| *** openstack has joined #openstack-ironic | 12:41 | |
| *** rakesh_hs has quit IRC | 12:53 | |
| lucas-hungry | dtantsur, oh, I haven't read the reply | 12:54 |
|---|---|---|
| lucas-hungry | will take a look | 12:54 |
| *** lucas-hungry is now known as lucasagomes | 12:54 | |
| *** pensu has joined #openstack-ironic | 13:01 | |
| *** rloo has joined #openstack-ironic | 13:07 | |
| derekh | lucasagomes: I've kicked off CI for https://review.openstack.org/#/c/124061/1 , assuming it passes your good to merge | 13:21 |
| lucasagomes | derekh, w00t! thanks! | 13:21 |
| lucasagomes | I will keep an eye on jenkins and remove the -2 if it's all good | 13:22 |
| *** ifarkas has quit IRC | 13:23 | |
| Shrews | lucasagomes: do you have the etherpad link with the ongoing hash ring work? | 13:23 |
| Shrews | also, morning | 13:23 |
| lucasagomes | Shrews, morning, hmm lemme find | 13:24 |
| lucasagomes | Shrews, https://etherpad.openstack.org/p/68IcuXSqzz | 13:25 |
| Shrews | lucasagomes: ah yes. thank you sir | 13:25 |
| lucasagomes | yvw :) | 13:26 |
| * Shrews updates whiteboard with that link | 13:26 | |
| *** ifarkas has joined #openstack-ironic | 13:31 | |
| Shrews | umm, did review.o.o go away? | 13:34 |
| GheRivero | not for me | 13:39 |
| GheRivero | but etherpad is... sometimes | 13:39 |
| Shrews | i'm working from the car shop this morning. not sure if it's their wifi or not | 13:40 |
| *** pensu has quit IRC | 13:55 | |
| *** teju has left #openstack-ironic | 13:57 | |
| openstackgerrit | David Shrewsbury proposed a change to openstack/ironic: Add "active" tracking to nodes and conductors https://review.openstack.org/124493 | 13:59 |
| *** jistr has quit IRC | 14:00 | |
| rloo | happy Monday Ironickers :) | 14:00 |
| Shrews | morning rloo | 14:00 |
| rloo | lucasagomes: you OK with approving this? : https://review.openstack.org/#/c/124258/ | 14:01 |
| lucasagomes | rloo, oh haven't looked into that | 14:01 |
| rloo | morning Shrews | 14:01 |
| lucasagomes | rloo, yes it's ok | 14:01 |
| lucasagomes | I agree with deva, if needed we can remove later | 14:01 |
| rloo | lucasagomes: thx | 14:01 |
| lucasagomes | rloo, thanks and morning :) | 14:01 |
| *** jistr has joined #openstack-ironic | 14:01 | |
| dtantsur | rloo, Shrews, morning | 14:01 |
| Shrews | hi dtantsur | 14:02 |
| rloo | Afternoon lucasagomes, dtantsur. | 14:02 |
| GheRivero | morning all | 14:02 |
| rloo | afternoon GheRivero | 14:03 |
| *** pensu has joined #openstack-ironic | 14:09 | |
| openstackgerrit | Dmitry Tantsur proposed a change to openstack/ironic: Fix typo in agent validation code https://review.openstack.org/124675 | 14:11 |
| Shrews | lucasagomes: nice catch on the missing FK on 124493. fixed that | 14:21 |
| lucasagomes | Shrews, cool, I'm still working on the periodic task thing | 14:21 |
| lucasagomes | I'm getting a 401 when prepare try to cache the image :( | 14:21 |
| NobodyCam | Good morning Ironic | 14:22 |
| rloo | morning NobodyCam | 14:25 |
| NobodyCam | morning rloo :) | 14:26 |
| jroll | morning ironic :) | 14:29 |
| jroll | dtantsur: oopssssssss | 14:29 |
| dtantsur | :) | 14:29 |
| *** jrist has joined #openstack-ironic | 14:29 | |
| jroll | yeah, I would rc1 that | 14:30 |
| dtantsur | morning jroll, NobodyCam, GheRivero | 14:30 |
| jroll | dtantsur: +2'd | 14:31 |
| NobodyCam | morning jroll and dtantsur; Shrews and GheRivero too :) | 14:31 |
| openstackgerrit | A change was merged to openstack/ironic: Updated from global requirements https://review.openstack.org/124258 | 14:31 |
| GheRivero | hi there | 14:31 |
| lucasagomes | NobodyCam, morning | 14:32 |
| NobodyCam | d'oh | 14:32 |
| NobodyCam | and ofc lucasagomes too!! | 14:32 |
| NobodyCam | :-p | 14:33 |
| NobodyCam | its a monday | 14:33 |
| lucasagomes | :) | 14:33 |
| *** rwsu has joined #openstack-ironic | 14:33 | |
| openstackgerrit | David Shrewsbury proposed a change to openstack/ironic: Add "active" tracking to nodes and conductors https://review.openstack.org/124493 | 14:33 |
| Shrews | NobodyCam: morning | 14:33 |
| NobodyCam | :) | 14:33 |
| romcheg | Morning Shrews, NobodyCam and everyone else | 14:34 |
| *** k4n0 has quit IRC | 14:34 | |
| romcheg | ! | 14:34 |
| Shrews | o/ | 14:34 |
| NobodyCam | mornign romcheg :) | 14:34 |
| dtantsur | romcheg, o/ | 14:34 |
| Shrews | Has anyone received their email for PTL voting yet? | 14:41 |
| NobodyCam | only for TripleO | 14:41 |
| openstackgerrit | Dmitry Tantsur proposed a change to openstack/python-ironicclient: Small fixes for utils/{common_filters,common_params_for_list} https://review.openstack.org/124040 | 14:42 |
| Shrews | NobodyCam: odd they wouldn't all go out at the same time | 14:42 |
| dtantsur | rloo, ^^^ | 14:42 |
| NobodyCam | Shrews: many project only had one ptl canadate | 14:42 |
| NobodyCam | no need to vote on only one | 14:43 |
| lucasagomes | Shrews, I did for TripleO only | 14:43 |
| lucasagomes | but as NobodyCam said, I think Ironic won't need a vote | 14:43 |
| NobodyCam | brb | 14:44 |
| dtantsur | heh I've seen only one candidate for Ironic PTL, guess who :D | 14:44 |
| jroll | hm, I didn't get any | 14:46 |
| jroll | did nova have multiple candidates? | 14:46 |
| *** pensu has quit IRC | 14:47 | |
| *** pcrews has joined #openstack-ironic | 14:48 | |
| *** foexle has quit IRC | 14:49 | |
| rloo | thx dtantsur | 14:52 |
| lucasagomes | jroll, I think only michael still applied for nova ptl | 14:54 |
| jroll | aha, cool | 14:54 |
| lucasagomes | haven't check tho, but as far as I remember | 14:54 |
| lucasagomes | and I didn't get any pool for the nova ptl etc... so I'm assuming that's correct :) | 14:54 |
| *** hemna has quit IRC | 14:55 | |
| jroll | yeah, figured :) | 14:55 |
| *** todd_dsm has joined #openstack-ironic | 14:56 | |
| *** todd_dsm has quit IRC | 14:56 | |
| NobodyCam | fyi: https://wiki.openstack.org/wiki/PTL_Elections_September/October_2014#Candidates | 14:56 |
| jroll | nice, ty NobodyCam :) | 14:58 |
| jroll | only two elections, heh | 14:58 |
| lucasagomes | I wonder what happens if nobody is willing to be ptl | 15:00 |
| jroll | rm -rf $project | 15:00 |
| lucasagomes | lol | 15:02 |
| lucasagomes | idk, it kinda sounds like they should have a strategy for that | 15:02 |
| jroll | I'm sure they do | 15:02 |
| dtantsur | choose a random person from ATC's | 15:02 |
| dtantsur | will be fun | 15:02 |
| lucasagomes | heh | 15:03 |
| rloo | I asked last week, and devananda said they had no policy wrt no one running for ptl of a project. | 15:03 |
| Shrews | that would be an odd situation | 15:03 |
| Shrews | i think it should be, of the ATC, the one with the oldest commit :) | 15:04 |
| * Shrews jokes, obviously | 15:04 | |
| lucasagomes | heh | 15:06 |
| lucasagomes | it would be an anarchy project :P | 15:06 |
| NobodyCam | no ptl would seem like a dead or die-ing project :( | 15:07 |
| lucasagomes | hmm I don't see it that way | 15:07 |
| NobodyCam | :-p hehehe | 15:07 |
| devananda | morning, all | 15:08 |
| rloo | I was thinking that no PTL would default to the previous/current PTL, thereby saving the PTL a few minutes or whatever from doing the paperwork or whatever ;) | 15:08 |
| NobodyCam | good mornig devananda | 15:08 |
| lucasagomes | someone have seem it from glance client http://paste.openstack.org/show/116745/? | 15:08 |
| lucasagomes | devananda, morning | 15:08 |
| Shrews | morning devananda | 15:08 |
| rloo | ahh, all hail the new PTL;) Morning devananda. | 15:08 |
| lucasagomes | devananda, have a min to talk about the hash_ring take_over() | 15:08 |
| devananda | lol. morning, all :) | 15:08 |
| devananda | lucasagomes: yep. just saw your comment | 15:08 |
| lucasagomes | devananda, right there's more, didn't commented yet | 15:08 |
| devananda | only have a minute tho - gotta drive in a bit | 15:08 |
| devananda | oh, k | 15:08 |
| lucasagomes | devananda, 1) the thing about storing the condutor.id | 15:09 |
| lucasagomes | everytime the conductor rejoins the cluster, even if it has an old entry we are unregistering it and registering again | 15:09 |
| lucasagomes | so take_over() will be trigged on the same conductor | 15:09 |
| lucasagomes | idk if it's desirable, we may want to make the rejoin of the conductor to just update the list of drivers and update_at fields | 15:10 |
| devananda | lucasagomes: so that's one of the things I changed by using the online field | 15:10 |
| lucasagomes | 2) I'm having a hard time with glanceclient to cache the images, if I use the context coming from the periodic task I get a 401, and with an admin token I'm getting a http://paste.openstack.org/show/116745/ | 15:10 |
| lucasagomes | looking into that | 15:10 |
| devananda | it no longer deletes itself on shutdown | 15:10 |
| devananda | in part because of the FK, we actually can't "delete" the old conductor record | 15:11 |
| lucasagomes | devananda, hmm... it still does, I'm using all the patches, I can check it | 15:11 |
| lucasagomes | so it shouldn't? | 15:11 |
| *** romcheg has left #openstack-ironic | 15:11 | |
| devananda | it calls unregister, which should be calling "set online=False" | 15:11 |
| devananda | and making it appear offline eg to the hash ring | 15:11 |
| devananda | however | 15:11 |
| devananda | in a restart, the window there should be small | 15:11 |
| devananda | so yes, also checking the heartbeat would help | 15:12 |
| lucasagomes | devananda, oh... yeah | 15:12 |
| lucasagomes | devananda, I commented also in another review about the "online" | 15:12 |
| devananda | the difference between just heartbeat and heartbeat+online is an indication of intention | 15:12 |
| lucasagomes | I think it's a bit redudant, now I see one use for it, but isn't the heartbeat alone enough? | 15:12 |
| lucasagomes | since in a crash online may not be updated | 15:12 |
| devananda | no recent hearbeat? => maybe crashed, net split, or something else? | 15:12 |
| lucasagomes | and will be a misleading field | 15:13 |
| lucasagomes | right | 15:13 |
| devananda | no recent heartbeat + online=False => it definitely went away and isnt' coming back | 15:13 |
| lucasagomes | so online is more about "the conductor was gracefully stopped/started" | 15:13 |
| devananda | recent heartbeat + online=False => restart | 15:13 |
| devananda | yes | 15:13 |
| devananda | that may not be the best column name | 15:13 |
| devananda | and clearly i didn't document THAT very wel inline | 15:13 |
| lucasagomes | hah yeah | 15:13 |
| lucasagomes | I was really like hmmm | 15:13 |
| lucasagomes | right it makes more sense now | 15:13 |
| devananda | I also didn't implement any of the heartbeat+online checking bits | 15:14 |
| dtantsur | stepping afk, see you later | 15:14 |
| *** dtantsur is now known as dtantsur|afk | 15:14 | |
| lucasagomes | right, ok... I will see if I can sort the glance thing fix | 15:14 |
| lucasagomes | since that is the main problem right now | 15:15 |
| lucasagomes | and then we can sort that online bits | 15:15 |
| *** vinbs has joined #openstack-ironic | 15:15 | |
| lucasagomes | devananda, thanks | 15:15 |
| dtantsur|afk | folks, easy one for your stats: https://review.openstack.org/#/c/124675/ 4x +2, +1 for Mr. J | 15:16 |
| lucasagomes | + | 15:16 |
| lucasagomes | +a | 15:16 |
| dtantsur|afk | thnx | 15:16 |
| devananda | lucasagomes: stepping afk for about an hour, but still available on gtalk | 15:17 |
| lucasagomes | devananda, ack, I will continue taking a look into it | 15:17 |
| lucasagomes | talk to u there if needed | 15:17 |
| lucasagomes | cheers | 15:17 |
| *** Poornima has joined #openstack-ironic | 15:17 | |
| *** comstud has joined #openstack-ironic | 15:20 | |
| *** jistr has quit IRC | 15:22 | |
| *** jistr has joined #openstack-ironic | 15:22 | |
| openstackgerrit | A change was merged to openstack/python-ironicclient: Add unit tests for "ironic node-show" shell cmd https://review.openstack.org/120001 | 15:25 |
| *** Guest73730 is now known as mgagne | 15:29 | |
| *** mgagne has quit IRC | 15:29 | |
| *** mgagne has joined #openstack-ironic | 15:29 | |
| *** vinbs has quit IRC | 15:32 | |
| *** dlaube has joined #openstack-ironic | 15:34 | |
| *** shakamunyi has joined #openstack-ironic | 15:36 | |
| openstackgerrit | A change was merged to openstack/ironic: Fix typo in agent validation code https://review.openstack.org/124675 | 15:49 |
| *** jcoufal has quit IRC | 15:56 | |
| *** foexle has joined #openstack-ironic | 15:59 | |
| *** marcoemorais has joined #openstack-ironic | 16:14 | |
| devananda | back | 16:20 |
| *** eghobo has joined #openstack-ironic | 16:27 | |
| lucasagomes | devananda, wb | 16:28 |
| lucasagomes | I will post what I've so u take a look | 16:28 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: WIP: add periodic task to rebuild hash rings https://review.openstack.org/124610 | 16:30 |
| lucasagomes | devananda, https://github.com/openstack/ironic/blob/master/ironic/common/images.py#L289-L292 is failing with http://paste.openstack.org/show/116753/ | 16:31 |
| *** rushiagr is now known as rushiagr_away | 16:31 | |
| lucasagomes | if I comment https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/deploy_utils.py#L420-L423 things works | 16:32 |
| lucasagomes | >.< | 16:32 |
| lucasagomes | it fetchs the image and all, but fail to look at the images metadata | 16:32 |
| lucasagomes | :/ | 16:32 |
| devananda | this is happening from prepare() ? | 16:33 |
| devananda | yea ... bad context perhaps? | 16:34 |
| lucasagomes | yeah from prepare | 16:34 |
| lucasagomes | yup, the thing is | 16:34 |
| devananda | normally this is called with the user's context | 16:34 |
| lucasagomes | the context from the periodic task was giving me a 401 | 16:34 |
| devananda | but that's nota vailable from the periodic task | 16:34 |
| lucasagomes | so I started using the admin context | 16:34 |
| lucasagomes | now I'm hitting that other problem | 16:34 |
| lucasagomes | yeah | 16:34 |
| devananda | right | 16:34 |
| lucasagomes | I think it should work with the admin context, idk if it's a glance bug | 16:35 |
| lucasagomes | I'm trying to check it now | 16:35 |
| devananda | ironic may now require a different privilege with glance to do this | 16:35 |
| *** comstud has quit IRC | 16:35 | |
| devananda | check the service catalog for ironic's permissions | 16:35 |
| devananda | devstack sets up an account for ironic, eg. to talk to neutron | 16:35 |
| lucasagomes | oh right, lemme check that | 16:35 |
| devananda | it might not be auth'd for glance, since ironic didn't need to do that (outside of the context passed from nova) | 16:35 |
| devananda | well - it needs to. but it wasn't | 16:36 |
| lucasagomes | it makes sense, lemme check here my devstack and catalog | 16:37 |
| *** jistr has quit IRC | 16:38 | |
| *** comstud has joined #openstack-ironic | 16:41 | |
| *** ifarkas has quit IRC | 16:46 | |
| *** eghobo has quit IRC | 16:46 | |
| *** chuckC has quit IRC | 16:54 | |
| *** ChuckC_ has joined #openstack-ironic | 16:56 | |
| *** jrist has quit IRC | 16:58 | |
| *** pensu has joined #openstack-ironic | 16:59 | |
| NobodyCam | quick run to the store... brb | 17:02 |
| lucasagomes | I will brb too, will be back for the meeting | 17:02 |
| *** lucasagomes is now known as lucas-afk | 17:03 | |
| *** derekh has quit IRC | 17:06 | |
| *** rushiagr_away is now known as rushiagr | 17:09 | |
| *** pensu has quit IRC | 17:12 | |
| *** athomas has quit IRC | 17:13 | |
| *** ndipanov is now known as ndipanov_gone | 17:15 | |
| NobodyCam | and back | 17:16 |
| *** harlowja has joined #openstack-ironic | 17:16 | |
| *** harlowja has quit IRC | 17:22 | |
| Shrews | devananda: i think we need to name the new FK in 124493 to match what the alembic upgrade would use. otherwise, we won't be able to downgrade correctly if 'ironic-dbsync create-schema' is used instead of 'upgrade'. trying to confirm that now... | 17:22 |
| Shrews | s/name the new FK/name the new FK in models.py/ | 17:23 |
| devananda | Shrews: ++ | 17:23 |
| *** jrist has joined #openstack-ironic | 17:24 | |
| *** ChuckC_ has quit IRC | 17:24 | |
| *** harlowja has joined #openstack-ironic | 17:24 | |
| *** marcoemorais has quit IRC | 17:30 | |
| *** athomas has joined #openstack-ironic | 17:31 | |
| *** marcoemorais has joined #openstack-ironic | 17:33 | |
| *** marcoemorais has quit IRC | 17:33 | |
| *** marcoemorais has joined #openstack-ironic | 17:34 | |
| *** pensu has joined #openstack-ironic | 17:35 | |
| Shrews | yup. confirmed | 17:38 |
| openstackgerrit | David Shrewsbury proposed a change to openstack/ironic: Add "active" tracking to nodes and conductors https://review.openstack.org/124493 | 17:38 |
| NobodyCam | humm Keystone open'd for Kilo development... when are we opening up | 17:40 |
| devananda | NobodyCam: after RC1 is tagged | 17:43 |
| devananda | NobodyCam: which is when ever these bugs are fixed | 17:43 |
| NobodyCam | :) should we open the spec repo? | 17:44 |
| devananda | NobodyCam: we could? though I'm going to remain focused on getting these bugs fixed until they're actually done | 17:44 |
| NobodyCam | devananda: oh ya ++, just figured it would allow folks to start preping their specs | 17:45 |
| devananda | NobodyCam: let's land any preparatory changes, like the new section, etc | 17:46 |
| NobodyCam | devananda: like https://review.openstack.org/#/c/103225/ | 17:46 |
| NobodyCam | :) | 17:46 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add HashRingManager to wrap hash ring singleton https://review.openstack.org/120951 | 17:48 |
| NobodyCam | dtantsur|afk: have you tested https://review.openstack.org/#/c/121297 | 17:48 |
| devananda | Shrews: working on 124493 right now? mind if I rebase and poke at the "online" bits? | 17:50 |
| Shrews | devananda: go ahead | 17:51 |
| *** tariqne has joined #openstack-ironic | 17:51 | |
| *** marcoemorais has quit IRC | 17:52 | |
| *** marcoemorais has joined #openstack-ironic | 17:53 | |
| *** marcoemorais has quit IRC | 17:55 | |
| NobodyCam | just posting spec reviews that others may wish to look over: https://review.openstack.org/#/c/121297 | 17:55 |
| *** marcoemorais has joined #openstack-ironic | 17:56 | |
| *** marcoemorais has quit IRC | 17:58 | |
| *** marcoemorais has joined #openstack-ironic | 17:58 | |
| *** marcoemorais has quit IRC | 17:58 | |
| * NobodyCam calls t-mobile to yell... brb | 17:58 | |
| *** marcoemorais has joined #openstack-ironic | 17:59 | |
| *** lsmola has quit IRC | 18:01 | |
| *** ChuckC_ has joined #openstack-ironic | 18:03 | |
| *** chuckC has joined #openstack-ironic | 18:04 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Fix hash ring stability https://review.openstack.org/118932 | 18:09 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add "active" tracking to nodes and conductors https://review.openstack.org/124493 | 18:09 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add HashRingManager to wrap hash ring singleton https://review.openstack.org/120951 | 18:09 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Fix hash ring stability https://review.openstack.org/118932 | 18:13 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add "active" tracking to nodes and conductors https://review.openstack.org/124493 | 18:13 |
| devananda | fixed a rebase error | 18:13 |
| devananda | doh. need to fix again | 18:13 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Fix hash ring stability https://review.openstack.org/118932 | 18:14 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add "active" tracking to nodes and conductors https://review.openstack.org/124493 | 18:14 |
| *** marcoemorais has quit IRC | 18:16 | |
| *** marcoemorais has joined #openstack-ironic | 18:18 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: WIP: add periodic task to rebuild hash rings https://review.openstack.org/124610 | 18:20 |
| devananda | afk a few | 18:20 |
| devananda | i think the first 3 patches in that series are good now. going to work on the periodic task when I get back | 18:21 |
| *** marcoemorais1 has joined #openstack-ironic | 18:28 | |
| *** Poornima has quit IRC | 18:30 | |
| *** marcoemorais has quit IRC | 18:30 | |
| devananda | jroll: is teh agent-nv test passing at all yet? | 18:39 |
| jroll | devananda: not yet :( | 18:40 |
| devananda | jroll: think it'll be passing soon? | 18:40 |
| jroll | devananda: the VMs don't have enough RAM in there, was trying to find the best way to solve it, got pulled into other stuff | 18:40 |
| jroll | devananda: we could quickly hack it to set up 1024MB VMs when testing agent driver, but that's not viable long term | 18:40 |
| JayF | maybe more useful to have something that works at all than one that just uses resources and failing | 18:41 |
| JayF | even if we know we'd have to get the ram usage back down | 18:41 |
| devananda | yea, the wasted test resources was my concern at the moment | 18:41 |
| devananda | seems like moving taht test to the experimental pipeline for now would be better | 18:42 |
| devananda | until you have a solution that allows it to pass | 18:42 |
| *** rushiagr is now known as rushiagr_away | 18:42 | |
| devananda | OR what JayF suggested -- but that's somewhat misleading, and we're liable to break it again when we enable parallel testing in our jobs | 18:43 |
| JayF | I'd rather if we're going to push any change, we push the one to make it work :) | 18:43 |
| devananda | (which I hope is early in Kilo) | 18:43 |
| jroll | I have a solution to make it (hopefully) pass, it's set VM_RAM to 1024 | 18:44 |
| jroll | but as you say | 18:44 |
| NobodyCam | popping out for quick smoke b4 meeting | 18:45 |
| jroll | the only real solution, I think, is to get DIB working | 18:45 |
| jroll | we'll still need to use 1024 when testing coreos image built from source | 18:46 |
| devananda | jroll: that rules out the parallel testing against coreos images, but i think that's OK | 18:47 |
| jroll | agree | 18:47 |
| *** bandicot has joined #openstack-ironic | 18:48 | |
| *** lucas-afk is now known as lucasagomes | 18:50 | |
| NobodyCam | last call for agenda changes | 18:51 |
| devananda | jroll: i don't see a toggle for that in infra/config or devstack-gate right now. sounsd like something we'll need eventualy anyways, to support your coreos images. | 18:52 |
| jroll | devananda: right, just need to add it to dsg | 18:52 |
| jroll | devananda: there's already an if for the agent_ssh driver | 18:52 |
| NobodyCam | is yuriyz around? | 18:53 |
| *** BertieFulton has joined #openstack-ironic | 18:54 | |
| *** dtantsur|afk is now known as dtantsur | 18:56 | |
| dtantsur | NobodyCam, did not test, but assumed it works, as it managed to actually build docs as part of docs job :) | 18:57 |
| NobodyCam | yep | 18:57 |
| *** romcheg has joined #openstack-ironic | 19:00 | |
| NobodyCam | BertieFulton: want to jion us in #openstack-meeting | 19:02 |
| jroll | are the SAP folks working on anything specific? :) | 19:05 |
| NobodyCam | jroll: not yet this is their first time joining. | 19:06 |
| jroll | right, curious if the goal is to deploy ironic at SAP, or work on some particular integration, etc | 19:07 |
| *** marcoemorais1 has quit IRC | 19:11 | |
| NobodyCam | jroll: I have a working POC env in the SAP lab that can deploy a Sles (Suse linux Enterprize server) with the use of the tripleo seed | 19:13 |
| jroll | NobodyCam: neat | 19:14 |
| NobodyCam | :) | 19:14 |
| *** marcoemorais has joined #openstack-ironic | 19:15 | |
| *** marcoemorais has quit IRC | 19:21 | |
| *** marcoemorais has joined #openstack-ironic | 19:21 | |
| *** dprince has quit IRC | 19:49 | |
| *** harlowja has quit IRC | 19:50 | |
| *** marcoemorais has quit IRC | 19:54 | |
| *** harlowja has joined #openstack-ironic | 19:55 | |
| *** marcoemorais has joined #openstack-ironic | 19:55 | |
| Shrews | devananda: fyi, you'll need to remove your -2 for https://review.openstack.org/103225 | 19:56 |
| devananda | Shrews: ack, ty | 19:58 |
| *** marcoemorais has quit IRC | 19:58 | |
| *** marcoemorais has joined #openstack-ironic | 19:59 | |
| lucasagomes | alright it's dinner time here, have a good night everyone! | 19:59 |
| devananda | ok - I also need to switch cafes. the wifi here is terrible | 19:59 |
| *** lucasagomes is now known as lucas-dinner | 19:59 | |
| NobodyCam | great meeting all Thnk you :) | 20:00 |
| NobodyCam | thank even | 20:00 |
| devananda | bbiab | 20:00 |
| dtantsur | g'night :) | 20:00 |
| *** dtantsur is now known as dtantsur|afk | 20:00 | |
| NobodyCam | night dtantsur|afk and lucas-dinner :) | 20:00 |
| *** BertieFulton has left #openstack-ironic | 20:00 | |
| NobodyCam | brb | 20:04 |
| *** tariqne has left #openstack-ironic | 20:07 | |
| *** pensu has quit IRC | 20:07 | |
| *** andreykurilin_ has joined #openstack-ironic | 20:10 | |
| *** kincl has joined #openstack-ironic | 20:19 | |
| *** marcoemorais has quit IRC | 20:21 | |
| *** igordcard has joined #openstack-ironic | 20:22 | |
| *** todd_dsm has joined #openstack-ironic | 20:26 | |
| *** stevebaker has joined #openstack-ironic | 20:27 | |
| *** comstud has quit IRC | 20:38 | |
| *** comstud has joined #openstack-ironic | 20:38 | |
| *** marcoemorais has joined #openstack-ironic | 20:39 | |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Fix issue where thread kill() can 'stick' ipmi communications https://review.openstack.org/124873 | 20:56 |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Recover from kill() while in command https://review.openstack.org/124873 | 20:58 |
| *** jcoufal has joined #openstack-ironic | 21:02 | |
| NobodyCam | I'm somewhat concerned about the "This will cause all existing | 21:08 |
| NobodyCam | node -> conductor mappings to be invalid" comment | 21:09 |
| openstackgerrit | A change was merged to stackforge/pyghmi: Recover from kill() while in command https://review.openstack.org/124873 | 21:09 |
| Shrews | I think we need some documentation on the hash ring. The fact that the 'help' text for the HR options is several sentences long is a good indicator of that. | 21:10 |
| rloo | Shrews: ++ | 21:10 |
| Shrews | I'm having a devil of a time trying to wrap my head around how it's supposed to work | 21:10 |
| NobodyCam | do we expose reblance in the cli? | 21:11 |
| Shrews | eewww. i hope not | 21:11 |
| Shrews | that seems like the wrong thing to do | 21:12 |
| jroll | yes | 21:12 |
| jroll | self-healing ftw | 21:12 |
| NobodyCam | Shrews: how do we handle the comment on 118932 | 21:12 |
| Shrews | NobodyCam: what comment? | 21:13 |
| *** todd_dsm has quit IRC | 21:14 | |
| NobodyCam | and unless a rebalance is triggered immediately will cause existing deployed nodes to be unable to reboot | 21:14 |
| *** todd_dsm has joined #openstack-ironic | 21:14 | |
| Shrews | NobodyCam: oh. i don't know what the proposal is for the retriggering | 21:15 |
| Shrews | kill or add a conductor? | 21:16 |
| NobodyCam | lol.... 21:11 | Shrews > eewww. i hope not | 21:17 |
| *** Guest78856 has joined #openstack-ironic | 21:18 | |
| Shrews | well, my "eewww" was my concern for too easily exposing a very disruptive operation | 21:19 |
| *** Guest78856 is now known as annegentle | 21:19 | |
| NobodyCam | my concern was that a operator would have to add then remove a fake conductor to rebalance after that lands | 21:19 |
| Shrews | NobodyCam: yeah, valid concern that needs to be addressed | 21:20 |
| jroll | where does a rebalance need to be triggered? | 21:20 |
| jroll | or like, what situation, I should say | 21:20 |
| NobodyCam | lol well not forthe IPA folks | 21:21 |
| NobodyCam | ajroll: https://review.openstack.org/#/c/118932 | 21:21 |
| Shrews | NobodyCam: i think this review will address it: https://review.openstack.org/#/c/124610/ | 21:21 |
| Shrews | just need to trigger the periodic task | 21:21 |
| jroll | NobodyCam: but *what situation* would need a manual rebalance? | 21:22 |
| *** linggao has quit IRC | 21:22 | |
| jroll | oh, you're talking about the commit message | 21:23 |
| jroll | so it's only when updating an ironic environment to that code | 21:23 |
| jroll | just... rolling restart? | 21:23 |
| NobodyCam | yea any existing env | 21:23 |
| jroll | probably would need like a double rolling restart | 21:23 |
| *** todd_dsm has quit IRC | 21:23 | |
| *** todd_dsm has joined #openstack-ironic | 21:24 | |
| jroll | dunno if there's really a way to avoid that | 21:27 |
| devananda | back | 21:28 |
| NobodyCam | wb devananda | 21:29 |
| *** cinerama` is now known as cinerama | 21:30 | |
| devananda | yes, https://review.openstack.org/#/c/118932/9 will require restarting all ironic services | 21:30 |
| devananda | api and conductor | 21:30 |
| * devananda edits the commit message | 21:31 | |
| jroll | I mean, deploying code in general requires that | 21:31 |
| jroll | unless there's some magic in ironic I don't know about :P | 21:31 |
| devananda | jroll: right. except this changes *where* the nodes map to | 21:31 |
| devananda | so you can't just update one half (eg the API services) | 21:31 |
| jroll | oh, yeah | 21:31 |
| jroll | idk, I generally assume I need to update both | 21:32 |
| devananda | also, this patch doesn't include the periodic task yet | 21:32 |
| devananda | we might want to invert the patch order if we are concerned about per-patch support | 21:32 |
| NobodyCam | devananda: can we make special note in the chane log or you thinkk thats overkill | 21:33 |
| NobodyCam | change log even | 21:33 |
| devananda | NobodyCam: define "special note" | 21:33 |
| devananda | change log == commit message history | 21:33 |
| *** krtaylor has quit IRC | 21:34 | |
| NobodyCam | humm note sure. just concerned aboutthe change and folks knowing to restart | 21:35 |
| NobodyCam | snote/not | 21:35 |
| devananda | if folks are doing a release upgrade, they should restart everything | 21:35 |
| devananda | folks running off of trunk will need to know | 21:35 |
| devananda | s/they should/we can assume they will/ | 21:35 |
| *** krtaylor has joined #openstack-ironic | 21:36 | |
| NobodyCam | ack :) | 21:36 |
| *** todd_dsm has quit IRC | 21:40 | |
| rloo | devananda: i was wondering about what/if we needed to do anything wrt string freeze cuz of the added string: https://review.openstack.org/#/c/124225/ | 21:41 |
| devananda | rloo: I commented on that on rev2 | 21:43 |
| rloo | devananda: "Below here, I'm adding a new string to _LW(). I think it's important to log something here." | 21:44 |
| rloo | devananda: that comment. Is that all you need to do? the stringfreeze page has some points but I don't know if they apply here: https://wiki.openstack.org/wiki/StringFreeze#Medium.2FLarge_changes_and_changes_after_the_RC_release | 21:45 |
| devananda | rloo: it is only one word different from an existing string in the same file | 21:47 |
| devananda | s/provision/power/ | 21:47 |
| devananda | rloo: so it seems like a "minor" change to me | 21:48 |
| rloo | devananda: ok ;) Shall I approve it then, or wait for others to provide feedback? | 21:49 |
| devananda | rloo: lgtm, already has 2 +2 | 21:49 |
| rloo | devananda: done. | 21:50 |
| devananda | ty | 21:51 |
| rloo | devananda: it had 2 +2 and Jenkins blessing before I even looked at it, so thought maybe folks were waiting for ... something... ;) | 21:51 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Make hash ring mapping be more consistent https://review.openstack.org/118932 | 21:52 |
| * devananda updates commit message on ^, blowing away a +2 | 21:52 | |
| devananda | actually, going to rebase it again | 21:52 |
| devananda | I think this should be on top of the periodic task. Shrews - does taht make sense to you? | 21:53 |
| *** krtaylor has quit IRC | 21:53 | |
| Shrews | devananda: yep | 21:55 |
| Shrews | devananda: actually... | 21:58 |
| Shrews | the periodic task depends on is_active_on changes. i don't think you can do that | 21:58 |
| devananda | Shrews: oh? I was going to put the ring distribution fix on top of the other three | 21:59 |
| Shrews | oh wait. that's a separate change. | 21:59 |
| devananda | I think we have two scenario | 21:59 |
| *** jcoufal has quit IRC | 21:59 | |
| devananda | - change mapping, then add periodic task ==> an upgrade between these is unrecoverable | 22:00 |
| devananda | - add periodic task, then change mapping ==> an upgrade between these could yield a lot of unnecessary rebalancing | 22:00 |
| devananda | so I think #2 is better | 22:00 |
| *** krtaylor has joined #openstack-ironic | 22:01 | |
| Shrews | agreed #2 is better. but we should merge these fairly closely together. i don't see anyone running an in-between version | 22:01 |
| devananda | right | 22:03 |
| devananda | we need both in RC1 | 22:03 |
| * jroll runs the in between just to spite Shrews | 22:06 | |
| *** todd_dsm has joined #openstack-ironic | 22:07 | |
| * Shrews avoids OnMetal | 22:09 | |
| NobodyCam | 118932 has a couple of pep8 issues | 22:09 |
| NobodyCam | :( | 22:09 |
| jroll | rude, Shrews | 22:10 |
| *** mikedillion has joined #openstack-ironic | 22:10 | |
| jroll | :P | 22:10 |
| Shrews | hehe | 22:11 |
| * Shrews needs to make dinner. ciao folks | 22:11 | |
| NobodyCam | enjoy Shrews | 22:11 |
| * devananda rebases things | 22:12 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Make hash ring mapping be more consistent https://review.openstack.org/118932 | 22:13 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add "active" tracking to nodes and conductors https://review.openstack.org/124493 | 22:13 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add periodic task to rebuild conductor local state https://review.openstack.org/124610 | 22:13 |
| * devananda fixes pep8 issues, hopefully | 22:13 | |
| devananda | lucas-dinner: comments on your changes to 124610 ^ | 22:13 |
| *** igordcard has quit IRC | 22:16 | |
| *** bandicot has quit IRC | 22:20 | |
| devananda | adam_g: looks like we're not precaching the trusty image for DIB, so it's fetching (and sometimes failing to do so) from the 'net | 22:21 |
| devananda | adam_g: http://logs.openstack.org/51/120951/10/check/check-tempest-dsvm-ironic-pxe_ssh-postgres/832b1ce/logs/devstacklog.txt.gz | 22:21 |
| adam_g | devananda, yeah i saw that just a bit ago | 22:22 |
| devananda | know if there's an open bug? | 22:22 |
| NobodyCam | humm anyone seen this error: http://logs.openstack.org/51/120951/10/check/check-tempest-dsvm-ironic-pxe_ssh-postgres/832b1ce/logs/screen-g-api.txt.gz#_2014-09-29_18_39_15_259 | 22:22 |
| NobodyCam | lol | 22:22 |
| adam_g | devananda, i haven't seen one. ill get one opened. theres likely some image caching stuff in place already that we can use. | 22:23 |
| * NobodyCam should read b4 posting | 22:23 | |
| devananda | adam_g: awesome, ty | 22:23 |
| *** mikedillion has quit IRC | 22:23 | |
| devananda | NobodyCam: i think that's the issue i was just looking at | 22:24 |
| NobodyCam | devananda: yep | 22:24 |
| devananda | NobodyCam: seems like it should have failed sooner | 22:24 |
| NobodyCam | I have not seen that before | 22:24 |
| adam_g | hmm | 22:24 |
| openstackgerrit | A change was merged to openstack/ironic: Conductor changes target_power_state before starting work https://review.openstack.org/124225 | 22:26 |
| openstackgerrit | A change was merged to openstack/ironic-specs: Add a section for RPC API impact https://review.openstack.org/103225 | 22:30 |
| *** mikedillion has joined #openstack-ironic | 22:31 | |
| openstackgerrit | A change was merged to openstack/ironic-specs: Stop using intersphinx https://review.openstack.org/121297 | 22:31 |
| NobodyCam | devananda: :( check_uptodate.sh: ironic.conf.sample is not up to date. (118932) | 22:35 |
| devananda | yurg. i fixed it in some of the rebases | 22:35 |
| * devananda wants to kill that check | 22:35 | |
| NobodyCam | kill the check but keep the sample file? | 22:36 |
| NobodyCam | or remove both | 22:36 |
| devananda | kill the check. see the ML. | 22:36 |
| devananda | going to post a fix shortly | 22:38 |
| *** penick has joined #openstack-ironic | 22:40 | |
| *** todd_dsm has quit IRC | 22:43 | |
| *** achanda has joined #openstack-ironic | 22:43 | |
| devananda | what do ya'll think -- need a new CONF option for the timing of the periodic task to recheck hash rings? | 22:48 |
| devananda | or not? | 22:48 |
| devananda | jroll, NobodyCam, rloo ^ | 22:48 |
| jroll | devananda: as opposed to? | 22:49 |
| jroll | the power timeout option? | 22:49 |
| jroll | err s/timeout/loop/ | 22:49 |
| devananda | as opposed to using the default global timing | 22:49 |
| jroll | very yes please | 22:49 |
| devananda | yes what? | 22:49 |
| devananda | yes new CONF? yes default? | 22:49 |
| jroll | we have our power loop pretty high because it can be slow | 22:49 |
| jroll | yes new conf | 22:49 |
| devananda | hm, ok | 22:50 |
| devananda | so this has the same properties as the power loop | 22:50 |
| devananda | it only fires off max_periodic_workers at once | 22:50 |
| jroll | one property it doesn't have is using the crappy ipmi protocol :) | 22:50 |
| devananda | sure, but this is liable to pull a bunch of data from glance | 22:51 |
| devananda | running a lot of workers in parallel might not be what you want | 22:51 |
| *** todd_dsm has joined #openstack-ironic | 22:51 | |
| jroll | mmm | 22:51 |
| devananda | granted, pxe images, not user images | 22:51 |
| jroll | right... | 22:52 |
| jroll | I meant we have our power loop set to a high interval, not high number of workers | 22:52 |
| *** ChuckC_ has quit IRC | 22:52 | |
| jroll | though I see your point | 22:52 |
| *** ChuckC_ has joined #openstack-ironic | 22:54 | |
| *** todd_dsm has quit IRC | 22:54 | |
| * devananda fixes conf file | 22:56 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Make hash ring mapping be more consistent https://review.openstack.org/118932 | 22:56 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add periodic task to rebuild conductor local state https://review.openstack.org/124610 | 22:56 |
| *** penick has quit IRC | 23:00 | |
| *** marcoemorais has quit IRC | 23:01 | |
| *** marcoemorais has joined #openstack-ironic | 23:01 | |
| *** marcoemorais has quit IRC | 23:02 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Stop running check_uptodate in the pep8 testenv https://review.openstack.org/124919 | 23:02 |
| *** marcoemorais has joined #openstack-ironic | 23:02 | |
| devananda | also, I'd be forever grateful if anyone has time to test the hash-ring patch series - especially the periodic tas - on a multi-node cluster | 23:03 |
| devananda | with the pxe driver | 23:03 |
| devananda | I will do it, ofc, but the mroe folks taht play with it the better | 23:03 |
| NobodyCam | devananda: is that a check that the gate should run too? | 23:04 |
| NobodyCam | it would costly | 23:04 |
| NobodyCam | *would be | 23:04 |
| devananda | NobodyCam: yes. see my commit message | 23:05 |
| devananda | it should run in the check pipeline, not the gate pipeline | 23:05 |
| NobodyCam | hummm py 26 failing on 118932 | 23:05 |
| NobodyCam | :) | 23:05 |
| NobodyCam | hummm https://jenkins06.openstack.org/job/gate-ironic-python26/850/console | 23:06 |
| NobodyCam | AttributeError: 'HashRingTestCase' object has no attribute 'assertLess' | 23:07 |
| devananda | oh, interesting | 23:08 |
| NobodyCam | 27 passes | 23:08 |
| devananda | some times I wish we weren't supporting old versions of redhat ... | 23:08 |
| devananda | https://docs.python.org/2/library/unittest.html | 23:09 |
| devananda | assertLess was added in 2.7 | 23:09 |
| NobodyCam | nice | 23:09 |
| jroll | and rhel didn't backport it like OrderedDict? | 23:09 |
| devananda | NobodyCam: I'm inclined, for right now, not to bother adding another test that non-votingly posts comments on our patches saying whether the config file is out of date | 23:14 |
| NobodyCam | jroll: hummm https://pypi.python.org/pypi/unittest2 | 23:14 |
| devananda | let's go back to the good ol' days of just doing that manually now and then | 23:14 |
| devananda | it's as easy as "tox -egenconfig" | 23:14 |
| jroll | NobodyCam: something something global requirements | 23:14 |
| NobodyCam | lol | 23:15 |
| NobodyCam | devananda: I'm ok with that.. | 23:15 |
| NobodyCam | we could also auto build it in to docs/ and remove it from the repo | 23:16 |
| devananda | NobodyCam: this is the standard location that all the projects share | 23:18 |
| NobodyCam | yea, though I thought several have already removed the sample | 23:19 |
| * NobodyCam is probibly wrong | 23:19 | |
| NobodyCam | oh /me never knew about assertAlmostEqual and assertNotAlmostEqual. | 23:20 |
| NobodyCam | for fuzzy math ? | 23:20 |
| *** andreykurilin_ has quit IRC | 23:21 | |
| devananda | nova removed teh sample | 23:21 |
| devananda | several other projects got (-) feedback when they went t o | 23:22 |
| devananda | as a consumer of other projects, i appreciate when they have a smple config for me | 23:22 |
| devananda | sample | 23:22 |
| devananda | so I don't want to remove it from ironic | 23:22 |
| NobodyCam | I also like having it! for in doc or repo is fine, just so long as I have access to it | 23:23 |
| * devananda rebases and fixes assertLess | 23:23 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Make hash ring mapping be more consistent https://review.openstack.org/118932 | 23:23 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add "active" tracking to nodes and conductors https://review.openstack.org/124493 | 23:23 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add periodic task to rebuild conductor local state https://review.openstack.org/124610 | 23:24 |
| NobodyCam | but ack +1 | 23:24 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Add HashRingManager to wrap hash ring singleton https://review.openstack.org/120951 | 23:24 |
| devananda | ok, gonna head home before it starts raining too much more | 23:24 |
| NobodyCam | ok stay dry | 23:24 |
| NobodyCam | :) | 23:24 |
| *** mikedillion has quit IRC | 23:25 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/ironic: Stop running check_uptodate in the pep8 testenv https://review.openstack.org/124919 | 23:28 |
| *** romcheg has left #openstack-ironic | 23:39 | |
| *** lucas-dinner has quit IRC | 23:42 | |
| *** shakamunyi has quit IRC | 23:51 | |
| *** achanda has quit IRC | 23:55 | |
| *** harlowja is now known as harlowja_away | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!