensu | Haomeng: Hey | 05:54 |
| pensu | pensu: still going through the stuff you gave me, wanted to change through dnsmasq, but there is no dnsmasq.conf....:( | 05:55 |
| Haomeng | pensu: I worked out the dnsmasq conf which can be used for ironic without neutron | 05:55 |
| Haomeng | pensu: share to you | 05:55 |
| pensu | Haomeng: awesome! yes, please....:) | 05:55 |
| Haomeng | http://paste.openstack.org/show/149196/ | 05:56 |
| Haomeng | pensu: yes, it is cool without neutron:) | 05:56 |
| Haomeng | pensu: and nova | 05:56 |
| Haomeng | pensu: now, my env depdnes on glance only | 05:56 |
| Haomeng | pensu: easy env:) | 05:56 |
| Haomeng | pensu: want to find some way to configure the white list for dnsmasq for special mac | 05:57 |
| *** rushiagr_away is now known as rushiagr | 06:00 | |
| *** Masahiro has quit IRC | 06:00 | |
| pensu | Haomeng: great....one thing, the /etc/ironic-discovered/.....my Ironic doesn't make that, you created that or should it be there? Coz my setup doesn't have any dnsmasq.conf... | 06:01 |
| pensu | Haomeng: Wow! That is amazing.... | 06:01 |
| *** pcrews has quit IRC | 06:01 | |
| Haomeng | pensu: :) | 06:01 |
| *** pradipta is now known as pradipta_away | 06:01 | |
| Haomeng | pensu: set dhcpprovider=none in ironic conf | 06:01 |
| Haomeng | pensu: do you want to run discovered? | 06:03 |
| Haomeng | pensu: my dnsmasq.conf is for the case without neutron control | 06:03 |
| *** Masahiro has joined #openstack-ironic | 06:03 | |
| Haomeng | pensu: run a new dnsmasq with my conf as input | 06:04 |
| Haomeng | pensu: it works in my env:) | 06:04 |
| pensu | Haomeng: I have done that already... | 06:04 |
| Haomeng | pensu: ok | 06:04 |
| *** achanda has quit IRC | 06:04 | |
| pensu | Haomeng: I actually want to find dnsmasq.conf......looks like it is running dnsmasq without a separate config file... | 06:04 |
| pensu | Haomeng: let me check with a custom file... | 06:04 |
| pensu | Haomeng: I am not using discovery... | 06:05 |
| Haomeng | pensu: no | 06:05 |
| Haomeng | pensu: we can run dnsmasq with mult-process with different dnsmasq.conf as input | 06:06 |
| pensu | Haomeng: oh, yeah, we can do that....thanks...:) | 06:07 |
| Haomeng | pensu: it is confused by me for the conf name | 06:07 |
| Haomeng | dont care the file name | 06:07 |
| Haomeng | but it is working with both discoverd and normal ironic bm provision | 06:09 |
| pensu | Haomeng: yeah, I am focusing on normal bm provision for now....let's hope it works...:) | 06:10 |
| Haomeng | sure, I verified already | 06:10 |
| Haomeng | make sure the ip range is same subnet with your interface in conf | 06:12 |
| *** achanda has joined #openstack-ironic | 06:12 | |
| pensu | Haomeng: sure...:) | 06:13 |
| Haomeng | in my env, 9.119.58.1 is gateway, 9.119.58.134 is the conductor's ip | 06:14 |
| *** achanda has quit IRC | 06:15 | |
| *** chenglch has quit IRC | 06:15 | |
| *** chenglch|2 has joined #openstack-ironic | 06:18 | |
| *** lazy_prince2 has joined #openstack-ironic | 06:18 | |
| *** kes has joined #openstack-ironic | 06:18 | |
| *** lazy_prince2 has quit IRC | 06:21 | |
| *** achanda has joined #openstack-ironic | 06:31 | |
| *** Masahiro has quit IRC | 06:34 | |
| *** achanda has quit IRC | 06:47 | |
| *** Marga_ has joined #openstack-ironic | 06:50 | |
| *** dlpartain has joined #openstack-ironic | 06:50 | |
| *** Marga_ has quit IRC | 06:51 | |
| *** achanda has joined #openstack-ironic | 06:59 | |
| *** Haomeng|2 has joined #openstack-ironic | 07:07 | |
| *** Haomeng has quit IRC | 07:07 | |
| *** achanda has quit IRC | 07:11 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/ironic: Updated from global requirements https://review.openstack.org/140941 | 07:13 |
| *** smoriya has joined #openstack-ironic | 07:18 | |
| *** holser has left #openstack-ironic | 07:20 | |
| *** dlpartain has left #openstack-ironic | 07:32 | |
| *** kes has quit IRC | 07:37 | |
| *** chenglch|2 has quit IRC | 07:37 | |
| *** ifarkas has joined #openstack-ironic | 07:40 | |
| *** kes has joined #openstack-ironic | 07:42 | |
| *** k4n0 has joined #openstack-ironic | 07:44 | |
| *** Masahiro has joined #openstack-ironic | 07:44 | |
| *** jcoufal has joined #openstack-ironic | 07:46 | |
| *** Masahiro has quit IRC | 07:51 | |
| *** Masahiro has joined #openstack-ironic | 07:56 | |
| *** chenglch has joined #openstack-ironic | 07:57 | |
| *** jerryz_ has joined #openstack-ironic | 08:03 | |
| *** jerryz has quit IRC | 08:06 | |
| *** achanda has joined #openstack-ironic | 08:15 | |
| *** ZhangYang has quit IRC | 08:24 | |
| *** mkerrin has quit IRC | 08:29 | |
| *** igordcard has joined #openstack-ironic | 08:40 | |
| *** jcoufal has quit IRC | 08:42 | |
| *** jcoufal has joined #openstack-ironic | 08:43 | |
| *** rushiagr has left #openstack-ironic | 08:46 | |
| *** dtantsur|afk is now known as dtantsur | 08:57 | |
| dtantsur | Morning! | 08:57 |
| *** jcoufal has quit IRC | 09:01 | |
| *** jcoufal has joined #openstack-ironic | 09:01 | |
| openstackgerrit | Merged stackforge/ironic-discoverd: Require manual power on if ipmi_setup_credentials is set in Node.extra https://review.openstack.org/140656 | 09:04 |
| *** jistr has joined #openstack-ironic | 09:06 | |
| *** bradjones has quit IRC | 09:08 | |
| *** erwan_taf_ has joined #openstack-ironic | 09:09 | |
| *** erwan_taf has joined #openstack-ironic | 09:09 | |
| *** bradjones has joined #openstack-ironic | 09:10 | |
| *** athomas has joined #openstack-ironic | 09:11 | |
| *** achanda has quit IRC | 09:11 | |
| *** achanda has joined #openstack-ironic | 09:12 | |
| *** derekh has joined #openstack-ironic | 09:13 | |
| *** achanda_ has joined #openstack-ironic | 09:13 | |
| *** masco has quit IRC | 09:14 | |
| *** masco has joined #openstack-ironic | 09:15 | |
| *** achanda has quit IRC | 09:16 | |
| *** achanda_ has quit IRC | 09:20 | |
| *** bradjones has quit IRC | 09:24 | |
| *** andreykurilin_ has joined #openstack-ironic | 09:24 | |
| *** mkerrin has joined #openstack-ironic | 09:27 | |
| *** Masahiro has quit IRC | 09:29 | |
| *** Masahiro has joined #openstack-ironic | 09:30 | |
| *** pradipta_away is now known as pradipta | 09:33 | |
| *** bradjones has joined #openstack-ironic | 09:43 | |
| *** Masahiro has quit IRC | 09:52 | |
| *** athomas has quit IRC | 09:53 | |
| *** athomas has joined #openstack-ironic | 10:04 | |
| *** Masahiro has joined #openstack-ironic | 10:08 | |
| *** naohirot has quit IRC | 10:16 | |
| *** chenglch has quit IRC | 10:28 | |
| *** igordcard has quit IRC | 10:30 | |
| sambetts | morning dtantsur | 10:31 |
| *** Masahiro has quit IRC | 10:33 | |
| *** Masahiro has joined #openstack-ironic | 10:34 | |
| *** andreykurilin_ has quit IRC | 10:36 | |
| *** pelix has joined #openstack-ironic | 10:39 | |
| dtantsur | sambetts, morning | 10:43 |
| *** dtantsur is now known as dtantsur|brb | 10:43 | |
| *** igordcard has joined #openstack-ironic | 10:47 | |
| ifarkas | dtantsur|brb, hi, what was wrong with the returning-to-ramdisk implementation? | 10:49 |
| *** ifarkas has quit IRC | 10:52 | |
| *** ifarkas has joined #openstack-ironic | 10:53 | |
| GheRivero | morning all | 10:54 |
| *** Masahiro has quit IRC | 10:59 | |
| *** ramineni has quit IRC | 11:12 | |
| *** pradipta is now known as pradipta_away | 11:13 | |
| *** naohirot has joined #openstack-ironic | 11:35 | |
| *** pensu has quit IRC | 11:39 | |
| *** ramineni has joined #openstack-ironic | 11:43 | |
| *** smoriya has quit IRC | 11:52 | |
| openstackgerrit | Tan Lin proposed openstack/ironic: Add AMT-PXE-Driver to deploy cloud on PC https://review.openstack.org/135184 | 11:59 |
| *** Masahiro has joined #openstack-ironic | 12:00 | |
| *** Masahiro has quit IRC | 12:04 | |
| *** enikanorov_ has joined #openstack-ironic | 12:13 | |
| *** afaranha has joined #openstack-ironic | 12:13 | |
| *** enikanorov has quit IRC | 12:15 | |
| *** masco has quit IRC | 12:16 | |
| *** Haomeng has joined #openstack-ironic | 12:19 | |
| *** Haomeng|2 has quit IRC | 12:20 | |
| *** ramineni1 has joined #openstack-ironic | 12:33 | |
| *** dprince has joined #openstack-ironic | 12:33 | |
| *** ramineni has quit IRC | 12:35 | |
| *** dtantsur|brb is now known as dtantsur | 12:43 | |
| *** jgrimm is now known as zz_jgrimm | 12:53 | |
| *** zz_jgrimm is now known as jgrimm | 13:01 | |
| *** jcoufal_ has joined #openstack-ironic | 13:09 | |
| *** jcoufal has quit IRC | 13:12 | |
| *** masco has joined #openstack-ironic | 13:16 | |
| *** Marga_ has joined #openstack-ironic | 13:26 | |
| *** rakesh_hs has quit IRC | 13:33 | |
| *** pawel-palucki has joined #openstack-ironic | 13:36 | |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: In-band hardware properties inspection via ironic-discoverd https://review.openstack.org/135605 | 13:38 |
| sambetts | dtantsur: if i want to test one of the drivers with real hardware, is running ironic with devstack a good place to start? | 13:40 |
| dtantsur | sambetts, not perfect, but probably the best. If you have Fedora, you may try https://openstack.redhat.com/Deploying_an_RDO_Undercloud_with_Instack too. | 13:41 |
| openstackgerrit | Merged stackforge/ironic-discoverd: Support updating IPMI credentials from within ramdisk https://review.openstack.org/140814 | 13:45 |
| sambetts | I run a ubuntu env for testing normally, so I guess devstack is probably the best place to start | 13:47 |
| *** Masahiro has joined #openstack-ironic | 13:48 | |
| sambetts | dtantsur: in the local.conf does the IRONIC_ENABLED_DRIVERS just take a list of the driver class names? the devstack docs only show senarios for agent and pxe | 13:51 |
| dtantsur | sambetts, these are not class names, there are entrypoints as found in setup.cfg | 13:51 |
| sambetts | dtantsur: I see! I didn't even know that was a thing | 13:52 |
| dtantsur | :) | 13:52 |
| sambetts | :-) | 13:52 |
| dtantsur | it's really non-obvious, unless you worked with stevedore before | 13:53 |
| *** Masahiro has quit IRC | 13:53 | |
| *** pradipta_away is now known as pradipta | 13:58 | |
| *** mitz- has joined #openstack-ironic | 13:58 | |
| *** mitz_ has quit IRC | 14:00 | |
| *** rloo has joined #openstack-ironic | 14:09 | |
| *** andreykurilin_ has joined #openstack-ironic | 14:10 | |
| naohirot | good night ironic | 14:15 |
| *** naohirot has quit IRC | 14:16 | |
| *** ryanpetrello has joined #openstack-ironic | 14:16 | |
| *** kbyrne has quit IRC | 14:28 | |
| *** pawel-palucki is now known as ppalucki | 14:28 | |
| *** pawel-palucki has joined #openstack-ironic | 14:29 | |
| *** kbyrne has joined #openstack-ironic | 14:29 | |
| *** pawel-palucki has quit IRC | 14:29 | |
| *** ppalucki has quit IRC | 14:29 | |
| *** pawel-palucki has joined #openstack-ironic | 14:29 | |
| *** pawel-palucki has quit IRC | 14:29 | |
| *** ppalucki has joined #openstack-ironic | 14:29 | |
| *** kbyrne has quit IRC | 14:42 | |
| *** kbyrne has joined #openstack-ironic | 14:43 | |
| lintan | yuriyz: Hi | 14:43 |
| yuriyz | lintan hi will be later | 14:43 |
| lintan | yuriyz: OK, ping me when you are free | 14:44 |
| NobodyCam | good morning Ironic | 14:45 |
| rloo | hey, morning NobodyCam. hi lintan | 14:48 |
| NobodyCam | good morning rloo :) | 14:49 |
| dtantsur | morning NobodyCam, rloo! Hi, lintan. | 14:49 |
| lintan | Hi morning | 14:49 |
| rloo | afternoon dtantsur | 14:50 |
| NobodyCam | morning dtantsur and lintan :) | 14:50 |
| lintan | :) I am at midnight actually... | 14:50 |
| NobodyCam | heheehe :-P | 14:51 |
| *** ryanpetrello has quit IRC | 14:51 | |
| lintan | But morning for you guys rloo, NobodyCam and dtantsur | 14:51 |
| rloo | btw, dtantsur, I haven't had a chance to look again at your in-band hw properties spec. are you ok if it waits til next week (or hopefully it gets approved before then?) | 14:51 |
| rloo | lintan: midnight partying? :-) | 14:52 |
| *** Marga_ has quit IRC | 14:52 | |
| lintan | rloo: hahaha I hope so | 14:52 |
| *** ryanpetrello has joined #openstack-ironic | 14:52 | |
| *** r-daneel has quit IRC | 14:54 | |
| ramineni1 | rloo : hi | 15:01 |
| ramineni1 | rloo : morning :) | 15:01 |
| rloo | hi ramineni1 | 15:01 |
| ramineni1 | rloo: still not convinced on this spec https://review.openstack.org/#/c/129529/ :) | 15:02 |
| ramineni1 | rloo: do you still believe adding them to ManagementInterface is not useful? | 15:03 |
| rloo | ramineni1: no, not convinced. I still don't see any good reason for having explicit get/set-boot-mode methods, vs get/set-capability methods that can handle boot mode and any other capabilities. | 15:03 |
| rloo | ramineni1: probably need to get some other folks (core-reviewer) to agree to it. I'm not a core-reviewer for specs, so you don't really need to convince me ;) | 15:04 |
| dtantsur | ... but I would recommend you DO convince rloo :) | 15:06 |
| rloo | ramineni1: or convince dtantsur, and if he is convinced ... ;) | 15:06 |
| ramineni1 | rloo,dtantsur : :) , | 15:07 |
| victor_lowther | morning all | 15:07 |
| dtantsur | victor_lowther, o/ | 15:08 |
| ramineni1 | I'm rather confused myself which is the best way , :P , could use more opinions on that spec though | 15:08 |
| NobodyCam | morning victor_lowther :) | 15:08 |
| NobodyCam | morning ramineni1 | 15:08 |
| rloo | morning victor_lowther | 15:08 |
| *** ppalucki has quit IRC | 15:08 | |
| ramineni1 | morning NobodyCam :) | 15:09 |
| NobodyCam | :) | 15:09 |
| yuriyz | litan hello | 15:12 |
| yuriyz | lintan hello | 15:13 |
| lintan | yuriyz: Hi | 15:13 |
| lintan | yuriyz: I want to discuss with your comments | 15:13 |
| yuriyz | ok | 15:14 |
| lintan | yuriyz: for boot_persistent value check, I am thinking of admin input a string value to driver_info, there is no problem for config. | 15:15 |
| lintan | yuriyz: this is the first comment in common.py | 15:15 |
| yuriyz | Yes but I can set only amt_boot_device via node update | 15:15 |
| yuriyz | amt_boot_persistent will be None | 15:16 |
| NobodyCam | morning yuriyz :) | 15:17 |
| yuriyz | morning NobodyCam Ironic | 15:17 |
| NobodyCam | :) | 15:17 |
| yuriyz | my amt_boot_device will be dropped | 15:18 |
| lintan | yuriyz: So you mean update via node update, but we have a default value from conf | 15:18 |
| *** ramineni has joined #openstack-ironic | 15:19 | |
| lintan | yuriyz: it is True, so we will not dropped amt_boot_device | 15:20 |
| yuriyz | I do not see parse_driver_info() in set_next_boot_device() | 15:20 |
| yuriyz | default will come from where? | 15:21 |
| lintan | yuriyz: OK, I get your point | 15:21 |
| *** ramineni1 has quit IRC | 15:22 | |
| lintan | yuriyz: You mean you initial a node, and only update boot_device for the node | 15:22 |
| yuriyz | yes | 15:22 |
| *** ramineni1 has joined #openstack-ironic | 15:23 | |
| lintan | yuriyz: OK, I will think up for this case. But for the first comment, "Oslo config casts bool options values to boolean type, this not needed. " I am thinking of admin input a string value to driver_info, there is no problem for config. | 15:25 |
| *** anderbubble has joined #openstack-ironic | 15:25 | |
| *** ramineni has quit IRC | 15:26 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Support for non-Glance image references https://review.openstack.org/136741 | 15:26 |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Do not strip 'glance://' prefix from image hrefs https://review.openstack.org/139057 | 15:26 |
| vdrok | rloo, hi! | 15:28 |
| rloo | hi vdrok | 15:29 |
| vdrok | rloo, I pushed a new review accidentally | 15:29 |
| vdrok | rloo, https://review.openstack.org/139057 | 15:29 |
| vdrok | rloo, could you reapprove? | 15:29 |
| rloo | vdrok: I dunno... let me think about it :-) | 15:29 |
| rloo | vdrok: no changes? | 15:29 |
| vdrok | rloo, ok :) | 15:29 |
| vdrok | there were changes in https://review.openstack.org/136741 and rebase | 15:30 |
| *** jrist has quit IRC | 15:30 | |
| *** jrist has joined #openstack-ironic | 15:30 | |
| vdrok | rloo, and seems that git review didn't get it :( | 15:30 |
| rloo | vdrok: ugh. ok, give me a few secs. | 15:31 |
| vdrok | rloo, sure | 15:31 |
| rloo | vdrok: should be good now. | 15:33 |
| vdrok | rloo, thanks :) | 15:33 |
| *** Masahiro has joined #openstack-ironic | 15:37 | |
| openstackgerrit | Vladyslav Drok proposed openstack/ironic: Remove Python 2.6 from setup.cfg https://review.openstack.org/141074 | 15:38 |
| *** Masahiro has quit IRC | 15:41 | |
| lintan | rloo:Hi I see your comments, do you have time? | 15:42 |
| rloo | lintan: never any time, but sure :-) | 15:43 |
| rloo | lintan: if those driver_info properties are optional, and someone specifies them, I don't see how it is OK to delete them. | 15:44 |
| *** shakamunyi has joined #openstack-ironic | 15:45 | |
| lintan | rloo:why do we need to delete them, I think this will be deleted when delete the node | 15:45 |
| rloo | lintan: your code deletes them. you tell me why? | 15:45 |
| *** jerryz_ has quit IRC | 15:46 | |
| rloo | lintan: in management.py, set_next_boot_device(). they are being deleted there, aren't they? amt_boot_device & amt_boot_persistent. | 15:46 |
| lintan | rloo:you mean boot_device, yes, because it will read the value to set boot_device | 15:46 |
| rloo | lintan: but what if the user specified those values? the code deletes them. | 15:47 |
| jroll | morning everybody :) | 15:47 |
| lintan | rloo:if it's non persistent, we will only used the boot_device once | 15:47 |
| rloo | lintan: does the user know that? and even if you only use that value once, I don't think it means the code can delete it. | 15:47 |
| rloo | lintan: if you were a user and specified that value, and then later, looked to see what the driver_info properties were and that was missing, would that be expected? | 15:48 |
| lintan | rloo:so it means we will boot with disk in second time if we boot with pxe at first time, if user set-boot-device with pxe and non-persistent | 15:48 |
| rloo | lintan: so it might mean that you need to find a different mechanism to do what you want to do. | 15:49 |
| rloo | lintan: why do you want to allow the user to specify those values? | 15:49 |
| NobodyCam | brb | 15:50 |
| lintan | rloo:OK. I didn't expect user to see the boot_device from ironic node-show, I only expect them to see from get-boot-device | 15:50 |
| lintan | rloo:you are right, that's something I didn't consider before | 15:51 |
| rloo | lintan: can't the user use the API to set the boot device? http://docs.openstack.org/developer/ironic/webapi/v1.html#nodes | 15:52 |
| rloo | lintan: GET/SET /v1/nodes/uuid/management/boot_device | 15:52 |
| lintan | rloo:yes, they can, I expect them behavior in this way. | 15:54 |
| *** ramineni1 has quit IRC | 15:57 | |
| lintan | rloo:what about this, I didnt delete the boot_device, but I will change the boot_device to the default one, disk | 15:57 |
| rloo | lintan: the code can't (or shouldn't anyway) modify anything in driver_info. | 15:58 |
| *** pcrews has joined #openstack-ironic | 15:58 | |
| *** ifarkas has quit IRC | 15:58 | |
| *** ifarkas has joined #openstack-ironic | 15:58 | |
| lintan | rloo: this will be tough, I have to think about this | 16:03 |
| *** subscope has quit IRC | 16:03 | |
| rloo | lintan: thx for thinking about it! | 16:03 |
| lintan | rloo: Thank you :) talk to you later | 16:06 |
| NobodyCam | question are references to pxe_* and iscsi_* the same drivers? | 16:08 |
| jroll | NobodyCam: for example? | 16:09 |
| jroll | probably not | 16:09 |
| rloo | NobodyCam: do we have iscsi_* drivers? where do you see references to that? | 16:09 |
| NobodyCam | line 31 of https://review.openstack.org/#/c/136020/3/specs/kilo/irmc-management-driver.rst | 16:10 |
| jroll | I think there's iscsi_ilo | 16:10 |
| jroll | NobodyCam: in general, some drivers boot the iscsi deploy ramdisk via not-pxe methods, those tend to be iscsi_* | 16:11 |
| jroll | but idk if irmc can do that | 16:11 |
| NobodyCam | ack. I was recalling some chat about renaming the PXE_ driver to Iscsi_ | 16:12 |
| *** shakamunyi has quit IRC | 16:12 | |
| jroll | ++ | 16:12 |
| rloo | NobodyCam: I seem to recall yesterday or sometime this week, there was some sort of 'pleeeee' from devananda to rename the pxe drivers. | 16:13 |
| jroll | it's really confusing | 16:13 |
| *** jcoufal_ has quit IRC | 16:14 | |
| *** anderbubble has quit IRC | 16:23 | |
| *** anderbubble has joined #openstack-ironic | 16:25 | |
| dtantsur | folks, maybe stupid question, but can node.get call cause a Conflict? | 16:32 |
| jroll | no? | 16:33 |
| jroll | don't believe so | 16:33 |
| *** pradipta is now known as pradipta_away | 16:34 | |
| NobodyCam | brb | 16:36 |
| dtantsur | jroll, and the same for all get() and list() calls? what about validate()? | 16:36 |
| * dtantsur trying to figure out what to retry | 16:36 | |
| jroll | validate() would | 16:36 |
| jroll | any GET request will not | 16:36 |
| jroll | anything else could, I think | 16:36 |
| dtantsur | cool thanks! | 16:36 |
| *** erwan_taf_ has quit IRC | 16:37 | |
| jroll | dtantsur: not 100% sure but that should be close enough :) | 16:37 |
| *** erwan_taf has quit IRC | 16:37 | |
| *** lazy_prince is now known as killer_prince | 16:39 | |
| SpamapS | before I go off blaming all of nova for this, have any of you experienced a problem where if you delete a flavor while things are deploying, the instances in nova are now undeletable? | 16:43 |
| SpamapS | | fault | {"message": "Flavor 7 could not be found.", "code": 404, "created": "2014-12-11T10:37:51Z"} | | 16:44 |
| SpamapS | | flavor | baremetal (82f959f5-b20e-44b2-9604-f68abc19e214) | | 16:44 |
| jroll | interesting | 16:44 |
| SpamapS | confusing thing: flavor is not 7, it is 'baremetal (that-uuid)' | 16:44 |
| jroll | that's almost certainly nova | 16:44 |
| jroll | db id :P | 16:44 |
| *** Marga_ has joined #openstack-ironic | 16:44 | |
| SpamapS | jroll: I think so too, but wondered if anybody else had seen this. | 16:44 |
| jroll | we don't delete flavors for a reason | 16:44 |
| SpamapS | jroll: noooooooo | 16:44 |
| SpamapS | db ids in your errors: rookie mistake | 16:45 |
| openstackgerrit | Dmitry Tantsur proposed stackforge/ironic-discoverd: Retry on Conflict exceptions from Ironic https://review.openstack.org/141096 | 16:45 |
| *** david-lyle_afk is now known as david-lyle | 16:46 | |
| rloo | SpamapS: this seems similar: https://bugs.launchpad.net/ironic/+bug/1400269 | 16:47 |
| rloo | SpamapS: but they were trying to delete, you're deploying. | 16:48 |
| rloo | SpamapS: so I'll blame it on nova anyway ;) | 16:48 |
| SpamapS | the nova state machine is a spooky, spooky place to poke around in | 16:57 |
| SpamapS | rloo: that is exactly the problem I'm having. I am trying to delete, I just happened to delete it while they were deploying (which is not why the deploy failed) | 16:57 |
| *** dlaube has joined #openstack-ironic | 16:57 | |
| rloo | SpamapS: so it looks like a problem with the ironic driver (based on the bug description, I haven't looked into it) | 16:58 |
| *** shakamunyi has joined #openstack-ironic | 16:59 | |
| jroll | whaddya know | 17:00 |
| jroll | https://github.com/openstack/nova/blob/master/nova/virt/ironic/driver.py#L311-312 | 17:00 |
| jroll | wait why do we need extra specs | 17:00 |
| jroll | https://github.com/openstack/nova/blob/master/nova/virt/ironic/patcher.py#L106-108 | 17:00 |
| jroll | actually, https://github.com/openstack/nova/blob/master/nova/virt/ironic/patcher.py#L134-139 | 17:01 |
| * jroll puts up a patch | 17:01 | |
| *** Marga_ has quit IRC | 17:01 | |
| rloo | jroll: that was to support code in icehouse, right? | 17:01 |
| dtantsur | ifarkas, https://review.openstack.org/#/c/135589/ | 17:01 |
| ifarkas | dtantsur, thanks | 17:01 |
| *** Marga_ has joined #openstack-ironic | 17:01 | |
| jroll | oh, this'll be great | 17:02 |
| openstackgerrit | Merged openstack/ironic: Use Literal Blocks to write code sample in docstring https://review.openstack.org/139965 | 17:03 |
| *** dtantsur is now known as dtantsur|afk | 17:04 | |
| *** achanda has joined #openstack-ironic | 17:05 | |
| openstackgerrit | Merged openstack/ironic: Do not strip 'glance://' prefix from image hrefs https://review.openstack.org/139057 | 17:05 |
| *** marcoemorais has joined #openstack-ironic | 17:08 | |
| *** marcoemorais1 has joined #openstack-ironic | 17:10 | |
| SpamapS | jroll: you got a patch? I still have the affected box.. about to tear it down though. | 17:11 |
| *** marcoemorais has quit IRC | 17:13 | |
| jroll | SpamapS: yeah, working on one right now | 17:15 |
| JayF | dtantsur|afk: How are you packaging the discoverd? | 17:15 |
| jroll | just cleaning up tests | 17:15 |
| jroll | SpamapS: does tripleo use the whole "flavor has deploy ramdisk IDs" thing? | 17:16 |
| SpamapS | jroll: | extra_specs | {} | | 17:17 |
| SpamapS | jroll: ramdisk id is in the image for us. | 17:17 |
| jroll | beautiful | 17:17 |
| jroll | well | 17:17 |
| jroll | deploy ramdisk, not instance | 17:17 |
| SpamapS | It's not in flavor | 17:18 |
| jroll | cool | 17:20 |
| * jroll waiting for git review | 17:20 | |
| jroll | https://review.openstack.org/141104 | 17:20 |
| jroll | I have in no way tested this | 17:20 |
| jroll | and it's large, but should work | 17:20 |
| jroll | this is my favorite patch in recent times, <3 deleting code | 17:21 |
| jroll | devananda: pls to review https://review.openstack.org/141104 :) | 17:21 |
| SpamapS | craaaap | 17:21 |
| SpamapS | it is | 17:21 |
| SpamapS | nova flavor-key baremetal set "cpu_arch"="$ARCH" \ "baremetal:deploy_kernel_id"="$deploy_kernel_id" \ "baremetal:deploy_ramdisk_id"="$deploy_ramdisk_id" | 17:21 |
| jroll | yeah :/ | 17:21 |
| SpamapS | jroll: whats the alternative? | 17:21 |
| jroll | SpamapS: node.driver_info['pxe_deploy_ramdisk'] = $deploy_kernel_id | 17:22 |
| jroll | etc | 17:22 |
| SpamapS | per node? | 17:22 |
| jroll | the flavor thing has been deprecated, should be removed for kilo | 17:22 |
| jroll | yes | 17:22 |
| SpamapS | so every node we have to restate it every time we create node? | 17:22 |
| jroll | yes | 17:23 |
| SpamapS | that seems like busywork, but ok, is there a deprecation bug I can attach a tripleo task to? | 17:23 |
| *** masco has quit IRC | 17:23 | |
| SpamapS | jroll: I actually take that back. It sounds like resilience work. :) | 17:23 |
| jroll | ha | 17:24 |
| jroll | there were reasons | 17:24 |
| jroll | let me see if I can dig it up | 17:24 |
| SpamapS | No I think I understand them on second thought. | 17:24 |
| SpamapS | I can just open a new one if it's not readily apparent from a git annotate or something. | 17:24 |
| *** anderbubble has quit IRC | 17:24 | |
| *** Masahiro has joined #openstack-ironic | 17:26 | |
| jroll | SpamapS: yeah, not finding much | 17:27 |
| jroll | sorry | 17:27 |
| SpamapS | jroll: how about docs for the keys? | 17:28 |
| *** Masahiro has quit IRC | 17:31 | |
| jroll | SpamapS: just above here: http://docs.openstack.org/developer/ironic/deploy/install-guide.html#setup-the-drivers-for-bare-metal-service | 17:31 |
| jroll | or step 3 here http://docs.openstack.org/developer/ironic/deploy/install-guide.html#flavor-creation | 17:31 |
| rloo | jroll: was just going to mention that. i think that may be the only place where it is documented. | 17:33 |
| *** alexpilotti has joined #openstack-ironic | 17:33 | |
| jroll | rloo: right | 17:34 |
| rloo | devananda: ^^ should we update juno release notes about flavor-extra-specs-baremetal* deprecation? | 17:34 |
| *** alexpilotti has quit IRC | 17:34 | |
| *** alexpilotti has joined #openstack-ironic | 17:34 | |
| rloo | jroll: what if someone created the instances using flavors-in-icehouse, and they upgrade to juno and then upgrade to kilo, and then try to delete the instance? | 17:35 |
| SpamapS | jroll: cool thanks, FYI https://bugs.launchpad.net/tripleo/+bug/1401617 | 17:36 |
| jroll | rloo: that will still work, it would just leave the pxe_deploy_ramdisk keys on the node | 17:37 |
| rloo | jroll: ok. or we could leave code in the cleanup, to use the flavor extraspecs if that exists. | 17:37 |
| jroll | rloo: the whole point is to get rid of that code :/ | 17:38 |
| rloo | jroll: well, it is to deprecate the use of those extraspecs... | 17:39 |
| jroll | right... | 17:39 |
| rloo | jroll: i'm fine either way, but wondering if we should try to document that they pxe_deploy_ramdisk keys are left on the node in this yucky situation. | 17:39 |
| jroll | I mean | 17:41 |
| jroll | I think that's actually the best case? | 17:41 |
| jroll | then they don't need to update the node ;) | 17:41 |
| rloo | jroll: forget what I said about looking/using the flavor extra specs when deleting. that could cause diff problem | 17:41 |
| rloo | jroll: old flavor-from-icehouse used with kilo (or juno ugh), where the extraspecs is not used but exists. when deleting, it would be wrong to use the extraspecs info. | 17:42 |
| jroll | rloo: I'm not sure what you mean | 17:43 |
| devananda | morning, all | 17:43 |
| devananda | in the office today, so more randomly afk than usual | 17:43 |
| NobodyCam | good morning devananda | 17:43 |
| jroll | morning devananda | 17:43 |
| rloo | jroll: i'm assuming with kilo, someone could specify a flavor that has extra-specs-baremetal* stuff. but that isn't/wouldn't be used for deploying. | 17:44 |
| jroll | rloo: right, is that a problem? | 17:44 |
| jroll | oh I see | 17:44 |
| devananda | mmm, lots of scrollback | 17:44 |
| jroll | you're saying we don't want cleanup code | 17:44 |
| rloo | jroll: so if we left code to use that flavor-extra-specs-baremetal* stuff when deleting/cleaning those new instances, it'd be wrong. SO yeah, your patch is ok. | 17:44 |
| jroll | devananda: present for you https://review.openstack.org/#/c/141104/1 | 17:45 |
| jroll | rloo: right, cool :) | 17:45 |
| rloo | jroll: i'm saying yeah, for old instances using flavors-extra-specs-baremetal*, cleaning will leave the pxe_deploy_ramdisk keys on the node, like you said :-) | 17:46 |
| jroll | right :) | 17:46 |
| * jroll runs off to eat breakfast | 17:47 | |
| devananda | rloo: the migration script from nova-bm should have copied the flavor bits into the node.driver_info | 17:48 |
| devananda | rloo: but I dont know if anyone did that in a migration | 17:48 |
| devananda | for eg. icehouse-juno | 17:48 |
| devananda | are we actually logging a deprecation notice from the nova virt driver? | 17:49 |
| rloo | devananda: oh right, we didn't provide support for migrating icehouse. | 17:50 |
| devananda | rloo: right | 17:50 |
| devananda | jroll: so I like that patch, but a) I think it's going to break tripleo, and b) it looks like we didn't actually LOG a deprecation notice for operators for a cycle :( | 17:51 |
| jroll | devananda: right... didn't think about (b) | 17:52 |
| jroll | SpamapS filed a bug for (a) and is aware | 17:52 |
| SpamapS | (b) is something we try to keep on top of in TripleO .. so please do add that. | 17:53 |
| SpamapS | what about falling back to the flavor id only if you don't have the settings in the node? | 17:53 |
| devananda | SpamapS: i think we do that ... lemme check | 17:53 |
| *** marcoemorais1 has quit IRC | 17:54 | |
| *** Marga_ has quit IRC | 17:55 | |
| jroll | devananda: so there's an outstanding bug where destroy() fails if the flavor no longer exists | 17:55 |
| jroll | which comes from this code | 17:55 |
| jroll | can probably fix it in a different way, I just love to delete code | 17:55 |
| devananda | SpamapS: ah. other way around. if the flavor contains those keys AND the node driver is a 'pxe*' driver, then we'll copy them from the flavor to the node | 17:55 |
| *** achanda has quit IRC | 17:55 | |
| devananda | SpamapS: and delete them from the node when the instance is deleted | 17:55 |
| devananda | jroll: yup. that bug makes sense | 17:56 |
| SpamapS | devananda: ah so it being in the flavor overrides the more specific per node value? | 17:56 |
| *** achanda has joined #openstack-ironic | 17:56 | |
| devananda | normally, nova caches all the info it needs, but it doesn' tcache extra specs -- so we fetch it from the db | 17:56 |
| devananda | we've circumvented nova's means of allowing flavors to be deleted or updated w/o affecting existing instances :( | 17:56 |
| devananda | SpamapS: yes | 17:57 |
| SpamapS | devananda: I suppose the runtime flavor choice is more specific than the registered node choice. | 17:57 |
| SpamapS | anyway, we just need to register nodes w/ ids and then we cease to be a blocker | 17:57 |
| *** r-daneel has joined #openstack-ironic | 17:57 | |
| *** anderbubble has joined #openstack-ironic | 17:58 | |
| SpamapS | It's pretty awkward btw.. pulling in what used to be a flavor thing, into the hardware register-time data. | 17:58 |
| SpamapS | They are done at completely different levels of TripleO currently. | 17:58 |
| devananda | SpamapS: that it was a flavor thing is a vestige of assumping that nova is the only user of ironic | 17:59 |
| devananda | SpamapS: we can leave this glue code in the nova driver if it's important to people to preserve taht use case | 17:59 |
| SpamapS | devananda: I'm saying that perhaps ironic should have its own generic information that gets used by nodes. | 17:59 |
| devananda | SpamapS: as in, "set this metadata for that group of nodes" ? | 18:00 |
| SpamapS | Yeah | 18:00 |
| NobodyCam | JoshNang: are you around? | 18:00 |
| *** jistr has quit IRC | 18:00 | |
| devananda | because it's very likely to have >1 deploy ramdisk in use in an environment | 18:00 |
| JoshNang | NobodyCam: yup! what's up? | 18:00 |
| *** achanda has quit IRC | 18:00 | |
| devananda | SpamapS: so we can't have it as a global config | 18:00 |
| SpamapS | I'm saying Yes to M:N, no to 1:1 | 18:01 |
| devananda | SpamapS: I don't see this as a chassis-level setting either, fwiw, so I'm not sure where to fit it | 18:01 |
| devananda | we don't have arbitrary tagging / grouping today | 18:01 |
| devananda | which would sort of cover this, maybe? | 18:01 |
| NobodyCam | JoshNang: I was looking at cleaning spec. https://review.openstack.org/#/c/102685 on line 150 what data would be stored there? | 18:01 |
| SpamapS | deploy groups maybe | 18:01 |
| devananda | group N = {key: deploy_ramdisk_id, value: nnnn} | 18:02 |
| devananda | group M = {key: ipmi_password, value: foo} | 18:02 |
| devananda | node UUUU is_a_member_of: N, M | 18:02 |
| devananda | the ilo driver folks have asked for something similar | 18:02 |
| SpamapS | devananda: Rather than trying to imagine all the reasons one might want a different deploy kernel/ramdisk for groups of servers sized between 1 and some unknown upper bound, we could just call it deploy groups and let people set values for the node groups that are going to be the same. | 18:02 |
| devananda | so they can associate credentials with all the cartridges in a chassis | 18:03 |
| *** harlowja_away is now known as harlowja | 18:03 | |
| SpamapS | devananda: this also helps with things like the SSH jump capability we had talked about. | 18:03 |
| devananda | yup | 18:03 |
| JoshNang | NobodyCam: hrm perhaps poorly worded. one example is the agent has a version string for its hardware manager, and if that changes (reboot to a new version part way through cleaning), we current restart cleaning | 18:03 |
| SpamapS | crap I hate when we have good ideas, because I have no time for good ideas | 18:03 |
| JoshNang | so storing that string would be useful, and i think makes sense for driver_fino | 18:04 |
| JoshNang | info, even | 18:04 |
| devananda | SpamapS: I don't imagine that any of this will happen in kilo ... yea, ENOTIME | 18:04 |
| devananda | SpamapS: but I like the idea | 18:04 |
| NobodyCam | JoshNang: I was thinking extra_info | 18:04 |
| SpamapS | devananda: we're going to need the SSH jump host thing very soon as a POC. | 18:04 |
| NobodyCam | JoshNang: but let think on it :) | 18:05 |
| devananda | SpamapS: that should just be a new driver | 18:05 |
| devananda | SpamapS: rather than trying to split the conductor into two services | 18:05 |
| NobodyCam | SpamapS: a bridging ssh driver | 18:05 |
| devananda | SpamapS: also, Fuel team seems to be doing something sort of similar -- they want to SSH into a host directly to control its power state | 18:05 |
| JoshNang | NobodyCam: ok! i chose driver_info because cleaning steps would be functions of the drivers, but i'm ok with either | 18:05 |
| SpamapS | devananda: it's a new driver that will be a superset of IPMI | 18:05 |
| SpamapS | since it needs to basically do ssh -L xxx:$ipmitarget:$ipmi_port $jumphost and then just do ipmi with $localhost:xxx as ipmi target | 18:06 |
| *** derekh has quit IRC | 18:07 | |
| SpamapS | devananda: ssh in and self-power-control seems a bid unrealistic.. but o-k. ;) | 18:07 |
| jroll | JoshNang: NobodyCam: I don't think ironic should touch node.extra, in general | 18:07 |
| devananda | SpamapS: right. either this should be a new power driver, or it might fit into the command-set of the existing ssh driver | 18:08 |
| devananda | SpamapS: that's basically a dict of commands to run after you SSH to a remote host to manage the power staet of other things | 18:09 |
| devananda | SpamapS: currently it only makes virsh/vbox/vmware calls on local host. i THINK it would work with ipmitool too | 18:09 |
| devananda | NobodyCam: ^^ ??? | 18:09 |
| SpamapS | devananda: hm, I wonder if we can dictate that the remote host has ipmitool on it. THat might actually be enough for POC | 18:09 |
| jroll | I don't see why it wouldn't work with ipmi | 18:09 |
| jroll | tool | 18:10 |
| devananda | gah. nope. | 18:12 |
| devananda | needs a new driver | 18:12 |
| devananda | the ssh power driver's control flow is predicated on fetching a list of all || running instances on a host | 18:12 |
| devananda | so that it knows what it can manage | 18:12 |
| jroll | that command could be "echo localhost" | 18:13 |
| *** achanda has joined #openstack-ironic | 18:14 | |
| devananda | jroll: https://git.openstack.org/cgit/openstack/ironic/tree/ironic/drivers/modules/ssh.py#n379 | 18:14 |
| jroll | ewwwww | 18:15 |
| devananda | new driver | 18:15 |
| *** pelix has quit IRC | 18:16 | |
| devananda | SpamapS: so. I think such a "remoet ipmitool" driver would be great. who's got time to write it? :) | 18:16 |
| *** pelix has joined #openstack-ironic | 18:16 | |
| SpamapS | devananda: I think I'll be writing it for a business need. | 18:17 |
| NobodyCam | devananda: SpamapS: would the ipmi listener daemon thingy that jjohnson2 is working on solve the issue? | 18:17 |
| SpamapS | devananda: or somebody on the Helion installer team. | 18:18 |
| SpamapS | NobodyCam: maybe? | 18:18 |
| devananda | NobodyCam: I don't think so? that listens for ipmi packets, right? | 18:18 |
| SpamapS | NobodyCam: we basically have a requirement to be able to do provisioning in places where ILO is only available behind jump hosts. | 18:18 |
| devananda | SpamapS: is the requirement specifically for iLO support, or just a basic set of ipmi? | 18:19 |
| SpamapS | SSH is always an option. IPMI listener relay packet things.. probably a stretch. | 18:19 |
| SpamapS | devananda: ipmi is mandated, even though ILO is likely to be there. | 18:19 |
| SpamapS | I just say ILO because I'm a company man. ;) | 18:19 |
| devananda | cool | 18:19 |
| devananda | *cough* | 18:19 |
| NobodyCam | oh so ssh bridging support like we have ipmi bridging support | 18:19 |
| *** Marga_ has joined #openstack-ironic | 18:19 | |
| devananda | SpamapS: ok. so, you probably know the drill. start with a spec. drivers which don't touch other code are usually pretty fast to review | 18:20 |
| *** Marga_ has quit IRC | 18:21 | |
| SpamapS | devananda: I'll actually do that in parallel with a POC in virt and then deploy in-situ for evaluation by intended consumer. :) | 18:21 |
| *** pelix has quit IRC | 18:22 | |
| SpamapS | which I'm also doing for the L3 stuff | 18:22 |
| NobodyCam | SpamapS: are you able to use ipmi on the jump host? | 18:22 |
| SpamapS | NobodyCam: perhaps. | 18:23 |
| devananda | SpamapS: ack | 18:24 |
| SpamapS | NobodyCam: probably the simplest way to go forward if that is possible. If not, we have to tunnel it. | 18:24 |
| NobodyCam | SpamapS: and the current bridging support will not work? https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/ipmitool.py#L79-L91 | 18:24 |
| SpamapS | NobodyCam: I don't think they want IPMI to be directly contacted from anywhere outside the management network. | 18:26 |
| NobodyCam | sorry if I came into the converstion late | 18:26 |
| SpamapS | NobodyCam: ICMB has never been mentioned so I don't think they've considered it. Might be worth asking. | 18:27 |
| SpamapS | IPMB? too many I acronyms | 18:27 |
| devananda | ipmi double-bridging is a separate thing | 18:28 |
| *** Nisha has joined #openstack-ironic | 18:28 | |
| devananda | where you federate BMCs | 18:28 |
| devananda | one BMC controls several others | 18:28 |
| *** vipul has quit IRC | 18:28 | |
| devananda | OR where one BMC controls many devices | 18:28 |
| devananda | (it does both ... hence why there are like 5 add'l options for double bridging...) | 18:29 |
| devananda | but I dont think that helps in the case where your IPMI network is walled off, and the only way in is to SSH to a bastion host | 18:29 |
| *** achanda has quit IRC | 18:30 | |
| *** achanda has joined #openstack-ironic | 18:30 | |
| *** athomas has quit IRC | 18:31 | |
| *** romcheg has quit IRC | 18:31 | |
| *** romcheg has joined #openstack-ironic | 18:32 | |
| *** jcoufal has joined #openstack-ironic | 18:32 | |
| rloo | hi NobodyCam, wrt 139215, the unit tests for fsm. how much/detail do we want to review those? Are those copied from taskflow? | 18:32 |
| *** achanda has quit IRC | 18:33 | |
| *** achanda has joined #openstack-ironic | 18:33 | |
| *** marcoemorais has joined #openstack-ironic | 18:33 | |
| NobodyCam | rloo: yes that is a port of https://github.com/openstack/taskflow/blob/master/taskflow/tests/unit/test_types.py | 18:33 |
| *** marcoemorais1 has joined #openstack-ironic | 18:35 | |
| *** vipul has joined #openstack-ironic | 18:35 | |
| Nisha | devananda, (off topic here). with respect to comment on introspection spec, one comment is there..."what is the BODY of the PUT request?" What is the expectation? is it the syntax of PUT request? or something else.... | 18:36 |
| *** romcheg has quit IRC | 18:36 | |
| rloo | NobodyCam: I wonder if we should mention where the original files are. I think I'll skip reviewing the tests stuff. for now anyway. i really want to see the changes to our code. | 18:37 |
| *** marcoemorais has quit IRC | 18:38 | |
| *** jb11211 has joined #openstack-ironic | 18:38 | |
| devananda | NobodyCam: I share dtantsur|afk 's sense that this is awkward | 18:38 |
| devananda | NobodyCam: but that might be more feedback we should give to harlowja :) | 18:38 |
| devananda | Nisha: yes. an HTTP PUT request must have a body | 18:39 |
| devananda | Nisha: according to the RFC, the body MAY be empty. but we should be very clear in the spec what is expected | 18:39 |
| harlowja | devananda awkward, whatttt | 18:40 |
| harlowja | lol | 18:40 |
| Nisha | devananda ok. i will mention that in the spec | 18:40 |
| Nisha | thanks | 18:40 |
| harlowja | u guys can just help https://github.com/harlowja/automaton get moving ;) | 18:40 |
| devananda | Nisha: do we expect a PUT with a BODY containing some JSON, eg {key: value}? or do we expect a PUT with Content-length: 0 and no body? | 18:40 |
| devananda | Nisha: thanks | 18:40 |
| harlowja | and then u don't need to duplicate stuff ;) | 18:40 |
| harlowja | cough https://github.com/harlowja/automaton/blob/master/automaton/machines.py cough | 18:40 |
| NobodyCam | devananda: the tests them self of the util file stuff? | 18:40 |
| harlowja | cough https://github.com/harlowja/automaton/tree/master/automaton/tests cough | 18:41 |
| devananda | harlowja: lol. i get it :) | 18:42 |
| *** yjiang5 is now known as yjiang5_away | 18:43 | |
| Nisha | devananda, it looks like def put(self, node_uuid, discover): | 18:44 |
| Nisha | where discover is actually None because there is nothing to pass from CLI in this case for patch | 18:44 |
| devananda | harlowja: oh good. its on pip | 18:44 |
| harlowja | devananda soemthing ; ya | 18:45 |
| devananda | harlowja: I'll take a shot at pulling that in. however, to do so, it'll need to be in requirements .... | 18:45 |
| Nisha | while PATCH expects a dictionary (key:value) to be passed | 18:45 |
| harlowja | devananda ya, i just made it last weekend, so .. | 18:45 |
| harlowja | for those who aren't getting the 'cough' reference | 18:45 |
| harlowja | http://i.imgur.com/Y0bn9HD.gif | 18:45 |
| harlowja | lol | 18:45 |
| Nisha | devananda, by BODY of PUT request, do you mean above syntax? | 18:45 |
| NobodyCam | harlowja: lol | 18:45 |
| *** achanda has quit IRC | 18:45 | |
| devananda | Nisha: no. that's the python method definition | 18:45 |
| devananda | Nisha: but I think you just said that there is no body | 18:46 |
| devananda | "is actually None because ..." | 18:46 |
| *** achanda has joined #openstack-ironic | 18:46 | |
| Nisha | you mean the passed arguments? | 18:46 |
| harlowja | devananda that code has the runners split off from the main FSM | 18:46 |
| *** Marga_ has joined #openstack-ironic | 18:46 | |
| harlowja | ex @ https://github.com/harlowja/automaton/blob/master/automaton/machines.py#L331 | 18:46 |
| devananda | harlowja: yup. already read it | 18:47 |
| harlowja | devananda i also added in ' _pre_process_event' | 18:47 |
| harlowja | and such | 18:47 |
| harlowja | since i think that will be needed for u | 18:47 |
| devananda | harlowja: yea. so I think librarification of this is great | 18:47 |
| harlowja | agreed | 18:47 |
| devananda | harlowja: but, you know, it's not ready to librarify after just 4 days :) | 18:48 |
| harlowja | sure, which is why i'm ok with not using it yet, but at least the possiblitiy is there... | 18:49 |
| harlowja | didn't santa claus create the world in 4 days, so seems like a library could be made in 4 days | 18:50 |
| harlowja | or was that the flying spahhettin monster | 18:50 |
| harlowja | i can't remember | 18:50 |
| devananda | NobodyCam: a note in the unit test (and utils file) that this is a port from TaskFlow, and we plan to continue to keep in sync and eventually libify it // delete it from ironic | 18:50 |
| harlowja | * spaghetti (how dare i) | 18:50 |
| devananda | NobodyCam: might help alleviate some of the concerns that the unit test structure isn't quite normal for our code ... | 18:50 |
| *** achanda has quit IRC | 18:51 | |
| devananda | harlowja: mmm, spaghetti | 18:51 |
| harlowja | lol | 18:51 |
| NobodyCam | ya fsm! (flying spaghetti monster) | 18:51 |
| * NobodyCam loves pastafarians | 18:51 | |
| harlowja | lol | 18:52 |
| harlowja | the fsm.py now has a new name | 18:53 |
| harlowja | ha | 18:53 |
| NobodyCam | lol... | 18:53 |
| rloo | harlowja: needs a mascot... | 18:53 |
| harlowja | http://i2.kym-cdn.com/entries/icons/original/000/000/206/Flying_Spaghetti_Monster.jpg ? | 18:54 |
| NobodyCam | http://upload.wikimedia.org/wikipedia/commons/7/76/FSM_Logo.svg | 18:54 |
| harlowja | but that one might already be taken | 18:54 |
| rloo | NobodyCam: would be good to put a note in fsm.py too, that it is a port... | 18:54 |
| NobodyCam | ya | 18:54 |
| rloo | thx NobodyCam | 18:54 |
| NobodyCam | will add right after a quick smoke... hhehehe...brb | 18:55 |
| devananda | stepping afk for a bit ... | 18:55 |
| *** achanda has joined #openstack-ironic | 18:56 | |
| *** harlowja has quit IRC | 19:00 | |
| *** Marga_ has quit IRC | 19:04 | |
| *** Marga_ has joined #openstack-ironic | 19:06 | |
| *** anderbubble has quit IRC | 19:09 | |
| NobodyCam | j* hows the weather up in the bay area.. news makes it look nasty | 19:10 |
| JayF | we're all huddled up in our house | 19:10 |
| JayF | most of my apt complex has no power, but I was spared | 19:10 |
| NobodyCam | oh wow | 19:11 |
| dlaube | road in front of my house in santa cruz mountains -> https://dl.dropboxusercontent.com/u/7057904/WinterStorm-Dec-11-2014-road1.jpg | 19:11 |
| JayF | dlaube: 280 near South SF + San Bruno is completely flooded in the northerly direction | 19:11 |
| NobodyCam | dlaube: wow thats a road and not a river | 19:11 |
| JayF | dlaube: and I've seen pohtos of El Camino Real in the same area in the same state | 19:12 |
| dlaube | JayF: omg! I used to drive that everyday | 19:12 |
| JayF | dlaube: I live in Parkmerced, so I take the right turn out of my complex onto 280 all the time :) I take that drive often as well | 19:12 |
| dlaube | NobodyCam: I know right.. almost makes you think about getting a fishing pole… or life raft | 19:13 |
| NobodyCam | :-p | 19:13 |
| NobodyCam | git status | 19:14 |
| NobodyCam | d'oh ww | 19:14 |
| *** Masahiro has joined #openstack-ironic | 19:15 | |
| openstackgerrit | Merged openstack/ironic: Updated from global requirements https://review.openstack.org/140941 | 19:17 |
| openstackgerrit | Merged openstack/ironic: Correct vmware ssh power manager https://review.openstack.org/139810 | 19:17 |
| *** marcoemorais1 has quit IRC | 19:18 | |
| *** Masahiro has quit IRC | 19:19 | |
| *** rlr0329 has joined #openstack-ironic | 19:21 | |
| *** romcheg has joined #openstack-ironic | 19:22 | |
| afaranha | Is the installation guide updated to latest version of ironic? http://docs.openstack.org/developer/ironic/deploy/install-guide.html | 19:25 |
| afaranha | It seems that ironic-dbsync doesn't have the create_schema function anymore | 19:26 |
| afaranha | do anyone knows about this? | 19:26 |
| NobodyCam | huh | 19:26 |
| NobodyCam | we did | 19:26 |
| NobodyCam | rloo: something like this? (or more detail) :# Utilities ported from TaskFlow project used by test_fsm.py. It is our plan to | 19:27 |
| NobodyCam | # continue to keep these in sync and remove once they have been Libified. | 19:27 |
| rloo | afaranha: I see 'create_schema' in cmd/dbsync.py | 19:27 |
| NobodyCam | afaranha: https://github.com/openstack/ironic/blob/master/ironic/cmd/dbsync.py#L77 | 19:28 |
| rloo | NobodyCam: yeah, that's fine with me. maybe instead of 'It is our plan', a TODO? | 19:28 |
| afaranha | I got error when tried this: ironic-dbsync: error: argument command: invalid choice: 'create_schema' (choose from 'upgrade', 'downgrade', 'stamp', 'revision', 'version') | 19:28 |
| rloo | NobodyCam: do people want links to the the original files? | 19:28 |
| rloo | afaranha: what version of ironic code are you using? | 19:28 |
| afaranha | rloo, 2014.1.rc1 | 19:29 |
| rloo | afaranha: i seem to recall that the 'create_schema' is relatively new. | 19:29 |
| afaranha | rloo, Ah, that's why | 19:29 |
| rloo | afaranha: ok. that's icehouse. which is an older version of ironic. | 19:29 |
| NobodyCam | https://github.com/openstack/ironic/commit/e76932a0730abfd98092b4b9cbadd3a90cb32fb6 | 19:29 |
| rloo | afaranha: the install guide reflects what is on master branch. | 19:29 |
| NobodyCam | 2014.2.rc3 | 19:29 |
| rloo | afaranha: i think you can just invoke ironic-dbsync | 19:29 |
| devananda | afaranha: that,s an icehouse release candidate. about 8 months old ... | 19:30 |
| devananda | afaranha: any reason you're not using 2014.2? | 19:30 |
| rloo | afaranha: i think we all suggest you use the latest and greatest ;) | 19:30 |
| afaranha | rloo, devananda I installed it following the guide: apt-get install ironic-api ironic-conductor python-ironicclient | 19:30 |
| afaranha | how can I download the latest one? | 19:30 |
| devananda | afaranha: ah. what distro/version? | 19:30 |
| NobodyCam | pip install :) | 19:30 |
| afaranha | NobodyCam, Gonna try that :) | 19:31 |
| afaranha | devananda, it's Ubuntu 14.04 | 19:31 |
| devananda | afaranha: ah. what distro/version? | 19:31 |
| devananda | hm, soubuntu 14.04 is probably going to stay at 2014.1.rc1 | 19:31 |
| devananda | (sorry for double send) | 19:32 |
| devananda | we should u\ | 19:32 |
| devananda | bah... keyboard.... | 19:32 |
| afaranha | devananda, I think I can install it using pip and informing the version | 19:32 |
| devananda | we should update the docs to indicate that that distro support is pinned | 19:32 |
| afaranha | devananda, Any compatibility issue with ubuntu 14.04? | 19:32 |
| devananda | afaranha: pip++ | 19:32 |
| afaranha | devananda, ok, Thanks! | 19:33 |
| afaranha | devananda, It's also good to indicate how to install with pip, I'm gonna help on that :) | 19:35 |
| rlr0329 | I have deployed ironic with devstack using the stable/juno branch. Can I use this to deploy images on a baremetal node? Any doc on how to do this? The install-guide at http://docs.openstack.org/developer/ironic/deploy/install-guide.html doesn't seem to cover the final steps. All I see in the http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html#deploying-ironic-with-devstack are instructions using VMs. Any | 19:35 |
| NobodyCam | afaranha: awesome I would love to rewview a doc patch :) | 19:37 |
| NobodyCam | rlr0329: configure a ipmi driven node? | 19:38 |
| rlr0329 | yes | 19:39 |
| NobodyCam | rlr0329: have you read http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html#exercising-the-services-locally | 19:40 |
| NobodyCam | see the example walkthrough of creating a node: in htat section | 19:41 |
| rlr0329 | I guess I missed that before. Will try that now. Thanks! | 19:43 |
| devananda | jroll: i khnow you'd like to see the fsm patch series land so we can keep that moving | 19:44 |
| devananda | jroll: have you seen what i added last night? is there something else on your mind you'd like to see right away? | 19:44 |
| jroll | devananda: I've been busy downstream, I haven't/don't have time to look right now | 19:44 |
| devananda | NobodyCam: what would you say to me stripping fsm.py and the test cases down to a bare minimum at this point? there is still a bunch of code in there we're not using, copied from taskflown | 19:48 |
| rloo | devananda: I added a comment to 139216 -- i might be wrong? anyway, i'll look more later. need to take a break. | 19:48 |
| rlr0329 | NobodyCam: Looking at that walkthrough, how do I specify the image information? Using the -i <key=value>? | 19:49 |
| NobodyCam | devananda: sure. ++ less porting == less upkeep | 19:49 |
| devananda | rloo: oh, interesting. i'll check | 19:49 |
| NobodyCam | rlr0329: see step #3 of http://docs.openstack.org/developer/ironic/deploy/install-guide.html#flavor-creation | 19:52 |
| NobodyCam | devananda: I'll stop working on the fix patch to fsm and the utils file untill your done | 19:53 |
| rlr0329 | NobodyCam: Got it thanks! | 19:54 |
| NobodyCam | :) | 19:55 |
| afaranha | NobodyCam, devananda what's the command to install the latest version form pip? I tried downloading from github, and sudo pip install ironic==2014.2.rc3 , but didn't work | 19:58 |
| afaranha | actually, the pypi page doesn't have any link for ironic: https://pypi.python.org/simple/ironic/ | 20:02 |
| *** dprince has quit IRC | 20:03 | |
| openstackgerrit | Adam Gandelman proposed openstack/ironic: Do not pass PXE net config from bootloader to ramdisk https://review.openstack.org/141148 | 20:04 |
| devananda | afaranha: if you want latest stable release from your distro, you would do that through apt, and should get 2014.2 (no "rc" in the name) | 20:04 |
| devananda | afaranha: if you want latest development code, git clone ... then pip install that | 20:05 |
| devananda | or python setup.py install | 20:05 |
| afaranha | all right, gonna try that | 20:05 |
| devananda | or python setup.py install only client libraries | 20:05 |
| devananda | sorry for the miscommunication earlier | 20:05 |
| *** Nisha_away has joined #openstack-ironic | 20:05 | |
| NobodyCam | ya... our quick start doc could use a install section | 20:06 |
| *** Nisha has quit IRC | 20:06 | |
| *** anderbubble has joined #openstack-ironic | 20:08 | |
| afaranha | strange... I did python setup.py intall and got the version: 2015.1.dev206.g2db2659 | 20:11 |
| afaranha | devananda, maybe ubuntu 14.04 have no compatibility with it..... | 20:11 |
| afaranha | but I quite sure it worked | 20:14 |
| NobodyCam | afaranha: youran that from the directory you cloned into? | 20:15 |
| afaranha | yes, the version is not the 2014.2, but ironic-dbsync has the create_scheme option | 20:15 |
| NobodyCam | yes cloning gets you the latest version | 20:16 |
| afaranha | As I already run ironic-dbsync, I just upgraded it | 20:16 |
| *** dprince has joined #openstack-ironic | 20:23 | |
| adam_g | do we make any guarantees with regard to the stability of the interface we provide to ramdisks? that is, the non-configurable kernel parameters set in the various template files | 20:26 |
| devananda | adam_g: no. afaik, operators often customize that anyway | 20:27 |
| lifeless | I think adam_g means the parameters ironic injects ? | 20:27 |
| lifeless | vs the ones in the templates which as you say are under effectively under operator control | 20:28 |
| *** mrda-away is now known as mrda | 20:28 | |
| adam_g | we allow user-specified things to be injected via config file | 20:28 |
| adam_g | afaics the actual templates we ship are source | 20:28 |
| mrda | Morning Ironic | 20:28 |
| NobodyCam | morning mrda | 20:28 |
| adam_g | we currently pass network config to ramdisks via an ip= kernel parameter, which i'm hoping to remove.. but worry that will break some users who may be using a custom ramdisk, and will definitely break people using out of date DIB deploy-ramdisk | 20:29 |
| adam_g | *that ip= parameter is hard-coded in templates | 20:30 |
| devananda | adam_g: where? | 20:30 |
| devananda | i dont see that in pxe config template | 20:30 |
| adam_g | devananda, that one uses 'ipappend 3' which causes syslinux to inject it, the otehr *pxe* ones hard-code it since those bootloaders dont support that | 20:32 |
| adam_g | the deploy-ironic ramdisk manually configures networking when ip= parameter is present, under the assumption its not a PXE/DHCP-able env, instead of DHCP'ing | 20:32 |
| adam_g | i *assume* the other templates followed suite in their templates since DHCP'ing from the ramdisk is currently broken and they couldn't get it to work | 20:33 |
| *** marcoemorais has joined #openstack-ironic | 20:34 | |
| devananda | adam_g: agent_config.template does not have ipappend line | 20:34 |
| devananda | ditto for the ipxe config | 20:34 |
| adam_g | devananda, because their ramdisk probably does the correct thing and DHCP for net config by default | 20:34 |
| devananda | which specifivcally calls out dhcp in there | 20:34 |
| devananda | right | 20:35 |
| adam_g | ipxe does | 20:35 |
| adam_g | ip=${ip}:${next-server}:${gateway}:${netmask} BOOTIF=${mac} | 20:35 |
| adam_g | which is effectively what syslinux injects when 'ipappend 3' is there | 20:36 |
| jroll | PaulCzar: ^ you're likely interested in this conversation | 20:36 |
| devananda | aah. yea, there it is | 20:36 |
| jroll | PaulCzar: specifically "since DHCP'ing from the ramdisk is currently broken" | 20:36 |
| adam_g | elilo_efi* does ip=%I:{{pxe_options.tftp_server}}:%G:%M:%H::on", and im not sure what the %H::on is used for, which worries me | 20:36 |
| adam_g | jroll, PaulCzar deploying when DHCP'ing from the current DIB-generated deploy-ironic doesn't work, but thats a DIB bug | 20:37 |
| jroll | adam_g: right, Paul was frustrated about it | 20:38 |
| *** marcoemorais1 has joined #openstack-ironic | 20:38 | |
| NobodyCam | brb | 20:38 |
| adam_g | i'd like to just fix that DIB bug, and remove the passing of ip= from our templates.. but that can break anyone using something custom, and would require others to generate a new deploy-ironic ramdisk with a minimum version of DIB during upgrade | 20:40 |
| *** marcoemorais has quit IRC | 20:41 | |
| devananda | adam_g: i'd like that, too. it should be captured in release notes, but i think that's a reasonable thing to expect anyone following trunk to handle | 20:43 |
| *** alexpilotti has quit IRC | 20:49 | |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Refactor async helper methods in conductor/manager.py https://review.openstack.org/139217 | 20:51 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Begin using the state machine for node deploy/teardown https://review.openstack.org/139216 | 20:51 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Convert drivers to use process_event() https://review.openstack.org/140883 | 20:51 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Enable async callbacks from task.process_event() https://review.openstack.org/140868 | 20:51 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Convert check_deploy_timeout to use process_event https://review.openstack.org/140869 | 20:51 |
| openstackgerrit | Devananda van der Veen proposed openstack/ironic: Add base state machine https://review.openstack.org/139215 | 20:51 |
| devananda | NobodyCam: if we add more tests, let's do it in a new patch | 20:51 |
| devananda | so we dont need to rebase all of those every time | 20:52 |
| devananda | rloo: i still need to address your comment on the second patch | 20:52 |
| devananda | but have a call in a few min | 20:52 |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic-specs: Discover node properties using new CLI node-inspect https://review.openstack.org/100951 | 20:52 |
| rloo | devananda: o | 20:53 |
| rloo | devananda: k | 20:53 |
| *** harlowja has joined #openstack-ironic | 20:53 | |
| NobodyCam | devananda: ack :) | 20:55 |
| NobodyCam | adam_g: your updating DIB to remove the need for IP kernel param | 20:56 |
| adam_g | NobodyCam, well, it looks like its currently meant to work without it but the only place it derives the boot servers address is from the ip= param, so deployment doesn't work. going to push something that finds that address via dhclient in the DHCP case | 21:00 |
| adam_g | NobodyCam, if that works, we can drop the parameter.. or see about making DHCP the default, first try behavior for DIB ramdisks, in which case we can keep it to be safe | 21:01 |
| *** ChuckC has quit IRC | 21:02 | |
| NobodyCam | ok :) | 21:02 |
| *** Masahiro has joined #openstack-ironic | 21:03 | |
| *** andreykurilin_ has quit IRC | 21:06 | |
| *** Masahiro has quit IRC | 21:08 | |
| *** dprince has quit IRC | 21:14 | |
| *** romcheg has quit IRC | 21:22 | |
| *** romcheg has joined #openstack-ironic | 21:22 | |
| dlaube | hey guys, can someone remind me which one is evil/unsupported in ironic? it is ipmitool or freeipmi? | 21:24 |
| *** ChuckC has joined #openstack-ironic | 21:25 | |
| dlaube | about to reinstall my ipmitool after seeing some strainge problems | 21:25 |
| NobodyCam | we use ipmitool | 21:30 |
| NobodyCam | there is a startup check for version support | 21:31 |
| jroll | how do I make tox (testr?) print stdout for test failures? | 21:31 |
| jroll | I feel like there's an option I might be missing | 21:32 |
| *** Marga_ has quit IRC | 21:34 | |
| NobodyCam | jroll: that should print out on each failure: http://paste.openstack.org/show/23j5T7QvFgn0BXZkVa9q | 21:34 |
| openstackgerrit | Yuriy Zveryanskyy proposed openstack/ironic-specs: Add a new driver for Fuel Agent https://review.openstack.org/138115 | 21:34 |
| NobodyCam | are you adding debuging print statments :-p | 21:34 |
| jroll | maaaaaybe | 21:35 |
| NobodyCam | lol | 21:35 |
| NobodyCam | they should show up in the stdout section | 21:35 |
| jroll | yeah | 21:36 |
| NobodyCam | are you adding to the test file or the code its testing? | 21:36 |
| *** PaulCzar has quit IRC | 21:36 | |
| clif_h | I'm doing both | 21:39 |
| *** EmilienM is now known as EmilienM|pto | 21:39 | |
| NobodyCam | hi clif_h :) do you see the stdout put in a tox test failure? | 21:40 |
| clif_h | this is in nova and not ironic if that makes any difference | 21:40 |
| clif_h | ah | 21:41 |
| clif_h | it does show me the output if it fails | 21:41 |
| clif_h | but not if it passes | 21:41 |
| clif_h | which makes some sense I suppsoe | 21:42 |
| jroll | oh, it showed up now? | 21:42 |
| jroll | lol | 21:42 |
| clif_h | a pass is meh, while a fail is omg | 21:42 |
| clif_h | NobodyCam: thanks :) | 21:43 |
| NobodyCam | :) | 21:43 |
| Shrews | jroll: i did that a long time ago. i think you have to set an env variable | 21:44 |
| Shrews | OS_STDOUT_CAPTURE or something | 21:44 |
| Shrews | though i could be wrong | 21:44 |
| jroll | could be | 21:44 |
| Shrews | s/could be/probably am/ | 21:45 |
| NobodyCam | bbiab ... food time | 21:51 |
| *** Marga_ has joined #openstack-ironic | 21:52 | |
| *** hemna has joined #openstack-ironic | 21:59 | |
| openstackgerrit | Yuriy Zveryanskyy proposed openstack/ironic-specs: Add a new driver for Fuel Agent https://review.openstack.org/138115 | 22:01 |
| *** yjiang5_away is now known as yjiang5 | 22:02 | |
| *** marcoemorais1 has quit IRC | 22:12 | |
| *** Nisha_away has quit IRC | 22:21 | |
| *** ryanpetrello has quit IRC | 22:22 | |
| *** ifarkas has quit IRC | 22:23 | |
| *** romcheg has quit IRC | 22:26 | |
| clif_h | To anyone familiar with the nova ironic driver, I'm trying to add behavior to driver to cache the ironic-client instead of reconstructing it for every call. Please take a look: https://review.openstack.org/#/c/141188/ | 22:26 |
| JayF | mrda: ^ you probably have a particular interest in https://review.openstack.org/#/c/141188/ | 22:27 |
| *** jrist has quit IRC | 22:29 | |
| *** jrist has joined #openstack-ironic | 22:30 | |
| mrda | Thanks JayF | 22:32 |
| *** ChuckC has quit IRC | 22:32 | |
| JayF | mrda: btw, if you haven't met him yet, clif_h is the newest 'dentist' | 22:33 |
| mrda | ahh, right | 22:33 |
| mrda | hi clif_h! | 22:33 |
| * mrda feels a little guilty for not closing this bug himself | 22:33 | |
| clif_h | mrda: greetings! | 22:33 |
| clif_h | don't feel guilty, it gave me something small to chew on | 22:34 |
| mrda | It fell off my plate, but I'm glad you picked it up :) | 22:34 |
| JayF | clif_h: don't absolve his guilt, redirect it to a harder bug! | 22:34 |
| JayF | ;) | 22:34 |
| mrda | JayF: There will be no more schwag for you ;) | 22:35 |
| *** marcoemorais has joined #openstack-ironic | 22:37 | |
| *** marcoemorais1 has joined #openstack-ironic | 22:40 | |
| *** marcoemorais has quit IRC | 22:41 | |
| *** achanda has quit IRC | 22:44 | |
| *** achanda_ has joined #openstack-ironic | 22:47 | |
| *** achanda_ has quit IRC | 22:51 | |
| *** andreykurilin_ has joined #openstack-ironic | 22:51 | |
| *** achanda has joined #openstack-ironic | 22:51 | |
| *** Masahiro has joined #openstack-ironic | 22:52 | |
| *** achanda has quit IRC | 22:56 | |
| *** Masahiro has quit IRC | 22:57 | |
| *** romcheg has joined #openstack-ironic | 22:58 | |
| JayF | https://review.openstack.org/#/c/138116/ could use another core review so it can land | 22:59 |
| JayF | just some oslo cleanup stuff for IPA we should get in before it gets stale (again, I already had to rebase and fix a couple of things once) | 22:59 |
| *** jcoufal has quit IRC | 23:02 | |
| NobodyCam | lol "bruning in nodes takes to long to do before a deploy.." | 23:09 |
| NobodyCam | nova boot ..... REPLY: your node will be ready in a week please come back then | 23:10 |
| jroll | lol | 23:10 |
| *** ChuckC has joined #openstack-ironic | 23:15 | |
| *** russell_h has quit IRC | 23:19 | |
| *** russell_h has joined #openstack-ironic | 23:19 | |
| *** russell_h has quit IRC | 23:20 | |
| *** russell_h has joined #openstack-ironic | 23:20 | |
| *** PaulCzar has joined #openstack-ironic | 23:31 | |
| *** anderbubble has quit IRC | 23:35 | |
| *** Marga_ has quit IRC | 23:36 | |
| *** Marga_ has joined #openstack-ironic | 23:37 | |
| *** Marga_ has quit IRC | 23:39 | |
| *** Marga_ has joined #openstack-ironic | 23:39 | |
| NobodyCam | JoshNang: left some comments on the clean/zapping specs | 23:43 |
| JoshNang | NobodyCam: thanks! | 23:44 |
| NobodyCam | lol don't thank me until you read them | 23:45 |
| NobodyCam | :-p | 23:45 |
| NobodyCam | JoshNang: I think your going to need three decarators: clean_step(cln_pri) / zap_step(zap_pri) / clean_and_zap_step(cln_pri, zap_pri) | 23:48 |
| *** shakamunyi has quit IRC | 23:49 | |
| *** andreykurilin_ has quit IRC | 23:49 | |
| *** andreykurilin__ has joined #openstack-ironic | 23:49 | |
| *** naohirot has joined #openstack-ironic | 23:52 | |
| *** naohirot has quit IRC | 23:52 | |
| *** naohirot has joined #openstack-ironic | 23:53 | |
| *** rlr0329 has left #openstack-ironic | 23:53 | |
| naohirot | good morning ironic | 23:53 |
| NobodyCam | good morning naohirot | 23:54 |
| naohirot | NobodyCam: Hi, thanks for the review again, I'll submit new patch set today for both specs. | 23:56 |
| NobodyCam | :) | 23:56 |
| * NobodyCam has done so many reviews he getting them mixed up | 23:56 | |
| naohirot | NobodyCam: :-) | 23:56 |
| SpamapS | is there any reason the ramdisk/kernel ID parameters can't have name equivalents that Ironic would try to look up in glance? | 23:56 |
| NobodyCam | because we like uuids | 23:57 |
| *** igordcard has quit IRC | 23:57 | |
| NobodyCam | we would need a look up function | 23:57 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
