| *** bradjones has joined #openstack-ironic | 00:01 | |
| *** Marga_ has quit IRC | 00:03 | |
| *** Marga_ has joined #openstack-ironic | 00:03 | |
| NobodyCam | devananda: question about that log.. worth adding the attempts left to the log.error line, or is that really useless info | 00:06 |
|---|---|---|
| devananda | IMO, useless | 00:08 |
| *** Marga_ has quit IRC | 00:08 | |
| NobodyCam | I agree | 00:08 |
| NobodyCam | removing | 00:08 |
| devananda | I would leave a single WARN line in there | 00:08 |
| NobodyCam | not error?? | 00:08 |
| devananda | especially for debugging purposes, it's helpful to know "that node didn't respond, but i'll try again" | 00:08 |
| devananda | yea. the system hasn't failed, so it's not an ERROR | 00:08 |
| devananda | but that's my opinion | 00:08 |
| NobodyCam | ahh I can see that | 00:08 |
| openstackgerrit | Michael Davies proposed openstack/ironic: Add logical name support to Ironic https://review.openstack.org/141737 | 00:09 |
| devananda | if it fails the final try and gets kicked to maintenance mode -- that's an ERROR | 00:09 |
| *** romcheg has quit IRC | 00:11 | |
| *** Masahiro has joined #openstack-ironic | 00:12 | |
| *** Marga_ has joined #openstack-ironic | 00:12 | |
| openstackgerrit | Chris Krelle proposed openstack/ironic: clean up manager logging https://review.openstack.org/141949 | 00:15 |
| NobodyCam | devananda: ^^^ something like that :-p | 00:16 |
| *** Masahiro has quit IRC | 00:17 | |
| devananda | yes | 00:17 |
| NobodyCam | :) | 00:17 |
| *** jerryz has joined #openstack-ironic | 00:33 | |
| *** david-lyle is now known as david-lyle_afk | 00:35 | |
| jroll | TIL `%(variable) f` == `%(variable)f` | 00:37 |
| NobodyCam | jroll: really | 00:41 |
| jroll | yeah | 00:41 |
| openstackgerrit | Josh Gachnang proposed openstack/ironic-python-agent: Fix missing string format symbol https://review.openstack.org/141954 | 00:41 |
| jroll | NobodyCam: ^^ | 00:41 |
| jroll | NobodyCam: https://gist.github.com/jimrollenhagen/b12b0f25fc4d4357e916 | 00:42 |
| jroll | yet it somehow didn't crash | 00:43 |
| NobodyCam | oh my so it made it a float | 00:47 |
| jroll | yeah | 00:47 |
| NobodyCam | thats messed up | 00:47 |
| jroll | https://github.com/openstack/oslo-incubator/blob/master/openstack/common/apiclient/exceptions.py#L25 | 00:48 |
| jroll | we still use this | 00:48 |
| jroll | in our client | 00:48 |
| *** Masahiro has joined #openstack-ironic | 00:51 | |
| JoshNang | Interesting that the tests exercise that code path, but still pass | 00:53 |
| JoshNang | (for the odd logging bug) | 00:54 |
| jroll | oh? | 00:54 |
| jroll | hmm | 00:54 |
| jroll | I guess it doesn't crash here either | 00:54 |
| *** alexpilotti has joined #openstack-ironic | 00:58 | |
| NobodyCam | did you see dougs comments on the refferenced etherpad: in case we can just abandon the code later. - dhellmann | 00:59 |
| jroll | yeah | 00:59 |
| jroll | we're supposed to move everything to openstacksdk | 00:59 |
| NobodyCam | anyone know the time frame we have to do this? | 01:00 |
| NobodyCam | devananda: ^^^^????? | 01:00 |
| jroll | NobodyCam: I don't think there's any mandate | 01:01 |
| NobodyCam | (yet) lol | 01:02 |
| jroll | :) | 01:02 |
| * NobodyCam steps afk ... Night all | 01:03 | |
| clif_h | dhellman: https://bugs.launchpad.net/python-ironicclient/+bug/1402840 this bug seems to originate from openstack.common.apiclient.exceptions | 01:04 |
| devananda | eh? | 01:05 |
| NobodyCam | lol | 01:05 |
| *** ChuckC has quit IRC | 01:05 | |
| devananda | unified client? nope. | 01:05 |
| jroll | nope as in "screw that" or "nope, I don't know the timeframe" | 01:05 |
| devananda | I mean, it's a good thing in itself, but its existence does not mean we are abandoning python-ironicclient | 01:05 |
| * jroll nods | 01:06 | |
| devananda | there are also several different efforts on a unified client going on | 01:06 |
| clif_h | It seems to assume that a response object that has a content type of 'text/*' necessarily has a 'text' attribute | 01:06 |
| clif_h | but that doesn't seem to always be the case | 01:06 |
| devananda | none of which would obviate the need for a reference implementation of our client library, IMO | 01:07 |
| clif_h | should I work around this in ironicclient or should I file a bug against oslo and fix it there? | 01:07 |
| devananda | clif_h: both | 01:07 |
| devananda | clif_h: if the code exists in oslo, it should be fixed there, and then that fix should be ported to ironic. so that needs a bug in both projects to track the work | 01:08 |
| devananda | clif_h: IFF it's a critical bug, it should be worked around locally in our client in the meantime | 01:09 |
| clif_h | devananda: I don't if it's critical | 01:09 |
| clif_h | I don't know if it's actually a bug in oslo | 01:09 |
| clif_h | although it is definitely a bug if ironicclient recieves a 401 unauthorized | 01:09 |
| clif_h | and doesn't raise an Unauthorized exception | 01:10 |
| devananda | right | 01:10 |
| devananda | that's a bug | 01:10 |
| clif_h | right now, for me, it is raising an attribute error exception | 01:10 |
| devananda | ironicclient should gracefully handle a keystone unauth'd error | 01:10 |
| clif_h | from a line in openstack.common.apiclient.exceptions | 01:10 |
| clif_h | I will file a bug in oslo then | 01:10 |
| clif_h | and point dhellmann to it | 01:10 |
| clif_h | and fix my bug report in ironicclient | 01:11 |
| devananda | clif_h: you can have one bug affect >1 project | 01:11 |
| clif_h | madness | 01:11 |
| clif_h | well affecting yes | 01:11 |
| devananda | clif_h: i'm not sure that the current description is good for that ... but if it is confirmed that the bug is in oslo, we'll want that bug to affect both projects so we can track when it's fixed in both | 01:11 |
| clif_h | ok | 01:12 |
| devananda | anyway, I need to run. heading to dinner and a local linux meetup | 01:12 |
| devananda | thanks for chasing it down :) | 01:12 |
| clif_h | devananda: thanks | 01:12 |
| clif_h | np | 01:12 |
| *** lucas-dinner has quit IRC | 01:12 | |
| *** Marga_ has quit IRC | 01:22 | |
| *** Marga_ has joined #openstack-ironic | 01:23 | |
| *** chenglch has joined #openstack-ironic | 01:26 | |
| *** Marga_ has quit IRC | 01:27 | |
| *** ryanpetrello has quit IRC | 01:27 | |
| openstackgerrit | Josh Gachnang proposed openstack/ironic-python-agent: Add standalone mode for IPA https://review.openstack.org/141957 | 01:30 |
| JoshNang | jroll: JayF ^ | 01:30 |
| JayF | I think we should get +2 from non-rackers on that one before it goes in | 01:30 |
| JayF | just make sure folks who aren't us are OK with it existing :) | 01:30 |
| JoshNang | agreed :) | 01:31 |
| *** ryanpetrello has joined #openstack-ironic | 01:32 | |
| *** Marga_ has joined #openstack-ironic | 01:41 | |
| openstackgerrit | Josh Gachnang proposed openstack/ironic-python-agent: Add standalone mode for IPA https://review.openstack.org/141957 | 01:41 |
| *** ChuckC has joined #openstack-ironic | 01:42 | |
| JayF | devananda: ^ I'd particularly like your thought on that change (not the technology part of the change ... but the what we're doing) | 01:44 |
| * JayF gone | 01:44 | |
| *** killer_prince is now known as lazy_prince | 01:45 | |
| *** anderbubble has joined #openstack-ironic | 01:46 | |
| *** alexpilotti has quit IRC | 01:47 | |
| *** BadCub02 has quit IRC | 01:51 | |
| *** marcoemorais has joined #openstack-ironic | 01:53 | |
| *** marcoemorais2 has joined #openstack-ironic | 01:55 | |
| *** marcoemorais1 has quit IRC | 01:56 | |
| *** marcoemorais1 has joined #openstack-ironic | 01:57 | |
| *** marcoemorais has quit IRC | 01:58 | |
| *** marcoemorais3 has joined #openstack-ironic | 01:58 | |
| *** marcoemorais1 has quit IRC | 01:58 | |
| *** marcoemorais2 has quit IRC | 02:00 | |
| *** marcoemorais3 has quit IRC | 02:02 | |
| *** marcoemorais has joined #openstack-ironic | 02:04 | |
| *** marcoemorais has quit IRC | 02:07 | |
| *** penick_ has joined #openstack-ironic | 02:10 | |
| *** penick has quit IRC | 02:10 | |
| *** penick_ is now known as penick | 02:10 | |
| harlowja | yo yo, https://review.openstack.org/#/c/141961/ (WIP, feel free to say, not yet, ha) | 02:14 |
| *** wanghua has joined #openstack-ironic | 02:15 | |
| *** Marga_ has quit IRC | 02:16 | |
| *** Marga_ has joined #openstack-ironic | 02:17 | |
| *** ryanpetrello has quit IRC | 02:19 | |
| *** Marga_ has quit IRC | 02:21 | |
| *** chen12 has joined #openstack-ironic | 02:30 | |
| *** Marga_ has joined #openstack-ironic | 02:33 | |
| *** penick has quit IRC | 02:35 | |
| *** rloo has quit IRC | 02:36 | |
| *** ramineni has joined #openstack-ironic | 02:43 | |
| *** eghobo has quit IRC | 03:01 | |
| *** Masahiro has quit IRC | 03:02 | |
| *** Masahiro has joined #openstack-ironic | 03:02 | |
| *** rwsu has quit IRC | 03:09 | |
| openstackgerrit | Gopi Krishna S proposed openstack/ironic-specs: Cisco Driver, Update after review changes https://review.openstack.org/139517 | 03:15 |
| *** Marga_ has quit IRC | 03:19 | |
| *** anderbubble has quit IRC | 03:20 | |
| *** Marga_ has joined #openstack-ironic | 03:25 | |
| *** anderbubble has joined #openstack-ironic | 03:31 | |
| *** anderbubble has quit IRC | 03:31 | |
| *** anderbubble has joined #openstack-ironic | 03:34 | |
| *** lazy_prince is now known as killer_prince | 03:34 | |
| *** bauzas has quit IRC | 03:38 | |
| *** Marga_ has quit IRC | 03:41 | |
| *** Nisha has joined #openstack-ironic | 03:44 | |
| *** masco has joined #openstack-ironic | 03:51 | |
| *** Marga_ has joined #openstack-ironic | 03:52 | |
| *** bauzas has joined #openstack-ironic | 03:52 | |
| *** Marga_ has quit IRC | 03:55 | |
| *** kes has joined #openstack-ironic | 03:58 | |
| *** bauzas has quit IRC | 04:00 | |
| *** bauzas has joined #openstack-ironic | 04:03 | |
| *** naohirot has joined #openstack-ironic | 04:15 | |
| naohirot | good afternoon ironic | 04:15 |
| *** Haomeng|2 has quit IRC | 04:16 | |
| *** Haomeng|2 has joined #openstack-ironic | 04:16 | |
| *** anderbubble has quit IRC | 04:22 | |
| *** aweeks has quit IRC | 04:34 | |
| *** Isotopp has quit IRC | 04:36 | |
| *** Isotopp has joined #openstack-ironic | 04:38 | |
| *** aweeks has joined #openstack-ironic | 04:39 | |
| *** killer_prince has quit IRC | 04:41 | |
| *** kes has quit IRC | 05:00 | |
| *** achanda has joined #openstack-ironic | 05:02 | |
| *** pensu has joined #openstack-ironic | 05:10 | |
| *** dlaube has quit IRC | 05:12 | |
| openstackgerrit | Merged openstack/ironic-python-agent: Fix missing string format symbol https://review.openstack.org/141954 | 05:16 |
| *** marcoemorais has joined #openstack-ironic | 05:41 | |
| *** harlowja is now known as harlowja_away | 05:42 | |
| *** marcoemorais1 has joined #openstack-ironic | 05:43 | |
| openstackgerrit | Zhongyue Luo proposed openstack/ironic-specs: Support DHCP service MAC address registration https://review.openstack.org/132511 | 05:45 |
| *** marcoemorais has quit IRC | 05:46 | |
| *** dlpartain has joined #openstack-ironic | 05:49 | |
| *** rameshg87 has joined #openstack-ironic | 05:51 | |
| *** achanda has quit IRC | 05:53 | |
| *** chenglch has quit IRC | 05:55 | |
| *** chenglch has joined #openstack-ironic | 05:55 | |
| openstackgerrit | Michael Davies proposed openstack/ironic: Add logical name support to Ironic https://review.openstack.org/141737 | 06:00 |
| rlr0329 | I am trying to pxe boot a baremetal node using the pxe_ipmi driver in a devstack installation. The baremetal node successfully gets an IP address but then the following message is displayed: TFTP Prefix: Unable to locate configuration file: | 06:03 |
| *** achanda has joined #openstack-ironic | 06:03 | |
| rlr0329 | The pxelinux.0 file exists in the tftp_root directory but the pxelinux.cfg directory is empty. Is this correct? What should the file ownership/permissions be? Other ideas? | 06:03 |
| *** bradjones has quit IRC | 06:07 | |
| *** Marga_ has joined #openstack-ironic | 06:08 | |
| *** bradjones has joined #openstack-ironic | 06:10 | |
| *** Marga_ has quit IRC | 06:13 | |
| *** Marga_ has joined #openstack-ironic | 06:27 | |
| *** anderbubble has joined #openstack-ironic | 06:31 | |
| *** anderbubble has quit IRC | 06:32 | |
| *** Marga__ has joined #openstack-ironic | 06:33 | |
| *** bradjones has quit IRC | 06:35 | |
| *** Marga__ has quit IRC | 06:35 | |
| *** anderbubble has joined #openstack-ironic | 06:37 | |
| *** Marga_ has quit IRC | 06:38 | |
| *** teju has joined #openstack-ironic | 06:39 | |
| *** Marga_ has joined #openstack-ironic | 06:39 | |
| *** killer_prince has joined #openstack-ironic | 06:41 | |
| *** killer_prince has quit IRC | 06:41 | |
| *** lazy_prince has joined #openstack-ironic | 06:42 | |
| *** lazy_prince has quit IRC | 06:42 | |
| *** lazy_prince has joined #openstack-ironic | 06:43 | |
| *** Marga_ has quit IRC | 06:50 | |
| *** Marga_ has joined #openstack-ironic | 06:50 | |
| *** dlpartain has quit IRC | 07:02 | |
| *** anderbubble has quit IRC | 07:02 | |
| *** mrda is now known as mrda-away | 07:05 | |
| *** Marga_ has quit IRC | 07:07 | |
| *** anderbubble has joined #openstack-ironic | 07:09 | |
| *** achanda has quit IRC | 07:10 | |
| *** achanda has joined #openstack-ironic | 07:11 | |
| *** Marga_ has joined #openstack-ironic | 07:12 | |
| *** achanda has quit IRC | 07:16 | |
| *** pcaruana has joined #openstack-ironic | 07:18 | |
| *** Marga_ has quit IRC | 07:23 | |
| *** achanda has joined #openstack-ironic | 07:34 | |
| *** penick has joined #openstack-ironic | 07:40 | |
| Haomeng|2 | rlr0329: hi | 07:44 |
| Haomeng|2 | rlr0329: the pxelinux.cfg directory will be populated by ironic | 07:45 |
| Haomeng|2 | rlr0329: did you enable tftp server? | 07:45 |
| Haomeng|2 | rlr0329: check tftp port, if it is linstening | 07:46 |
| Haomeng|2 | rlr0329: and check the tftp folder permision | 07:57 |
| *** ifarkas has joined #openstack-ironic | 08:01 | |
| cuihaozhi | r|r0329: i resolved same problem by edit /etc/xinet.d/tftp with option -m /tftp/map-file | 08:04 |
| Haomeng|2 | cuihaozhi: cool | 08:04 |
| *** erwan_taf has joined #openstack-ironic | 08:07 | |
| *** achanda has quit IRC | 08:08 | |
| *** achanda has joined #openstack-ironic | 08:15 | |
| *** anderbubble has quit IRC | 08:16 | |
| *** dlpartain has joined #openstack-ironic | 08:16 | |
| openstackgerrit | jiangfei proposed openstack/python-ironicclient: ironicclient handle faultstring when using SessionClient https://review.openstack.org/142021 | 08:18 |
| *** chenglch has quit IRC | 08:32 | |
| *** rakesh_hs has joined #openstack-ironic | 08:36 | |
| *** anderbubble has joined #openstack-ironic | 08:38 | |
| *** Krast has joined #openstack-ironic | 08:42 | |
| *** k4n0 has joined #openstack-ironic | 08:47 | |
| *** jcoufal has joined #openstack-ironic | 08:51 | |
| *** Nisha_away has joined #openstack-ironic | 08:54 | |
| *** Nisha has quit IRC | 08:54 | |
| *** yarker has quit IRC | 08:58 | |
| *** dlpartain has left #openstack-ironic | 09:08 | |
| *** ZhangYang has joined #openstack-ironic | 09:09 | |
| *** derekh has joined #openstack-ironic | 09:10 | |
| *** romcheg has joined #openstack-ironic | 09:10 | |
| *** marcoemorais1 has quit IRC | 09:14 | |
| *** yarker has joined #openstack-ironic | 09:14 | |
| *** masco has quit IRC | 09:18 | |
| *** anderbubble has quit IRC | 09:19 | |
| *** penick has quit IRC | 09:31 | |
| *** achanda has quit IRC | 09:32 | |
| *** achanda has joined #openstack-ironic | 09:32 | |
| *** dtantsur|afk is now known as dtantsur | 09:33 | |
| dtantsur | Morning! | 09:33 |
| openstackgerrit | Merged openstack/ironic: Add base state machine https://review.openstack.org/139215 | 09:34 |
| openstackgerrit | Merged openstack/ironic: Begin using the state machine for node deploy/teardown https://review.openstack.org/139216 | 09:34 |
| vdrok | morning ironic | 09:36 |
| vdrok | morning dtantsur | 09:36 |
| *** achanda has quit IRC | 09:37 | |
| dtantsur | vdrok, morning | 09:37 |
| dtantsur | NobodyCam, devananda, sorry, was feeling really bad yesterday afternoon :( +1 for bug review on Thu | 09:38 |
| ekarlso- | aare you guys gonan use for statemachine thing that was made outside taskflow btw? | 09:38 |
| *** MattMan has quit IRC | 09:41 | |
| *** MattMan has joined #openstack-ironic | 09:41 | |
| *** achanda has joined #openstack-ironic | 09:41 | |
| dtantsur | ekarlso-, some code was copy-pasted IIRC | 09:41 |
| ekarlso- | ah ok | 09:42 |
| *** achanda has quit IRC | 09:44 | |
| *** achanda has joined #openstack-ironic | 09:44 | |
| openstackgerrit | Yuriy Zveryanskyy proposed openstack/ironic: Improve Agent deploy driver validation https://review.openstack.org/141766 | 09:45 |
| yuriyz | morning Ironic | 09:47 |
| *** Nisha_away is now known as Nisha | 09:48 | |
| dtantsur | yuriyz, morning | 09:48 |
| Nisha | morning dtantsur | 09:49 |
| dtantsur | Nisha, morning | 09:49 |
| *** achanda has quit IRC | 09:50 | |
| *** achanda has joined #openstack-ironic | 09:50 | |
| yuriyz | dtantsur, in the meeting kozhukalov asked about Fuel Agent driver, can you review new version of the spec if you have some time? | 09:52 |
| dtantsur | on my radar | 09:52 |
| yuriyz | dtantsur, thanks | 09:52 |
| openstackgerrit | Imre Farkas proposed openstack/ironic-specs: New driver interface for RAID configuration https://review.openstack.org/135899 | 09:53 |
| ifarkas | rameshg87, hi, ^ | 09:54 |
| rameshg87 | ifarkas, hi | 09:54 |
| *** erwan_taf has quit IRC | 09:54 | |
| ifarkas | rameshg87, I fixed the use cases in the spec as those were incorrect | 09:54 |
| ifarkas | rameshg87, I will also comment on a few stuff that need to be discussed | 09:54 |
| rameshg87 | ifarkas, great | 09:55 |
| rameshg87 | ifarkas, i will have a look | 09:55 |
| ifarkas | rameshg87, thanks | 09:55 |
| rameshg87 | ifarkas, is there something major that you wanted to discuss ? | 09:56 |
| rameshg87 | ifarkas, if so, we can discuss in irc, otherwise we can use gerrit | 09:56 |
| Nisha | dtantsur, discovery spec too.... :) | 09:58 |
| dtantsur | yeah sure :) | 09:58 |
| dtantsur | lemme fix up my stuff and hand it to ifarkas for review, if should not take soon :) | 09:58 |
| ifarkas | rameshg87, I need to investigate a few things first on DRAC side | 09:59 |
| rameshg87 | ifarkas, okay | 10:00 |
| ifarkas | rameshg87, I will ping you if you are still online | 10:00 |
| rameshg87 | ifarkas, okay | 10:00 |
| dtantsur | Nisha, lgtm | 10:03 |
| *** erwan_taf has joined #openstack-ironic | 10:07 | |
| *** ekarlso- has quit IRC | 10:07 | |
| *** ekarlso- has joined #openstack-ironic | 10:07 | |
| openstackgerrit | Dmitry Tantsur proposed stackforge/ironic-discoverd: Simple functional testing for discoverd and the reference ramdisk https://review.openstack.org/141777 | 10:07 |
| openstackgerrit | Dmitry Tantsur proposed stackforge/ironic-discoverd: Add forgotten modules to setup.py (shame on me) https://review.openstack.org/142052 | 10:07 |
| dtantsur | ifarkas, could you review these ^^^? | 10:08 |
| ifarkas | dtantsur, yep, I will | 10:09 |
| *** jerryz has quit IRC | 10:14 | |
| *** jerryz has joined #openstack-ironic | 10:14 | |
| openstackgerrit | Imre Farkas proposed openstack/ironic-specs: New driver interface for RAID configuration https://review.openstack.org/135899 | 10:15 |
| *** pelix has joined #openstack-ironic | 10:20 | |
| *** ZZhangYang has joined #openstack-ironic | 10:22 | |
| *** ZhangYang has quit IRC | 10:24 | |
| openstackgerrit | Merged openstack/ironic: Add new enrollment and troubleshooting doc sections https://review.openstack.org/136202 | 10:27 |
| GheRivero | morning all | 10:32 |
| *** Nisha has quit IRC | 10:33 | |
| *** achanda has quit IRC | 10:40 | |
| *** athomas has quit IRC | 10:40 | |
| dtantsur | GheRivero, morning | 10:43 |
| *** Nisha has joined #openstack-ironic | 10:43 | |
| *** chenglch|2 has joined #openstack-ironic | 10:47 | |
| *** Nisha has quit IRC | 10:49 | |
| *** ramineni has quit IRC | 10:51 | |
| *** athomas has joined #openstack-ironic | 10:54 | |
| *** chenglch|2 has quit IRC | 10:54 | |
| sambetts | morning GheRivero, dtantsur | 10:56 |
| dtantsur | sambetts, morning | 10:57 |
| *** lucasagomes has joined #openstack-ironic | 10:58 | |
| GheRivero | romcheg: ping | 10:59 |
| *** wanghua has quit IRC | 10:59 | |
| *** wanghua has joined #openstack-ironic | 11:00 | |
| ifarkas | dtantsur, quick question re: #141777 | 11:07 |
| ifarkas | dtantsur, why you needed to create net/em1 and net/em2? | 11:07 |
| dtantsur | ifarkas, the ramdisk enumerates dirs in /sys/class/net to learn network devices | 11:08 |
| dtantsur | ifarkas, I simulate this process by using these files | 11:08 |
| dtantsur | is it what you asked? | 11:08 |
| ifarkas | dtantsur, yeah, that's what I asked | 11:09 |
| ifarkas | dtantsur, just trying to figure out how this simulation works | 11:09 |
| ifarkas | dtantsur, but really useful stuff ;-) | 11:09 |
| openstackgerrit | Merged stackforge/ironic-discoverd: Add forgotten modules to setup.py (shame on me) https://review.openstack.org/142052 | 11:09 |
| dtantsur | ifarkas, actually, there's no point in having them in git though... once you finish reviewing, I'll make them generated in rumtime | 11:09 |
| dtantsur | ifarkas, also merged ^^^ it was not interesting | 11:10 |
| ifarkas | dtantsur, alright | 11:10 |
| dtantsur | ifarkas, better have a look at https://review.openstack.org/#/c/141400/ | 11:10 |
| ifarkas | dtantsur, OK | 11:10 |
| openstackgerrit | Dmitry Tantsur proposed stackforge/ironic-discoverd: Simple functional testing for discoverd and the reference ramdisk https://review.openstack.org/141777 | 11:14 |
| dtantsur | dropped /net ^^^ will just approve now | 11:15 |
| *** bradjones has joined #openstack-ironic | 11:15 | |
| romcheg | GheRivero: pong | 11:18 |
| GheRivero | hi romcheg. I was looking to your oslo.policy patch. Are you planning to work on it again/soon? | 11:19 |
| romcheg | GheRivero: yup, I will update it today, had to be off the things for some time | 11:19 |
| GheRivero | ok. thanks! | 11:20 |
| *** Masahiro has quit IRC | 11:23 | |
| *** Masahiro has joined #openstack-ironic | 11:39 | |
| ifarkas | dtantsur, 141400 also acked. a quick question: will you move ironic-discoverd element under diskimage-buidler/elements? | 11:41 |
| dtantsur | ifarkas, you mean not the ramdisk, but element for discoverd service installation? | 11:45 |
| dtantsur | because the ramdisk is already there | 11:45 |
| ifarkas | dtantsur, yes, I meant the service | 11:45 |
| dtantsur | ifarkas, hmm... don't know. first of all it should go to tripleo-image-elements, not DIB, right? | 11:46 |
| dtantsur | then I probably need someone interested in it being there | 11:46 |
| ifarkas | dtantsur, not sure, what's the requirement for each of the element repos... | 11:46 |
| dtantsur | for now I only got requests for the ramdisk | 11:46 |
| *** rakesh_hs has quit IRC | 12:02 | |
| *** teju has left #openstack-ironic | 12:04 | |
| *** Masahiro has quit IRC | 12:07 | |
| ifarkas | rameshg87, are you still available? | 12:14 |
| rameshg87 | ifarkas, i am but i might need to leave for home in a few minutes :( | 12:14 |
| rameshg87 | ifarkas, may be we can discuss till then ? have about 10 mins .. | 12:14 |
| ifarkas | rameshg87, no worries, I can also comment on the spec | 12:15 |
| rameshg87 | ifarkas, okay .. | 12:15 |
| rameshg87 | ifarkas, let's discuss on the spec then .. | 12:15 |
| ifarkas | rameshg87, ok | 12:15 |
| *** arif-ali has quit IRC | 12:16 | |
| *** EmilienM is now known as EmilienM|afk | 12:17 | |
| *** arif-ali has joined #openstack-ironic | 12:17 | |
| *** romcheg has quit IRC | 12:18 | |
| *** pcrews has quit IRC | 12:19 | |
| *** dlpartain has joined #openstack-ironic | 12:19 | |
| *** kevinbenton has quit IRC | 12:20 | |
| *** pcrews has joined #openstack-ironic | 12:20 | |
| *** pelix has quit IRC | 12:22 | |
| *** kevinbenton has joined #openstack-ironic | 12:22 | |
| *** pelix has joined #openstack-ironic | 12:22 | |
| *** romcheg has joined #openstack-ironic | 12:23 | |
| openstackgerrit | Merged stackforge/ironic-discoverd: Simple functional testing for discoverd and the reference ramdisk https://review.openstack.org/141777 | 12:23 |
| *** ZhangYang has joined #openstack-ironic | 12:24 | |
| *** rameshg87 has quit IRC | 12:27 | |
| *** dlpartain has quit IRC | 12:29 | |
| *** dlpartain has joined #openstack-ironic | 12:29 | |
| *** ZZhangYang has quit IRC | 12:31 | |
| *** bauzas has quit IRC | 12:31 | |
| *** sambetts has quit IRC | 12:31 | |
| *** kevinbenton has quit IRC | 12:37 | |
| *** kevinbenton has joined #openstack-ironic | 12:40 | |
| *** pensu has quit IRC | 12:43 | |
| *** sambetts has joined #openstack-ironic | 12:49 | |
| *** bauzas has joined #openstack-ironic | 12:49 | |
| *** dprince has joined #openstack-ironic | 12:57 | |
| *** alexpilotti has joined #openstack-ironic | 13:02 | |
| *** alexpilotti_ has joined #openstack-ironic | 13:06 | |
| *** sambetts_ has joined #openstack-ironic | 13:07 | |
| *** alexpilotti has quit IRC | 13:07 | |
| *** alexpilotti_ is now known as alexpilotti | 13:07 | |
| *** sambetts has quit IRC | 13:07 | |
| *** sambetts_ is now known as sambetts | 13:07 | |
| *** Masahiro has joined #openstack-ironic | 13:07 | |
| *** bauzas has quit IRC | 13:08 | |
| *** bauzas has joined #openstack-ironic | 13:12 | |
| *** Masahiro has quit IRC | 13:12 | |
| openstackgerrit | Dmitry Tantsur proposed stackforge/ironic-discoverd: Update README and stop posting full changelogs https://review.openstack.org/142097 | 13:14 |
| dtantsur | ifarkas, ^^^ forward port of the same patch from stable into master, have a look if you have time | 13:16 |
| openstackgerrit | Yuriy Zveryanskyy proposed openstack/ironic-specs: Add a new driver for Fuel Agent https://review.openstack.org/138115 | 13:17 |
| *** pensu has joined #openstack-ironic | 13:17 | |
| ifarkas | dtantsur, lgtm! | 13:22 |
| dtantsur | cool | 13:23 |
| openstackgerrit | Dmitry Tantsur proposed openstack/ironic-specs: In-band hardware properties inspection via ironic-discoverd https://review.openstack.org/135605 | 13:26 |
| dtantsur | lucasagomes, mind also having a look at ^^^ ? | 13:27 |
| lucasagomes | dtantsur, nop 1 sec | 13:27 |
| *** EmilienM|afk is now known as EmilienM | 13:28 | |
| *** lifeless_ has joined #openstack-ironic | 13:33 | |
| *** lifeless has quit IRC | 13:35 | |
| *** dtantsur is now known as dtantsur|brb | 13:35 | |
| *** pensu has quit IRC | 13:43 | |
| openstackgerrit | AKAHIRA Junya proposed openstack/ironic: Update command options in the Installation Guide https://review.openstack.org/142107 | 13:43 |
| *** rlr0329 has quit IRC | 13:43 | |
| *** pensu has joined #openstack-ironic | 13:46 | |
| *** pensu has quit IRC | 13:53 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: Extend API multivalue fields https://review.openstack.org/137762 | 13:58 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: PXE: Pass root device hints via kernel cmdline https://review.openstack.org/142109 | 13:58 |
| *** dlpartain has left #openstack-ironic | 14:01 | |
| *** bradjones has quit IRC | 14:01 | |
| *** bradjones has joined #openstack-ironic | 14:05 | |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: PXE: Pass root device hints via kernel cmdline https://review.openstack.org/142109 | 14:05 |
| openstackgerrit | AKAHIRA Junya proposed openstack/ironic: Update command options in the Installation Guide https://review.openstack.org/142107 | 14:07 |
| openstackgerrit | Naohiro Tamura proposed openstack/python-ironicclient: Removed http proxy environment variable so that httpretty can work. https://review.openstack.org/142112 | 14:14 |
| naohirot | dtantsur|brb: Hi, I submitted mistakenly a ironicclient bug into ironic bug db, thank you for forwarding it into the ironicclient bug db. | 14:18 |
| *** rloo has joined #openstack-ironic | 14:18 | |
| naohirot | Shrews: Hi, I finally figured out why ironicclient tox unit test fails in my environment. | 14:20 |
| *** ZZhangYang has joined #openstack-ironic | 14:20 | |
| Shrews | naohirot: i forget the conversation, but great! :) | 14:20 |
| naohirot | Shrews: somehow httpretty didn't work in my environment, you showed me that ironicclient tox tests worked by the gate log. | 14:22 |
| naohirot | Shrews: thanks. If you have a time, please review the https://review.openstack.org/142112 | 14:23 |
| *** pelix1 has joined #openstack-ironic | 14:23 | |
| *** pelix has quit IRC | 14:23 | |
| *** trown|call has joined #openstack-ironic | 14:25 | |
| *** jeblair_ has joined #openstack-ironic | 14:25 | |
| naohirot | good night ironic | 14:28 |
| *** HenryG_ has joined #openstack-ironic | 14:28 | |
| *** naohirot has quit IRC | 14:28 | |
| *** ZhangYang has quit IRC | 14:30 | |
| *** k4n0 has quit IRC | 14:30 | |
| *** trown has quit IRC | 14:30 | |
| *** HenryG has quit IRC | 14:30 | |
| *** jeblair has quit IRC | 14:30 | |
| *** jerryz has quit IRC | 14:31 | |
| *** k4n0 has joined #openstack-ironic | 14:31 | |
| openstackgerrit | Merged stackforge/ironic-discoverd: Update README and stop posting full changelogs https://review.openstack.org/142097 | 14:33 |
| *** Masahiro has joined #openstack-ironic | 14:40 | |
| *** lazy_prince is now known as killer_prince | 14:40 | |
| *** jerryz has joined #openstack-ironic | 14:44 | |
| *** Masahiro has quit IRC | 14:44 | |
| *** ndipanov has quit IRC | 14:49 | |
| *** ndipanov has joined #openstack-ironic | 14:49 | |
| *** dtantsur|brb is now known as dtantsur | 14:50 | |
| *** athomas has quit IRC | 14:51 | |
| *** jerryz has quit IRC | 14:56 | |
| *** jerryz has joined #openstack-ironic | 14:57 | |
| *** ZZhangYang has quit IRC | 14:57 | |
| *** ZZhangYang has joined #openstack-ironic | 14:58 | |
| *** athomas has joined #openstack-ironic | 14:59 | |
| *** dprince has quit IRC | 15:00 | |
| *** dprince has joined #openstack-ironic | 15:01 | |
| NobodyCam | good morning Ironic | 15:12 |
| dtantsur | NobodyCam, morning | 15:18 |
| NobodyCam | morning dtantsur | 15:18 |
| *** viktors|afk has joined #openstack-ironic | 15:18 | |
| *** jcoufal_ has joined #openstack-ironic | 15:19 | |
| *** jcoufal has quit IRC | 15:21 | |
| BadCub_ | morning Ironic | 15:21 |
| jroll | morning everybody :) | 15:21 |
| NobodyCam | morning BadCub_ | 15:21 |
| NobodyCam | morning jroll | 15:22 |
| jroll | who was it that had problems with the deleted flavor? | 15:22 |
| NobodyCam | :) | 15:22 |
| jroll | this is a thing https://review.openstack.org/#/c/142123/ | 15:22 |
| dtantsur | NobodyCam, jroll: sync power states strikes back! any ideas if it is the correct thing to happen: http://paste.openstack.org/show/151894/ ? | 15:23 |
| rloo | jroll: wasn't it clint byrum? | 15:23 |
| rloo | hi everyone :-) | 15:23 |
| NobodyCam | morning rloo :) | 15:23 |
| jroll | dtantsur: lol, wat | 15:24 |
| jroll | SpamapS: https://review.openstack.org/#/c/142123/ | 15:24 |
| jroll | rloo: thanks and good morning :) | 15:24 |
| jroll | dtantsur: did these have instances on them? | 15:24 |
| jroll | specifically, the ones that turned on? | 15:24 |
| dtantsur | jroll, nope, they were under inspection (and had maintenance mode = True) | 15:24 |
| *** pelix1 has quit IRC | 15:25 | |
| *** pelix1 has joined #openstack-ironic | 15:25 | |
| jroll | dtantsur: fwiw, I think force_state_power_sync True is insane | 15:25 |
| *** trown|call has quit IRC | 15:25 | |
| *** trown|call has joined #openstack-ironic | 15:25 | |
| dtantsur | also morning, jroll, rloo! | 15:25 |
| *** HenryG_ has quit IRC | 15:25 | |
| *** HenryG_ has joined #openstack-ironic | 15:25 | |
| *** k4n0 has quit IRC | 15:25 | |
| *** k4n0 has joined #openstack-ironic | 15:25 | |
| jroll | huh, interesting | 15:25 |
| *** ndipanov has quit IRC | 15:25 | |
| *** ndipanov has joined #openstack-ironic | 15:25 | |
| dtantsur | and think what happened is that maintenance mode was disabled, while power state action were not finished. and our smart power sync feature kicked in | 15:25 |
| jroll | probably | 15:25 |
| jroll | but why would 'power off' take 20 minutes? | 15:26 |
| dtantsur | though I though we use locking exactly for that... | 15:26 |
| jroll | and yeah, should be locked at that point | 15:26 |
| NobodyCam | dtantsur: oh that could bee | 15:26 |
| *** jcoufal_ has quit IRC | 15:26 | |
| *** jcoufal_ has joined #openstack-ironic | 15:26 | |
| TheJulia | BadCub_, NobodyCam: Good morning :) | 15:26 |
| NobodyCam | morning TheJulia | 15:26 |
| NobodyCam | welcome to Ironic | 15:27 |
| dtantsur | morning also to BadCub_, TheJulia :) | 15:27 |
| dtantsur | it's a hot hour for this channel :D | 15:27 |
| NobodyCam | :) | 15:27 |
| * dtantsur lacks the new state machine in Juno... | 15:28 | |
| dtantsur | folks, do you think it's a bug? I mean, leaving maintenance probably should not touch the power state... | 15:29 |
| jroll | I think force_power_state_during_sync is a bug | 15:30 |
| jroll | but that's just me | 15:30 |
| NobodyCam | jroll: d'oh | 15:30 |
| jroll | :) | 15:31 |
| dtantsur | https://bugs.launchpad.net/ironic/+bug/1403106 | 15:36 |
| dtantsur | now we'll try to set force_power_state_during_sync=False and see | 15:36 |
| jroll | dtantsur: ohhhhh | 15:36 |
| jroll | it probably didn't run the sync loop between 4 and 5 | 15:37 |
| jroll | and so didn't pick up that it was supposed to be off | 15:37 |
| jroll | err no | 15:37 |
| jroll | ignore me | 15:37 |
| * jroll goes away to drink coffee | 15:37 | |
| dtantsur | :) | 15:37 |
| jroll | "using api" should set it in the db | 15:38 |
| dtantsur | sweet, something locked nodes in DB, so now we can't retry >_< | 15:39 |
| * dtantsur +2 to having API to break locks... | 15:39 | |
| jroll | :) | 15:39 |
| *** jerryz has quit IRC | 15:39 | |
| *** pcaruana has quit IRC | 15:42 | |
| *** PaulCzar has joined #openstack-ironic | 15:42 | |
| dtantsur | so, setting force_power_state_during_sync=false helped... | 15:51 |
| *** penick has joined #openstack-ironic | 15:53 | |
| *** lucasagomes is now known as lucas-hungry | 15:53 | |
| *** achanda has joined #openstack-ironic | 15:57 | |
| *** jdob has joined #openstack-ironic | 16:02 | |
| jdob | hey all, I'm getting a bunch of errors about nodes being locked when I try to create a heat stack on top of ironic. is there anyone around who can give me a hand debugging this? | 16:04 |
| *** achanda has quit IRC | 16:05 | |
| NobodyCam | jdob: how many nodes? is the power state correct on them before you started the deploy | 16:06 |
| NobodyCam | ? | 16:06 |
| NobodyCam | and good morning :) | 16:07 |
| jdob | funny you ask about the power state, that was previously my problem :) | 16:07 |
| dtantsur | NobodyCam, it's exactly after the process I was talking about above | 16:07 |
| jdob | and morning :) | 16:07 |
| jdob | ok, so here's my situation | 16:07 |
| jdob | oh, good, so you have a rough idea | 16:07 |
| dtantsur | jdob, I breifly described what you encountered | 16:07 |
| jdob | dtantsur got me past that with the conf file change | 16:07 |
| dtantsur | * briefly | 16:07 |
| jdob | and i waited a bit between discovery and heat create | 16:07 |
| jdob | so that all of them reflected powered off | 16:07 |
| jdob | i started the stack create and it was definitely in better shape | 16:08 |
| jdob | the node showed an instance ID and the state was deploying | 16:08 |
| jdob | then it jumped to deploy failed and everything cleaned itself up back to powered off | 16:08 |
| jdob | lemme pastebin the log | 16:08 |
| jdob | http://paste.fedoraproject.org/160248/41874616/ | 16:09 |
| jdob | another update (dtantsur this is new to you too) | 16:10 |
| jdob | nova tanked after the retries | 16:10 |
| jdob | but I'm guessing it tried again, either on its own or because heat triggered it | 16:10 |
| jdob | now one of the nodes is active and another is at deploying | 16:10 |
| jdob | in other words, it looks like the next time it ran it's having some more success | 16:11 |
| jdob | which is both good and disconcerting at the same time :) | 16:11 |
| jdob | i suppose my question is if a. the locked thing isn't as much of a problem as it sounds, since it'll retry (in other words, is that expected/acceptable) | 16:12 |
| jdob | and b. what would cause a node to be locked? | 16:13 |
| NobodyCam | jdob: most likly the power status task | 16:13 |
| jdob | NobodyCam: in the sense that it's in the process of being powered on and ironic tried to do other stuff to it too quickly? | 16:14 |
| NobodyCam | jdob: a breif lock it taken to poll power state | 16:15 |
| NobodyCam | jdob: are these values still at there default values? https://github.com/openstack/ironic/blob/master/etc/ironic/ironic.conf.sample#L560-L564 | 16:15 |
| *** achanda has joined #openstack-ironic | 16:15 | |
| jdob | NobodyCam: i don't have them overridden in ironic.conf, so I'm guessing yet | 16:16 |
| jdob | yes even | 16:16 |
| *** Nisha has joined #openstack-ironic | 16:17 | |
| jdob | this is on the Juno release | 16:17 |
| NobodyCam | how many nodes are you working with, and which driver? | 16:17 |
| jdob | 4 nodes, the pxe_ssh driver; these are VMs, in case I didn't specify that before | 16:18 |
| jdob | (in hindsight, that's probably a pretty relevant distinction) | 16:18 |
| NobodyCam | controling virsh i assume | 16:19 |
| jdob | yes | 16:20 |
| devananda | g'morning, all | 16:21 |
| NobodyCam | good morning devananda :) | 16:21 |
| *** penick has quit IRC | 16:21 | |
| jdob | NobodyCam: ok, it looks like my stack is continuing part the ironic parts, both of which reflect active. i'll see if I can reproduce it, but as a final question can you think of a reason why the first two nodes it tried to create would fail (on locked errors) but then stand up fine on the retry (again, not sure if that's an ironic retry or nova) | 16:26 |
| *** lsmola has quit IRC | 16:26 | |
| rloo | morning devananda, just the person I wanted to talk to. | 16:26 |
| rloo | devananda: wrt the state machine, https://review.openstack.org/#/c/139216/, I just added a comment even though it merged | 16:26 |
| rloo | devananda: I think having NOSTATE being NONE causes some problems with the fsm. | 16:27 |
| *** pelix1 is now known as pelix | 16:27 | |
| NobodyCam | jdob: its kinda seeming like the power status task is polling to agressivly | 16:27 |
| rloo | devananda: wrt the fsm's _target_state being set to None too. You have a NOTE about that manager.py | 16:28 |
| *** ndipanov has quit IRC | 16:28 | |
| jdob | NobodyCam: oh interesting; i misread what you meant earlier. so you think one power state request is hitting into another? | 16:28 |
| *** Marga_ has joined #openstack-ironic | 16:28 | |
| *** Masahiro has joined #openstack-ironic | 16:29 | |
| *** ndipanov has joined #openstack-ironic | 16:29 | |
| jdob | actually, that could make sense. i just tried SSHing from the VM my ironic is running on to the virt host and it's taking a surprisingly long amount of time to resolve | 16:29 |
| NobodyCam | jdob: that would def cause something like what your seeing | 16:30 |
| jdob | more than 1 second, so if that's the retry, i could see why they'd be bombing out | 16:30 |
| NobodyCam | is there anything you can do yo improve performance of hte network | 16:31 |
| jdob | ok, that's a good starting point, trying to figure out wtf is wrong with this box that SSH is being so weird | 16:31 |
| jdob | ya, lemme start there. there's no reason at all SSH should take that long | 16:31 |
| jdob | gotta be something really stupid | 16:31 |
| jdob | i'll let you know if things run better when I've debugged that | 16:31 |
| jdob | thanks NobodyCam and dtantsur :D | 16:31 |
| NobodyCam | else if this is a test env maybe bump up the timeout values | 16:31 |
| devananda | rloo: you are correct that NOSTATE being None causes problems | 16:33 |
| *** Masahiro has quit IRC | 16:34 | |
| devananda | rloo: I think I worked around the only such problem, with an inline comment | 16:34 |
| devananda | *and included an inline comment for the workaround | 16:34 |
| *** EmilienM is now known as EmilienM|afk | 16:34 | |
| devananda | rloo: but as soon as possible, we should be moving away from NOSTATE | 16:34 |
| rloo | devananda: take a look at my comment. The unit test isn't quite right, so it passes when it shouldn't. | 16:34 |
| devananda | rloo: :( | 16:35 |
| rloo | devananda: I think we may need to review the unit tests to make sure the fsm is set up correctly. | 16:35 |
| *** dwalleck has joined #openstack-ironic | 16:35 | |
| *** dlaube has joined #openstack-ironic | 16:36 | |
| *** lucas-hungry is now known as lucasagomes | 16:38 | |
| NobodyCam | morning lucasagomes | 16:38 |
| lucasagomes | NobodyCam, morning :D | 16:38 |
| lucasagomes | devananda, rloo morning | 16:38 |
| rloo | devananda: I haven't thought this through, but if task_manager.process_event is the only thing dealing with the fsm, I was wondering if we could add some internal fsmNOSTATE='nostate' and have the task_manager xlate that to NOSTATE for the node's [target_]provision_state values | 16:39 |
| rloo | hi lucasagomes | 16:39 |
| *** ryanpetrello has joined #openstack-ironic | 16:39 | |
| lucasagomes | rloo, oh that's a good point. Tho if we are adding a value 'nostate' perhaps we should already call it 'available' or something to make the NOSTATE->AVAILABLE transition easier? | 16:40 |
| rloo | lucasagomes: I would prefer to mimic the existing NOSTATE right now, just to get over this bump and not confuse things with 'the future' ;) | 16:41 |
| rloo | lucasagomes: but if it makes sense to do as you say (I haven't thought beyond capturing our existing states with the fsm) | 16:42 |
| NobodyCam | oh /my initial reaction is no.. stick with current states then switch to the new states | 16:42 |
| lucasagomes | rloo, I see, actually I'm out-of-date with the current plan for the state machine | 16:42 |
| lucasagomes | I did review some code today | 16:42 |
| lucasagomes | rloo, but I'm not 100% sure what the strategy is. Seems like we are using the new fsm module to model the current state machine | 16:43 |
| lucasagomes | and once that's finished we are going to add the new states etc | 16:43 |
| lucasagomes | is that correct? | 16:43 |
| rloo | lucasagomes: yes, I think so, based on what was mentioned in yesterday's ironic meeting | 16:43 |
| NobodyCam | thats my understanding of our direction | 16:43 |
| devananda | yes, more or less | 16:43 |
| devananda | there will be at least two "bumps" inthe road | 16:43 |
| devananda | - changing NOSTATE==None to AVAILABLE | 16:44 |
| *** david-lyle_afk is now known as david-lyle | 16:44 | |
| devananda | - adding callbacks / hooks for the new states which no-op today | 16:44 |
| lucasagomes | right | 16:44 |
| devananda | the first one is going to require a db migration and some code in the API layer | 16:44 |
| devananda | so that we can accept the inputs from an older client (ie, a non-upgraded Nova driver) | 16:45 |
| rloo | but the current NOSTATE may not always == AVAILABLE in the new world? | 16:45 |
| lucasagomes | but if the current NOSTATE=None is problematic, maybe we should start tackling the AVAILABLE part as part of this effort already? | 16:45 |
| devananda | first I think we should get all the code using the FSM, so nothing is directly returning states / changign the state in the db | 16:46 |
| devananda | my patches are close, but I haven't gone through everything yet | 16:46 |
| devananda | (too many other things this week I couldn't postpone) | 16:46 |
| lucasagomes | I see, yeah it's pretty close indeed | 16:46 |
| lucasagomes | mostly now seems to be to use the fsm on the drivers itself | 16:46 |
| devananda | then yes, we can address AVAILABLE | 16:46 |
| lucasagomes | like pass_deploy_info that sets it to ACTIVE from DEPLOYWAIT | 16:46 |
| devananda | rloo: you're correct. NOSTATE today means AVAILABLE only if it is validated | 16:47 |
| lucasagomes | devananda, off-topic, left a comment on https://review.openstack.org/#/c/137762/ | 16:47 |
| NobodyCam | brb | 16:48 |
| *** Marga_ has quit IRC | 16:49 | |
| *** Marga_ has joined #openstack-ironic | 16:49 | |
| *** pensu has joined #openstack-ironic | 16:53 | |
| *** alexpilotti has quit IRC | 16:55 | |
| dtantsur | calling it a day for today, see you on Thu (tomorrow is birthday PTO :) | 16:58 |
| jroll | happy birthday!!!!! | 16:59 |
| Nisha | advance happy birthday dtantsur | 16:59 |
| dtantsur | thanks :) | 16:59 |
| *** dtantsur is now known as dtantsur|afk | 16:59 | |
| lucasagomes | dtantsur|afk, all the best, have a good night | 17:01 |
| lucasagomes | dtantsur|afk, happy birthday! | 17:01 |
| *** ChuckC_ has joined #openstack-ironic | 17:06 | |
| NobodyCam | oh happy b-day dtantsur|afk | 17:07 |
| *** ChuckC has quit IRC | 17:07 | |
| *** romcheg has quit IRC | 17:08 | |
| *** enikanorov has quit IRC | 17:17 | |
| NobodyCam | Nisha: are you around? | 17:18 |
| Nisha | hi NobodyCam | 17:19 |
| NobodyCam | Hi Nisha :) | 17:19 |
| *** ChuckC_ has quit IRC | 17:19 | |
| Nisha | NobodyCam, :) | 17:19 |
| NobodyCam | I was just looking over 100951 | 17:19 |
| Nisha | Yes | 17:19 |
| *** ndipanov has quit IRC | 17:20 | |
| Nisha | NobodyCam, Does it look fine to get merged | 17:20 |
| Nisha | ;) | 17:20 |
| NobodyCam | on lines 227-228 you point out fake driver will need to be updated, but it not called out in the work items section, are you planning on working on the fake driver too? | 17:21 |
| *** achanda has quit IRC | 17:21 | |
| Nisha | Not exactly, the code doesnt explicitly do any changes to fake driver. | 17:23 |
| *** achanda_ has joined #openstack-ironic | 17:24 | |
| *** dwalleck has quit IRC | 17:24 | |
| *** derekh has quit IRC | 17:27 | |
| *** alexpilotti has joined #openstack-ironic | 17:27 | |
| *** jcoufal_ has quit IRC | 17:27 | |
| *** achanda_ has quit IRC | 17:28 | |
| *** achanda has joined #openstack-ironic | 17:29 | |
| *** Marga_ has quit IRC | 17:29 | |
| *** mkerrin has quit IRC | 17:32 | |
| *** ndipanov has joined #openstack-ironic | 17:33 | |
| *** achanda has quit IRC | 17:33 | |
| adam_g | JayF, oh hey, i dusted this off and got it working last night https://review.openstack.org/#/c/129099/ | 17:35 |
| adam_g | http://logs.openstack.org/99/129099/7/check//check-tempest-dsvm-ironic-pxe_ssh/8494e82/logs/ironic-bm-logs/ | 17:35 |
| * Shrews learns he had more vacation time to burn than he thought. Tomorrow is my last day for the year. | 17:36 | |
| JayF | adam_g: niiice | 17:37 |
| NobodyCam | Shrews: awesome-ness | 17:38 |
| *** jiangfei has quit IRC | 17:43 | |
| *** jiangfei has joined #openstack-ironic | 17:44 | |
| NobodyCam | Nisha: that is so close. I added one spelling nit and two questions that I hope are easly answered. | 17:46 |
| Nisha | :( ok. Means again need to wait for rereview of the spec once i repost it | 17:46 |
| Nisha | anyway i will just look at | 17:46 |
| *** ChuckC_ has joined #openstack-ironic | 17:47 | |
| NobodyCam | Nisha: if you can answer the questions I'll give it a +2 and maybe we can poke some of the other spec cores to take a look at it today | 17:48 |
| Nisha | for fake driver it was required when we add abstract method | 17:49 |
| openstackgerrit | Lucas Alvares Gomes proposed openstack/ironic: PXE: Pass root device hints via kernel cmdline https://review.openstack.org/142109 | 17:50 |
| *** dwalleck has joined #openstack-ironic | 17:52 | |
| Nisha | NobodyCam, it doesnt require any changes to fake driver now | 17:52 |
| Nisha | it was required in previous implementation | 17:52 |
| NobodyCam | Nisha: so you can remove that statment? | 17:52 |
| Nisha | yes | 17:52 |
| NobodyCam | :) | 17:53 |
| Nisha | :) | 17:53 |
| Nisha | regarding dependencies, it is dependent on state machine but for my code to work i just added the required states in states.py | 17:53 |
| *** Marga_ has joined #openstack-ironic | 17:54 | |
| *** ChuckC_ has quit IRC | 17:54 | |
| *** romcheg has joined #openstack-ironic | 17:54 | |
| *** jiangfei has quit IRC | 17:54 | |
| Nisha | should i mention it in the spec? | 17:55 |
| *** jiangfei has joined #openstack-ironic | 17:55 | |
| *** achanda has joined #openstack-ironic | 17:56 | |
| NobodyCam | Nisha: ya. | 17:56 |
| *** ChuckC_ has joined #openstack-ironic | 17:56 | |
| Nisha | k | 17:56 |
| NobodyCam | :) | 17:56 |
| NobodyCam | thank you Nisha :) | 17:58 |
| Nisha | just giving link to state spec will be fine? | 18:00 |
| NobodyCam | Nisha: where? | 18:00 |
| Nisha | in the dependencies | 18:01 |
| *** dwalleck has quit IRC | 18:01 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic-specs: Discover node properties using new CLI node-inspect https://review.openstack.org/100951 | 18:03 |
| NobodyCam | for me I would spicifilly call out what states you need and how your planning on working around the case where the states do not yet exist | 18:03 |
| *** jeblair_ is now known as jeblair | 18:04 | |
| *** penick has joined #openstack-ironic | 18:04 | |
| Nisha | means whatever i said here? | 18:05 |
| Nisha | NobodyCam, ^^^ | 18:06 |
| *** k4n0 has quit IRC | 18:06 | |
| Nisha | NobodyCam, i can actually say that | 18:06 |
| Nisha | in the spec | 18:06 |
| *** marcoemorais has joined #openstack-ironic | 18:10 | |
| *** erwan_taf has quit IRC | 18:11 | |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic-specs: Discover node properties using new CLI node-inspect https://review.openstack.org/100951 | 18:12 |
| *** achanda has quit IRC | 18:12 | |
| Nisha | NobodyCam, ^^^^ | 18:12 |
| *** harlowja_away is now known as harlowja | 18:12 | |
| Nisha | NobodyCam, i had one ques | 18:12 |
| *** dwalleck has joined #openstack-ironic | 18:12 | |
| NobodyCam | sure | 18:13 |
| *** achanda has joined #openstack-ironic | 18:13 | |
| Nisha | Since for introspection we plan to implement iLO driver as a reference implementation, is it fine to disocver those properties which can be just used for scheduling and requires no action from the driver explicitly | 18:13 |
| Nisha | example Server Model | 18:14 |
| *** achanda has quit IRC | 18:14 | |
| Nisha | This is just a static data, and will be same for that hardware | 18:14 |
| *** achanda has joined #openstack-ironic | 18:14 | |
| NobodyCam | Nisha: like a copy these properites to this other node? (I may be miss understanding the question) | 18:16 |
| *** jrist has quit IRC | 18:16 | |
| Nisha | NobodyCam, means discover with introspection and add to node.properties as capabilities | 18:17 |
| Nisha | it requires no explicit action from driver | 18:17 |
| *** Masahiro has joined #openstack-ironic | 18:17 | |
| JayF | You mean you want to fill node.properties with lots of info from iLO | 18:18 |
| JayF | that you can schedule against | 18:18 |
| NobodyCam | Nisha: spec spec only covers limited properites (mem,disk,# of cores,arch) | 18:19 |
| NobodyCam | anything eles I would think would be out of scope? | 18:20 |
| NobodyCam | morning JayF :) | 18:20 |
| Nisha | yes, actually i added a statement in the reference implemnetation for this | 18:20 |
| Nisha | but dmitry said no so i just removed it | 18:20 |
| Nisha | NobodyCam, so i can have this in ilo driver spec? and can that be also targetted for kilo | 18:22 |
| Nisha | > | 18:22 |
| Nisha | ? | 18:22 |
| *** Masahiro has quit IRC | 18:22 | |
| NobodyCam | Nisha: I think I would initialy agree with dmitry here. | 18:23 |
| *** ChuckC_ is now known as ChuckC | 18:23 | |
| NobodyCam | that said, there is nothing saying we can't revisit that at a later date. Ie implament this spec and then come back and expand upon it | 18:24 |
| *** jrist has joined #openstack-ironic | 18:28 | |
| lucasagomes | NobodyCam, mind looking at https://review.openstack.org/#/c/138729/ when u get some time? | 18:28 |
| * lucasagomes replied to the comments | 18:29 | |
| NobodyCam | Nisha: added a comment and gave it a +2 | 18:29 |
| lucasagomes | NobodyCam, oh never mind I will update the spec accoringly to ramesh comment | 18:29 |
| lucasagomes | 1 sec | 18:29 |
| NobodyCam | lucasagomes: ack | 18:29 |
| *** EmilienM|afk is now known as EmilienM | 18:35 | |
| Nisha | NobodyCam, Thanks | 18:35 |
| *** faizan has joined #openstack-ironic | 18:36 | |
| Nisha | lucasagomes, introspection spec https://review.openstack.org/#/c/100951/ | 18:37 |
| * lucasagomes clicks | 18:37 | |
| *** HenryG_ has quit IRC | 18:48 | |
| *** HenryG has joined #openstack-ironic | 18:49 | |
| *** Marga_ has quit IRC | 18:55 | |
| *** afaranha has quit IRC | 18:55 | |
| lucasagomes | Nisha, is the spec based on the state machine spec that was merged? | 18:56 |
| lucasagomes | Nisha, cause there's some stuff like "puts node.provision_state as INSPECTED" which doesn't seem to be true, in the state machine it goes from INSPECTING to MANAGED | 18:57 |
| lucasagomes | Nisha, see L304 | 18:57 |
| lucasagomes | https://review.openstack.org/#/c/133828/14/specs/kilo/new-ironic-state-machine.rst | 18:57 |
| *** Marga_ has joined #openstack-ironic | 18:57 | |
| NobodyCam | lucasagomes: INSPECTED is listed as an Intermediate State | 18:58 |
| NobodyCam | so that is our target? | 18:59 |
| lucasagomes | NobodyCam, yup it should be target but never provision_state | 18:59 |
| lucasagomes | AFAIUI | 19:00 |
| openstackgerrit | Josh Gachnang proposed openstack/ironic-python-agent: Add standalone mode for IPA https://review.openstack.org/141957 | 19:00 |
| *** marcoemorais has quit IRC | 19:01 | |
| *** penick has quit IRC | 19:02 | |
| *** marcoemorais has joined #openstack-ironic | 19:03 | |
| NobodyCam | lucasagomes: see lines 127 - 131 of the state machine spec https://review.openstack.org/#/c/133828/14/specs/kilo/new-ironic-state-machine.rst | 19:04 |
| lucasagomes | NobodyCam, oh so Ironic will provision_state = *ED states?! | 19:05 |
| lucasagomes | hmm but for how long? I mean it sounds a bit useless | 19:06 |
| lucasagomes | devananda, got a sec ^ | 19:06 |
| devananda | those are transitive states | 19:06 |
| NobodyCam | that was my understandng | 19:07 |
| *** Marga_ has quit IRC | 19:07 | |
| devananda | also, why are you both linking to the review, rather than the doc? | 19:07 |
| *** Marga_ has joined #openstack-ironic | 19:07 | |
| *** afaranha has joined #openstack-ironic | 19:07 | |
| devananda | http://specs.openstack.org/openstack/ironic-specs/specs/kilo/new-ironic-state-machine.html | 19:07 |
| NobodyCam | for line numbers | 19:07 |
| NobodyCam | :-p | 19:08 |
| lucasagomes | devananda, right, so for the inspecting thing. it starts as provision_state = INSPECTING target_provision_state = INSPECTED. And finishs as provision_state = MANAGED and target_provision_state = none | 19:08 |
| lucasagomes | right? | 19:08 |
| devananda | oh | 19:08 |
| devananda | yes | 19:08 |
| lucasagomes | devananda, spec is easier to point which line :D | 19:08 |
| lucasagomes | but indeed rendered looks better | 19:08 |
| devananda | managed//none -> (inspect) -> inspecting//inspected -> inspected//inspected -> managed/none | 19:08 |
| devananda | inspected//inspected may never be visible or saved in the database | 19:09 |
| lucasagomes | yeah +1 to never save it in the db | 19:09 |
| lucasagomes | because it will be super quick | 19:09 |
| devananda | but it's part of the model, so the FSM queues off of "desired == goal" | 19:09 |
| devananda | sorry | 19:09 |
| devananda | when desired == goal, clear the goal | 19:09 |
| devananda | that's where we're having a problem with NOSTATE==None today | 19:10 |
| lucasagomes | I see | 19:10 |
| lucasagomes | yeah, None complicate things there indeed | 19:10 |
| devananda | but if NOSTATE were stored as "nostate", it would work just like every other transition | 19:10 |
| devananda | yep | 19:10 |
| lucasagomes | devananda, another nit, MANAGED sounds like a momentary states (since it's ends with -ED) | 19:10 |
| lucasagomes | wouldn;t MANAGEABLE be a better final state? | 19:11 |
| devananda | you are correct | 19:11 |
| rloo | devananda: I think that the problem I found with the target_provision_state may not have to do with NOSTATE but with some other logic :-( | 19:11 |
| devananda | or just MANAGE | 19:11 |
| lucasagomes | yeah | 19:11 |
| devananda | no -ED or -ING | 19:11 |
| lucasagomes | MANAGE seems better, shorter | 19:11 |
| devananda | rloo: oh? | 19:11 |
| lucasagomes | rloo, mind sharing the problem? | 19:12 |
| rloo | devananda: I think conceptually, the fsm's _target_state is to indicate the visible state, but eg, if deploying ->done, and an error occurs | 19:12 |
| rloo | you have deploying->FAIL and target is still deploydone. There needs to be smoething to say change the target if something fails. I think. | 19:13 |
| *** Marga_ has quit IRC | 19:13 | |
| *** Marga_ has joined #openstack-ironic | 19:13 | |
| devananda | rloo: I *think* that I merely preserved the current behavior. that was my goal, at least | 19:14 |
| devananda | rloo: even when the current behavior is illogical | 19:14 |
| NobodyCam | brb | 19:14 |
| rloo | devananda: but it isn't preserving the current behavior. that's the problem. | 19:14 |
| devananda | rloo: hm. ok, that's a problem | 19:14 |
| rloo | devananda: I'm going through the unit tests now, found another case where the target_provision_state is not right. | 19:14 |
| rloo | devananda: am thinking that I want to make sure our unit tests cover all the states/transitions in the fsm. | 19:15 |
| lucasagomes | rloo, yeah sounds good to have very good coverage of unittests on those states transitions | 19:16 |
| devananda | rloo: for the current model, or the new one? | 19:16 |
| rloo | devananda: there's the case of deploying -> target is deploydone. event 'done' happens, transition is deploying -> active. but active is not deploydone, so fsm._target stays deploydone :-( | 19:16 |
| rloo | devananda: for the current model. | 19:17 |
| devananda | rloo: I se | 19:22 |
| lucasagomes | devananda, re https://review.openstack.org/#/c/137762/. I created ApiMultiType which is basically a MultiType field with pre-defined types, otherwise I would have to pass the same list of types for all json-like fields in our API | 19:26 |
| devananda | rloo: isn't that already covered by test__do_node_deply? | 19:26 |
| devananda | 941 self.assertEqual(states.ACTIVE, node.provision_state) | 19:26 |
| devananda | 942 self.assertEqual(states.NOSTATE, node.target_provision_state) | 19:26 |
| *** achanda has quit IRC | 19:26 | |
| rloo | devananda: the problem is that the test is incorrect. it needs to call task.process_event('deploy') to get the fsm in the right (ugh) state. | 19:27 |
| * NobodyCam is back | 19:27 | |
| *** marcoemorais has quit IRC | 19:27 | |
| *** marcoemorais has joined #openstack-ironic | 19:27 | |
| rloo | devananda: the test right now, the fsm._target_state = None, which is why it passes. but it should be ._target_state = 'deploy complete' before the _do_node_deploy() is called. | 19:28 |
| *** faizan has quit IRC | 19:28 | |
| *** dwalleck has quit IRC | 19:29 | |
| rloo | devananda: might be easier if I 'showed you'. I'll push up a patch in 5 min or so with some updated unit tests. | 19:30 |
| devananda | I see it | 19:30 |
| *** ryanpetrello_ has joined #openstack-ironic | 19:30 | |
| devananda | has taken me a bit to context switch and realize that this is mid-way through the patch series | 19:30 |
| devananda | do_node_deploy gets refactored significantly in the next patch | 19:30 |
| devananda | so fixing it here is kinda annoying | 19:31 |
| devananda | fixing it on top of the next patch will make much more sense | 19:31 |
| *** ryanpetrello has quit IRC | 19:31 | |
| *** ryanpetrello_ is now known as ryanpetrello | 19:31 | |
| devananda | to me, at least | 19:31 |
| rloo | devananda: oh. I'm still stuck on the 139216 patch. I figured if I didn't understand that one, there was no use going on. | 19:31 |
| NobodyCam | rloo: it helped me to start at patch #6 and work backwards | 19:32 |
| rloo | NobodyCam: hmm. the problem is that the code/tests aren't correct in 139216. ok, let me take a quick stroll through the rest of the patches ;) | 19:33 |
| *** dprince has quit IRC | 19:34 | |
| devananda | rloo: as I said to Shrews yesterday, you won't really see the -effect- of these changes until the fourth and sixth patches | 19:36 |
| devananda | 140868 is a significant change to task_manager | 19:36 |
| devananda | which should make things much clearer | 19:36 |
| devananda | 140883 starts changing the individual drivers. that'll need more work to get all the states, but it's close | 19:37 |
| openstackgerrit | Clif Houck proposed openstack/python-ironicclient: Fix to properly issue an Unauthorized exception https://review.openstack.org/142204 | 19:38 |
| *** ifarkas has quit IRC | 19:39 | |
| rloo | devananda: ok, I'll look at all the changes (the last patch) then. It's all or nothin' ;) | 19:39 |
| *** marcoemorais has quit IRC | 19:42 | |
| rloo | devananda: except that https://review.openstack.org/#/c/140883/ fails the unit tests and pep8 :-( | 19:42 |
| *** marcoemorais has joined #openstack-ironic | 19:42 | |
| *** marcoemorais has quit IRC | 19:42 | |
| *** marcoemorais has joined #openstack-ironic | 19:43 | |
| rloo | devananda: failures seem to start with https://review.openstack.org/#/c/140868/ | 19:43 |
| *** marcoemorais has quit IRC | 19:44 | |
| *** marcoemorais has joined #openstack-ironic | 19:44 | |
| lucasagomes | NobodyCam, re SIZE for root device hints | 19:53 |
| NobodyCam | huh | 19:54 |
| lucasagomes | NobodyCam, I was thinking about how conflicting it would be with the local_gb stuff and was wondering whether we should remove it :/ | 19:54 |
| NobodyCam | that actualy kinda makes sense | 19:54 |
| Nisha | lucasagomes, i was looking at introspection comments | 19:55 |
| Nisha | | Verb | Initial State| Intermediate States | End State | | 19:55 |
| Nisha | +=======================================================+ | 19:55 |
| Nisha | | inspect | MANAGED | INSPECTING -> INSPECTED | MANAGED | | 19:55 |
| * NobodyCam is about to jump on a call | 19:55 | |
| lucasagomes | NobodyCam, yeah I will leave it out on the next patch-set | 19:55 |
| lucasagomes | Nisha, right, we were talking about it ^^ | 19:56 |
| NobodyCam | lucasagomes: ++ | 19:56 |
| Nisha | as i understand the introspetion should just transition state from INSPECTING to INSPECTED. The MANAGED is not done by introspection. It should be done via another API call to chaneg the state from INSPECTED to MANAGED | 19:57 |
| lucasagomes | Nisha, so the thing is, provision_state = INSPECTED may never get to the DB | 19:57 |
| lucasagomes | when the inspection is finished the privision_state will be MANAGED | 19:57 |
| Nisha | then when is INSPECTED set? | 19:58 |
| lucasagomes | afaiui never to provision_state | 19:58 |
| lucasagomes | it will be set to the target_provision_state | 19:58 |
| lucasagomes | provision_state == MANAGED. User requests inspection. provision_state == INSPECTING, target_provision_state == INSPECTED. Once finished provision_state == MANAGED, target_provision_state == NONE | 19:59 |
| Nisha | ohk | 19:59 |
| lucasagomes | the INSPECTING -> INSPECTED indicates that the node is mid-inspection | 19:59 |
| lucasagomes | but at the end it will move back to MANAGED | 20:00 |
| lucasagomes | (or *FAIL if failed) | 20:00 |
| Nisha | ok | 20:00 |
| Nisha | lucasagomes, regarding this comment "Isn't the `updated_at` field enough for this? What are the reason/use case to have a specific field indicating the last time the node was inspected?" | 20:01 |
| Nisha | if the failure case updated_at will not be changed as nothing is saved to node object | 20:02 |
| harlowja | yo yo; devananda NobodyCam rloo https://review.openstack.org/#/c/141961/ if u get bored | 20:03 |
| harlowja | might as well get that process going forward... | 20:03 |
| lucasagomes | Nisha, right, ok that makes sense | 20:03 |
| Nisha | lucasagomes, secondly updated_at will be changed whenever node-update is done but that doesnt mean that it has inspected the node | 20:04 |
| lucasagomes | Nisha, yeah that one I get, but that was the reason why I comment | 20:04 |
| Nisha | so last_inspected should be different than updated_at | 20:04 |
| lucasagomes | since inspecting == updating the node | 20:04 |
| lucasagomes | I mean, at a higher level | 20:04 |
| Nisha | lucasagomes, i didnt get | 20:05 |
| Nisha | its true when inspection is don ejust now | 20:06 |
| lucasagomes | Nisha, but I see the value of keeping the last time the node was inspected | 20:06 |
| *** Masahiro has joined #openstack-ironic | 20:06 | |
| Nisha | but updated_at will updated if anyone executes node-update after inspection | 20:06 |
| lucasagomes | Nisha, right, I mean. The result of having a node inspecting is that the node is updated accordingly with what was found during inspection | 20:07 |
| Nisha | Yes | 20:07 |
| lucasagomes | that's why I was thinking that updated_at might be enough for it | 20:07 |
| Nisha | No but updated_at is changed evn when we just do node-update | 20:08 |
| lucasagomes | Nisha, but it's ok to keep the last_inspected (or inspected_at) field to save the time of the last inspection | 20:08 |
| Nisha | Thanks | 20:08 |
| Nisha | so only state related change is expected? | 20:08 |
| Nisha | in the spec | 20:08 |
| lucasagomes | yup | 20:08 |
| Nisha | ok will just post the new patch. | 20:09 |
| *** pensu has quit IRC | 20:10 | |
| *** Masahiro has quit IRC | 20:11 | |
| lucasagomes | thanks | 20:11 |
| openstackgerrit | Nisha Agarwal proposed openstack/ironic-specs: Discover node properties using new CLI node-inspect https://review.openstack.org/100951 | 20:13 |
| *** marcoemorais has quit IRC | 20:13 | |
| Nisha | lucasagomes, NobodyCam ^^^^ | 20:13 |
| *** marcoemorais has joined #openstack-ironic | 20:14 | |
| *** Marga_ has quit IRC | 20:15 | |
| lucasagomes | I will call it a day | 20:19 |
| lucasagomes | have a good night everyone | 20:19 |
| JayF | night | 20:19 |
| *** lucasagomes is now known as lucas-dinner | 20:19 | |
| lucas-dinner | NobodyCam, last thing re hints... I was using a dict in the spec, but was thinking about using the same syntax as we have for capabilies | 20:20 |
| lucas-dinner | e.g opt1:value1,opt2:value2 | 20:20 |
| *** penick has joined #openstack-ironic | 20:20 | |
| lucas-dinner | so we can share the code to parse those | 20:20 |
| lucas-dinner | lemme know what you think | 20:20 |
| jroll | lucas-dinner: I mean, that's not really a valid reason | 20:22 |
| jroll | lucas-dinner: you don't need to parse a dict at all :P | 20:22 |
| lucas-dinner | jroll, right, yeah indeed | 20:22 |
| lucas-dinner | jroll, but looks incosistent | 20:23 |
| jroll | lucas-dinner: right, I can see other reasons | 20:23 |
| lucas-dinner | idk whether it would be better to capabilities to also use dicts instead of that ',' list | 20:23 |
| jroll | yeah idk | 20:23 |
| jroll | I think the thing is capabilities can have multiple values for one thing | 20:23 |
| jroll | like opt1:val1,opt1:val2 | 20:23 |
| lucas-dinner | jroll, right, my idea was that hints could also be a combination of things | 20:25 |
| lucas-dinner | size:value,vendor:value | 20:25 |
| jroll | lucas-dinner: yeah, I could see that | 20:25 |
| lucas-dinner | maybe not for this spec tho | 20:25 |
| jroll | idk | 20:25 |
| *** andreykurilin_ has joined #openstack-ironic | 20:25 | |
| jroll | I personally would make everything a dict | 20:25 |
| jroll | I think the capabilities thing comes from nova | 20:25 |
| jroll | not for any actual reason | 20:26 |
| lucas-dinner | jroll, yeah I also prefer dicts, since it's a json-like fields | 20:26 |
| lucas-dinner | and it should support dicts and lists | 20:26 |
| jroll | ++ | 20:26 |
| JayF | Curious what other Ironic folk think of this: https://review.openstack.org/#/c/141957/ going into IPA | 20:26 |
| lucas-dinner | idk why we made it a string in the first place | 20:26 |
| jroll | right, I think it's a nova thing | 20:26 |
| lucas-dinner | jroll, I can open a bug about the capabilities and yet maintain backward compat (in case it's a string we parse it as-is today) | 20:27 |
| jroll | lucas-dinner: right, would be nice | 20:27 |
| lucas-dinner | jroll, ack, thanks for the input | 20:27 |
| jroll | np :) | 20:27 |
| lucas-dinner | jroll, ah last thing... re size | 20:27 |
| jroll | now go eat! | 20:27 |
| lucas-dinner | ramesh had a good point, about conflicting with local_gb | 20:27 |
| lucas-dinner | jroll, hah yeah... just lemme sort the size out too | 20:28 |
| jroll | hmm | 20:28 |
| jroll | so the problem is that local_gb should be involved in that? | 20:28 |
| lucas-dinner | jroll, I was wondering whether we should keep size as a valid hint, I like it. But I understand it may conflict with our current model | 20:28 |
| jroll | at first glance, I would think local_gb takes precedence, if more than one disk matches, the size hint could be used | 20:29 |
| lucas-dinner | so local_gb is trick, idk how we would fill that out when inspecting a node with multiple disks for e.g | 20:29 |
| jroll | like say local_gb=20, you have a 40gb and 4000gb disk, size:40gb would be a valid hint imo | 20:29 |
| *** igordcard has joined #openstack-ironic | 20:29 | |
| lucas-dinner | jroll, right, so use the local_gb as a hint. Or still allow operators to set size as hint separately ? | 20:30 |
| * lucas-dinner finds it confusing | 20:30 | |
| jroll | well, actually | 20:30 |
| lucas-dinner | I was tihnking about leaving size out until we sort that out | 20:30 |
| jroll | we only use that in nova for scheduling | 20:30 |
| lucas-dinner | yeah, and we can/should use ironic without nova as well | 20:30 |
| jroll | nothing about choosing a disk once a node is claimed by the scheduler | 20:30 |
| jroll | sure, but ironic itself doesn't look at local_gb | 20:31 |
| lucas-dinner | jroll, oh actually not. we use local_gb to create the partitions | 20:31 |
| jroll | mmmmmmmmm | 20:31 |
| jroll | but not to decide anything | 20:31 |
| lucas-dinner | jroll, that's the confusing part that comes from nova bm etc... local_gb == root partition size | 20:31 |
| lucas-dinner | not disk size | 20:31 |
| jroll | oh | 20:32 |
| jroll | well | 20:32 |
| jroll | we should filter out any disks smaller than local_gb | 20:32 |
| *** jdob has left #openstack-ironic | 20:32 | |
| jroll | and after that use the size hint | 20:32 |
| jroll | (imho) | 20:32 |
| lucas-dinner | jroll, right, yeah makes sense to fail validation or so if disk hint < local_gb | 20:32 |
| jroll | yeah, ++ | 20:33 |
| lucas-dinner | jroll, ack I will update the spec accordingly later | 20:33 |
| lucas-dinner | jroll, ta much for all the inputs | 20:33 |
| jroll | no problem :) | 20:33 |
| * lucas-dinner go eat :) | 20:33 | |
| jroll | have a good night o/ | 20:33 |
| lucas-dinner | jroll, night | 20:34 |
| mrda-away | Morning ironic | 20:39 |
| *** mrda-away is now known as mrda | 20:39 | |
| *** marcoemorais has quit IRC | 20:41 | |
| *** marcoemorais has joined #openstack-ironic | 20:42 | |
| *** pelix has quit IRC | 20:44 | |
| *** Marga_ has joined #openstack-ironic | 20:45 | |
| *** ChuckC has quit IRC | 20:46 | |
| *** Marga_ has quit IRC | 20:50 | |
| *** chuckC_ has joined #openstack-ironic | 21:00 | |
| NobodyCam | night lucas-dinner | 21:01 |
| Nisha | NobodyCam, thanks for reviews. jroll, JayF, devananda reviews on https://review.openstack.org/#/c/100951/ | 21:06 |
| NobodyCam | :) | 21:06 |
| NobodyCam | Nisha: Thank you! | 21:07 |
| Nisha | NobodyCam, :) | 21:07 |
| *** ChuckC has joined #openstack-ironic | 21:11 | |
| jroll | Nisha: +2, leaving time for others | 21:21 |
| Nisha | jroll, thanks | 21:25 |
| Nisha | JayF, devananda reviews on https://review.openstack.org/#/c/100951/ | 21:27 |
| JayF | If I wasn't busy, I would've reviewed it 20 minutes ago | 21:28 |
| Nisha | JayF, ok :) | 21:29 |
| *** Marga_ has joined #openstack-ironic | 21:30 | |
| *** lucas-dinner has quit IRC | 21:33 | |
| NobodyCam | jroll: (and all IPA'ers) your generally okay with the direction 137363 is stating? | 21:35 |
| jroll | NobodyCam: I want it to happen, I need to review again | 21:35 |
| jroll | does that answer your question or should I read it? | 21:35 |
| * JayF adds it to his review list | 21:36 | |
| NobodyCam | no no I think is a good thing .. my question was really at the lib-if-a-ca-tion action items | 21:37 |
| jroll | oh yeah, I've been pushing for that | 21:37 |
| jroll | I don't want duplicate code :) | 21:37 |
| NobodyCam | ack :) | 21:38 |
| *** dprince has joined #openstack-ironic | 21:43 | |
| *** Masahiro has joined #openstack-ironic | 21:55 | |
| *** Masahiro has quit IRC | 22:00 | |
| PaulCzar | is there a driver for systems without power controls? something that just sets up pxe and waits for the user to power the machine up ? | 22:12 |
| NobodyCam | PaulCzar: https://github.com/openstack/ironic/blob/master/setup.cfg#L46 | 22:13 |
| devananda | PaulCzar: not upstream, no. | 22:14 |
| devananda | PaulCzar: a "logpower" driver was proposed a while back, but abandoned | 22:14 |
| NobodyCam | devananda: wouldn't fake_pxe do that? | 22:14 |
| devananda | NobodyCam: it'll return success to any power request (on or off or reboot) without notifying in any way that the user should do that | 22:15 |
| afaranha | Hello, I'm using devstack to run Keystone, Nova, Glance, Neutron, etc, in order to run ironic and communicates with them | 22:15 |
| afaranha | I'm following this guide: http://docs.openstack.org/developer/ironic/deploy/install-guide.html , but I having some issues | 22:15 |
| devananda | PaulCzar: do your systems have some other out of band power control (eg, a smart PDU)? or do you actually need a person to push a button? | 22:15 |
| afaranha | I don't have nova-scheduler service running on my machine, when I try service nova-scheduler restart I got an error: nova-compute: unrecognized service | 22:16 |
| PaulCzar | devananda: we have ipmi, but my hands are tied waiting on somebody else to wire up access | 22:16 |
| afaranha | do anyone knows how to solve this? | 22:16 |
| afaranha | I'm using Ubuntu 14.04 | 22:16 |
| NobodyCam | so if I "watch ironic node-show" I could see when to push the button? | 22:16 |
| JayF | afaranha: if you're using devstack, you should use our devstack docs | 22:17 |
| PaulCzar | yeah that was my plan to watch ironic node-show so I can test it as is | 22:17 |
| PaulCzar | then switch to the ipmi driver when it gets wired up | 22:17 |
| PaulCzar | ipmi is there, but the networks can't talk currently | 22:17 |
| PaulCzar | so I can manually ipmi it | 22:17 |
| JayF | afaranha: http://docs.openstack.org/developer/ironic/dev/dev-quickstart.html#deploying-ironic-with-devstack | 22:17 |
| devananda | NobodyCam: huh. I suppose taht'll just reflect the power state that Ironic /wants/ it to be in, not what it actually is | 22:17 |
| afaranha | JayF, but I'm don't intend to use devstack to run Ironic, I want that Ironic gets real machine nodes to deploy on them | 22:18 |
| afaranha | JayF, I'm using devtsack just to run the basic services | 22:18 |
| JayF | afaranha: I'm not sure a "half devstack" install of Ironic is something we support or desire to support :/ | 22:18 |
| devananda | PaulCzar: just tossing an idea out -- what if ironic-conductor SSH'd to some other jump host that had access to ipmi? | 22:18 |
| devananda | PaulCzar: would that address the issue? | 22:18 |
| devananda | afaranha: why are you using devstack at all, then? | 22:18 |
| devananda | afaranha: it does not preserve state across restarts | 22:19 |
| afaranha | I just wanna test Ironic with actual machines | 22:19 |
| devananda | afaranha: so use devstack, and have Ironic *manage* other machines | 22:19 |
| afaranha | I was suppose to be a fast deploy just to study ironic | 22:19 |
| devananda | afaranha: you don't need to run Ironic outside of devstack | 22:19 |
| PaulCzar | devananda: yeah I can figure something like that out, was just hoping for something like NobodyCam is suggesting to do a real quick boot test | 22:19 |
| devananda | PaulCzar: gotcha. what Nobodycam suggests will probably work for that | 22:20 |
| PaulCzar | perfect! thanks devananda and NobodyCam | 22:20 |
| afaranha | devananda, I tried with devstack, but it simulated a real machine, I want to make it run on machines, is there a way to do this using devstack? | 22:20 |
| NobodyCam | PaulCzar: note you'll be fighting timeouts so be quick with the power button | 22:20 |
| NobodyCam | :-p | 22:20 |
| devananda | afaranha: yes. change the configuration so that ironic uses the pxe_ipmitool driver | 22:20 |
| devananda | afaranha: and you may need to change the networking -- but all this is covered in the docs which JayF already linked to you | 22:21 |
| afaranha | devananda, Great! I'm gonna check that | 22:21 |
| afaranha | thanks :) | 22:21 |
| jroll | afaranha: to be clear, ironic does not run services on the nodes it manages (which is what your email to the list sounded like) | 22:23 |
| *** anderbubble has joined #openstack-ironic | 22:33 | |
| NobodyCam | are there any other openstack projects have a shared LIB like the one described in 137363? | 22:34 |
| *** marcoemorais has quit IRC | 22:35 | |
| NobodyCam | s/projects have/projects that have/ | 22:35 |
| *** marcoemorais has joined #openstack-ironic | 22:35 | |
| *** anderbubble has quit IRC | 22:37 | |
| JayF | NobodyCam: I think that lib should be oslo.* | 22:39 |
| *** romcheg has quit IRC | 22:41 | |
| NobodyCam | lol.. i'm sitting here worring that we'll end up with another ironic repo | 22:41 |
| * NobodyCam has a hard time keeping up as it is now | 22:41 | |
| *** dprince has quit IRC | 22:41 | |
| *** romcheg has joined #openstack-ironic | 22:47 | |
| *** chuckC_ has quit IRC | 22:49 | |
| *** alexpilotti has quit IRC | 22:50 | |
| *** krtaylor has quit IRC | 23:13 | |
| openstackgerrit | AKAHIRA Junya proposed openstack/ironic: Update command options in the Installation Guide https://review.openstack.org/142107 | 23:14 |
| openstackgerrit | Roman Prykhodchenko proposed openstack/ironic: Refactoring for Ironic policy https://review.openstack.org/126265 | 23:17 |
| *** penick has quit IRC | 23:23 | |
| *** krtaylor has joined #openstack-ironic | 23:26 | |
| *** andreykurilin_ has quit IRC | 23:30 | |
| *** Marga_ has quit IRC | 23:42 | |
| *** Masahiro has joined #openstack-ironic | 23:44 | |
| *** penick has joined #openstack-ironic | 23:44 | |
| *** Nisha has quit IRC | 23:47 | |
| *** Masahiro has quit IRC | 23:48 | |
| *** naohirot has joined #openstack-ironic | 23:52 | |
| naohirot | good morning ironic | 23:52 |
| NobodyCam | morning naohirot | 23:52 |
| naohirot | hi, good evening :) | 23:53 |
| *** anderbubble has joined #openstack-ironic | 23:53 | |
| * naohirot I'm reading this, http://www.rackspace.com/blog/openpower-opening-the-stack-all-the-way-down/ | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!