| *** openstackgerrit has quit IRC | 00:19 | |
| *** openstackgerrit has joined #openstack-ironic | 00:20 | |
| NobodyCam | Shrews: rloo: I will dig in to that math. and have something for the morning :) | 00:22 |
|---|---|---|
| NobodyCam | Thnank you both :) | 00:22 |
| rloo | thx NobodyCam. (Sorry about that.) | 00:23 |
| NobodyCam | no no nothinhg to be sorry about | 00:23 |
| NobodyCam | I actully very much agree :) | 00:23 |
| NobodyCam | actually even | 00:23 |
| rloo | heh | 00:24 |
| *** matsuhashi has joined #openstack-ironic | 00:29 | |
| *** openstackgerrit has quit IRC | 00:34 | |
| *** openstackgerrit has joined #openstack-ironic | 00:35 | |
| adam_g | while provisioning multiple nodes at once, does dd'ing from the conductor happen in serial? | 00:38 |
| devananda | adam_g: I believe it happens in parallel. But IMBW ... | 00:42 |
| devananda | actually i'm fairly confident it happens in parallel | 00:43 |
| adam_g | devananda, ya, looks like it. might test some locking there to see if it helps deploying 30 nodes at once | 00:43 |
| adam_g | (from a single conductor) | 00:43 |
| devananda | adam_g: lifeless' data on time-to-deploy 30 nodes strongly says it happens in parallel | 00:43 |
| devananda | I'd love to see that data become public reference at some point :) | 00:43 |
| devananda | also, I need to run to dinner .... bbl! | 00:44 |
| adam_g | devananda, ive only been messing with this test rack this week and seeing lots of different kinds of things fall over during deployment, that causes instances to get rescheduled and later fail | 00:44 |
| adam_g | random tftp timeouts, iscsi timeouts, i/o errors on dd'ing, etc | 00:44 |
| *** openstackgerrit has quit IRC | 00:49 | |
| *** openstackgerrit has joined #openstack-ironic | 00:50 | |
| *** zdin0bot has quit IRC | 00:51 | |
| *** zdin0bot has joined #openstack-ironic | 00:52 | |
| *** zdin0bot has quit IRC | 01:03 | |
| *** foliveira has joined #openstack-ironic | 01:33 | |
| *** foliveira has left #openstack-ironic | 01:33 | |
| *** hipster has joined #openstack-ironic | 01:36 | |
| Shrews | NobodyCam: yeah, sorry. :( but i was thinking, are we trying to do this the hard way? isn't there a tool that will do this for us? never used it, but sgdisk seems to be a good candidate | 01:39 |
| Shrews | NobodyCam: sgdisk --zap-all : "Zap (destroy) the GPT and MBR data structures and then exit." | 01:40 |
| rloo | Shrews: there's a comment about a bug with sgdisk :-( | 01:59 |
| *** martyntaylor has quit IRC | 02:01 | |
| *** dkehn_ has joined #openstack-ironic | 02:24 | |
| *** rloo has quit IRC | 02:24 | |
| *** hipster has quit IRC | 02:25 | |
| *** dkehnx has quit IRC | 02:27 | |
| *** dkehn_ is now known as dkehnx | 02:35 | |
| *** matsuhas_ has joined #openstack-ironic | 02:55 | |
| *** sanek12 has joined #openstack-ironic | 02:56 | |
| *** nosnos_ has joined #openstack-ironic | 02:57 | |
| *** GheRiver1 has joined #openstack-ironic | 02:57 | |
| *** linggao has quit IRC | 02:57 | |
| lifeless | devananda: I'm not sure the data says the dd happens in parallel. Its slow enough that I'd suspect network contention in parallel, or you could say its fast enough .. arguments either way. | 02:58 |
| *** mkerrin1 has joined #openstack-ironic | 02:58 | |
| lifeless | devananda: but 2GB image, 10Gb network, the dd is easily a candidate for the inter-node gap | 02:58 |
| *** christop1eraedo has joined #openstack-ironic | 02:58 | |
| lifeless | devananda: (And I'd really lke to keep it serialised for mean time-to-deploy minimisation) | 02:59 |
| *** sanek11 has quit IRC | 03:03 | |
| *** blamar has quit IRC | 03:03 | |
| *** matsuhashi has quit IRC | 03:03 | |
| *** GheRivero has quit IRC | 03:03 | |
| *** pquerna has quit IRC | 03:03 | |
| *** nosnos has quit IRC | 03:03 | |
| *** anteaya has quit IRC | 03:03 | |
| *** mkerrin has quit IRC | 03:03 | |
| *** christopheraedo has quit IRC | 03:03 | |
| *** russell_h has quit IRC | 03:03 | |
| *** russell_h has joined #openstack-ironic | 03:03 | |
| *** russell_h has quit IRC | 03:03 | |
| *** russell_h has joined #openstack-ironic | 03:04 | |
| *** anteaya has joined #openstack-ironic | 03:04 | |
| *** russell_h has quit IRC | 03:05 | |
| *** russell_h has joined #openstack-ironic | 03:05 | |
| *** uberj has quit IRC | 03:05 | |
| *** uberj_ has joined #openstack-ironic | 03:06 | |
| *** pquerna has joined #openstack-ironic | 03:09 | |
| *** hipster has joined #openstack-ironic | 03:10 | |
| *** blamar has joined #openstack-ironic | 03:14 | |
| *** ramineni has joined #openstack-ironic | 03:21 | |
| *** boris-42 has quit IRC | 03:23 | |
| *** boris-42 has joined #openstack-ironic | 03:29 | |
| *** romcheg has joined #openstack-ironic | 03:29 | |
| *** romcheg has left #openstack-ironic | 03:29 | |
| *** Mikhail_D_ltp has joined #openstack-ironic | 03:34 | |
| *** eghobo has joined #openstack-ironic | 03:35 | |
| *** coolsvap|afk is now known as coolsvap | 03:49 | |
| *** Mikhail_D_ltp has quit IRC | 03:52 | |
| *** nosnos_ has quit IRC | 03:52 | |
| *** hipster has quit IRC | 03:55 | |
| *** rameshg87 has joined #openstack-ironic | 04:19 | |
| openstackgerrit | Sirushti Murugesan proposed a change to openstack/ironic: Adds support for deploying whole disk images https://review.openstack.org/73054 | 04:22 |
| *** eghobo has quit IRC | 04:25 | |
| openstackgerrit | Sirushti Murugesan proposed a change to openstack/ironic: Adds support for deploying whole disk images https://review.openstack.org/73054 | 04:27 |
| *** eghobo has joined #openstack-ironic | 04:31 | |
| *** nosnos has joined #openstack-ironic | 04:35 | |
| *** lazy_prince has joined #openstack-ironic | 04:41 | |
| *** lazy_prince is now known as killer_prince | 04:41 | |
| *** krtaylor has joined #openstack-ironic | 04:41 | |
| *** igor_ has joined #openstack-ironic | 04:59 | |
| *** igor__ has quit IRC | 05:02 | |
| *** harlowja_ is now known as harlowja_away | 05:17 | |
| *** vkdrao has joined #openstack-ironic | 05:32 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/ironic: Imported Translations from Transifex https://review.openstack.org/88508 | 06:03 |
| *** datajerk has joined #openstack-ironic | 06:05 | |
| *** eghobo has quit IRC | 06:06 | |
| *** radsy has quit IRC | 06:08 | |
| *** ifarkas has joined #openstack-ironic | 06:17 | |
| *** geertj has joined #openstack-ironic | 06:29 | |
| Mikhail_D_wk | Good morning Ironic !!! :) | 06:49 |
| *** jcoufal has joined #openstack-ironic | 06:53 | |
| mrda | hey Mikhail_D_wk | 07:02 |
| *** ndipanov has joined #openstack-ironic | 07:13 | |
| openstackgerrit | A change was merged to openstack/ironic: Fix wrong usage of mock methods https://review.openstack.org/93532 | 07:14 |
| *** vkdrao has quit IRC | 07:23 | |
| *** rameshg87 has quit IRC | 07:38 | |
| *** martyntaylor has joined #openstack-ironic | 07:39 | |
| dtantsur | Morning Ironic, morning mrda, Mikhail_D_wk :) | 07:51 |
| Mikhail_D_wk | dtantsur morning :) | 07:51 |
| mrda | hi dtantsur | 07:51 |
| dtantsur | ifarkas, morning :) based on yesterday's conversation with devananda, we should move downloading instance images to deploy(). Do you have time for it? If you don't, I can do. | 07:52 |
| dtantsur | ifarkas, I suggest not touching existing patch, as it's ready for merging | 07:52 |
| dtantsur | (or was it merged already?) | 07:53 |
| *** martyntaylor has joined #openstack-ironic | 07:53 | |
| *** foexle has joined #openstack-ironic | 07:55 | |
| ifarkas | dtantsur, morning! which one is the existing patch? | 07:56 |
| dtantsur | ifarkas, https://review.openstack.org/#/c/92340/ | 07:56 |
| dtantsur | ifarkas, I see it was -1'ed, so maybe add there? | 07:56 |
| *** rameshg87 has joined #openstack-ironic | 07:57 | |
| ifarkas | dtantsur, ahh, yeah. ok, I will add it, it makes sense to update the patch this way | 07:57 |
| dtantsur | thank you :) left a comment as a reminder | 07:58 |
| *** vkdrao has joined #openstack-ironic | 07:58 | |
| *** jistr has joined #openstack-ironic | 07:59 | |
| mrda | ok, enough for today. Until tomorrow! | 08:03 |
| *** mrda is now known as mrda_away | 08:03 | |
| *** vkdrao has quit IRC | 08:03 | |
| *** vkdrao has joined #openstack-ironic | 08:04 | |
| ifarkas | dtantsur, lol, thanks :-) | 08:06 |
| *** pelix has joined #openstack-ironic | 08:15 | |
| *** derekh_ has joined #openstack-ironic | 08:15 | |
| *** GheRiver1 is now known as GheRivero | 08:34 | |
| *** rameshg87 has quit IRC | 08:37 | |
| *** rameshg87 has joined #openstack-ironic | 08:55 | |
| *** athomas has joined #openstack-ironic | 09:01 | |
| *** sanek12 is now known as agordeev | 09:05 | |
| *** lucasagomes has joined #openstack-ironic | 09:25 | |
| *** vkdrao has quit IRC | 09:39 | |
| openstackgerrit | A change was merged to openstack/ironic: Renaming ironicclient exceptions in nova driver https://review.openstack.org/90040 | 10:34 |
| *** athomas has quit IRC | 10:37 | |
| openstackgerrit | A change was merged to openstack/ironic: Using system call for downloading files https://review.openstack.org/92560 | 10:39 |
| *** coolsvap is now known as coolsvap|afk | 10:51 | |
| *** ramineni has quit IRC | 10:58 | |
| *** blamar has quit IRC | 11:04 | |
| *** athomas has joined #openstack-ironic | 11:15 | |
| *** zul has joined #openstack-ironic | 11:18 | |
| *** vkdrao has joined #openstack-ironic | 11:21 | |
| *** vkdrao has quit IRC | 11:21 | |
| *** matsuhas_ has quit IRC | 11:27 | |
| openstackgerrit | Dmitry Tantsur proposed a change to openstack/ironic: Add some real-world testing on DiskPartitioner https://review.openstack.org/94620 | 11:33 |
| *** rameshg87 has quit IRC | 11:43 | |
| *** athomas has quit IRC | 11:55 | |
| openstackgerrit | Imre Farkas proposed a change to openstack/ironic: Cleanup images after deployment https://review.openstack.org/92340 | 11:56 |
| openstackgerrit | Imre Farkas proposed a change to openstack/ironic: Cleanup images after deployment https://review.openstack.org/92340 | 12:01 |
| sirushti | dtantsur, ping | 12:02 |
| dtantsur | sirushti, hi | 12:02 |
| sirushti | dtantsur, thanks for the review :-) had a question | 12:02 |
| dtantsur | sure | 12:03 |
| sirushti | so if I understand correctly, you're asking me to have two API calls to deploy_utils? | 12:03 |
| dtantsur | it's re https://review.openstack.org/#/c/73054/ right? | 12:04 |
| sirushti | yes | 12:04 |
| sirushti | sorry | 12:04 |
| sirushti | that | 12:04 |
| dtantsur | yes. my suggestion is to split 1 deploy() function into 2 for 2 cases, as they require somewhat different arguments do somewhat different things | 12:05 |
| sirushti | ah, okay so we call whichever appropriate from pxe.py | 12:05 |
| dtantsur | I think it makes sense, yes | 12:05 |
| sirushti | is the decorator pattern still necessary? | 12:06 |
| dtantsur | I don't think so | 12:07 |
| sirushti | Cool, sounds good, thanks! :-) | 12:07 |
| dtantsur | ur welcome :) | 12:07 |
| *** athomas has joined #openstack-ironic | 12:11 | |
| openstackgerrit | Imre Farkas proposed a change to openstack/ironic: Cleanup images after deployment https://review.openstack.org/92340 | 12:17 |
| *** geertj has quit IRC | 12:18 | |
| openstackgerrit | Imre Farkas proposed a change to openstack/ironic: Cleanup images after deployment https://review.openstack.org/92340 | 12:23 |
| dtantsur | Hi folks! Anyone has idea, why we don't have `parted` tool on Python 2.6 gate boxes? | 12:25 |
| *** geertj has joined #openstack-ironic | 12:25 | |
| dtantsur | I also highly suspect that parted behaves differently on Precise and Fedora 20... | 12:29 |
| dtantsur | Can someone with precise run this test: https://review.openstack.org/#/c/94620 please? | 12:30 |
| *** nosnos has quit IRC | 12:31 | |
| openstackgerrit | Dmitry Tantsur proposed a change to openstack/ironic: Add some real-world testing on DiskPartitioner https://review.openstack.org/94620 | 12:40 |
| *** blamar has joined #openstack-ironic | 12:42 | |
| jroll | lucasagomes: about instance_info, I started on it but haven't done much yet. if you want to take it, go ahead, otherwise I'll probably have something up middle of next week | 12:43 |
| lucasagomes | jroll, cool, I got a local patch moving it here | 12:44 |
| lucasagomes | jroll, testing it right now | 12:44 |
| lucasagomes | moving it == moving the pxe parameters to the instance info | 12:44 |
| *** ramineni has joined #openstack-ironic | 12:45 | |
| jroll | lucasagomes: wow, you rule | 12:45 |
| dtantsur | lucasagomes, hi! Know what I noticed re partitioning? On Precise parted, when told to have the last partition at SIZE-1 actually expands it to SIZE! | 12:46 |
| lucasagomes | jroll, nah, probably I will have some headaches with the unittests yet :P | 12:46 |
| *** jgrimm has joined #openstack-ironic | 12:46 | |
| dtantsur | that's not the case on Fedora | 12:46 |
| lucasagomes | dtantsur, lol what!? | 12:46 |
| lucasagomes | dtantsur, ok | 12:46 |
| lucasagomes | it might be some syntax problem | 12:46 |
| lucasagomes | parted accept negative numbers | 12:46 |
| dtantsur | lucasagomes, see how my test now looks like : https://review.openstack.org/#/c/94620/6/ironic/tests/drivers/test_deploy_utils.py | 12:47 |
| dtantsur | test_whole_disk in the end | 12:47 |
| lucasagomes | so if you create a partition which starts at 10 and ends at -1 | 12:47 |
| lucasagomes | -1 here means end of the disk | 12:47 |
| dtantsur | lucasagomes, but I create mkpart XX 19. On Fedora 19 means 19, on Ubuntu it means 20 :) | 12:47 |
| lucasagomes | lol | 12:47 |
| dtantsur | I suspect it's related to optimal layout | 12:47 |
| lucasagomes | omg | 12:48 |
| lucasagomes | yeah | 12:48 |
| jroll | lucasagomes: oh, right, that's the hard part :P | 12:48 |
| lucasagomes | jroll, yup | 12:48 |
| lucasagomes | dtantsur, yeah maybe we should just get rid of the optmial alignment then... | 12:48 |
| dtantsur | lucasagomes, what I still see is: root + eph + swap + 1 MiB == total_size - 1 | 12:48 |
| dtantsur | lucasagomes, I would not call it a problem, it's actually better :) | 12:49 |
| dtantsur | but it's less predictable | 12:49 |
| dtantsur | lucasagomes, I don't know what to do. The only suggestion from me is to check `root + eph + swap < size - 1 MiB` | 12:50 |
| dtantsur | because that's how I see it works | 12:50 |
| lucasagomes | yeah | 12:50 |
| dtantsur | and than on Ubuntu the "magic" 1 MiB will be "eaten" by parted, on Fedora - not, but that's probably gonna work | 12:50 |
| lucasagomes | wondering if that's have something to do with the version of parted | 12:51 |
| dtantsur | do we already have some checks somewhere? | 12:51 |
| dtantsur | I need someone with Ubuntu Precise to conduct experiments :) | 12:51 |
| *** rloo has joined #openstack-ironic | 12:55 | |
| *** jdob has joined #openstack-ironic | 12:57 | |
| lucasagomes | dtantsur, :( I don't | 12:57 |
| lucasagomes | although I think I have a precise vm because I was testing something there | 12:57 |
| NobodyCam | good morning Iroinic | 12:57 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Move PXE instance level parameters to instance_info https://review.openstack.org/94855 | 12:57 |
| lucasagomes | jroll, ^ | 12:58 |
| lucasagomes | NobodyCam, morning | 12:58 |
| NobodyCam | morning lucasagomes :) | 12:58 |
| jroll | lucasagomes: sweet :) | 12:59 |
| NobodyCam | morning jroll :) | 12:59 |
| *** _dguerri has quit IRC | 12:59 | |
| NobodyCam | and dtantsur morning | 13:00 |
| jroll | hey NobodyCam :) | 13:00 |
| jroll | I'm on vacation so pretend I'm not here ;) | 13:00 |
| lucasagomes | lol | 13:00 |
| NobodyCam | ahh LOL | 13:00 |
| *** dguerri has joined #openstack-ironic | 13:01 | |
| dtantsur | morning NobodyCam :) we're still having some dirty fun with partitioning :D | 13:01 |
| *** jbjohnso has joined #openstack-ironic | 13:01 | |
| NobodyCam | lol | 13:02 |
| dtantsur | btw morning jroll! I hope we're a nice company for vacation :) | 13:02 |
| jroll | haha | 13:02 |
| jroll | morning dtantsur :) | 13:02 |
| *** rloo has quit IRC | 13:02 | |
| *** rloo has joined #openstack-ironic | 13:03 | |
| dtantsur | lucasagomes, you may want to sync your efforts with https://review.openstack.org/#/c/90233/ | 13:05 |
| dtantsur | you have some things overlapping, I guess | 13:05 |
| lucasagomes | dtantsur, oh true | 13:05 |
| lucasagomes | yeah... the reason why I didn't do that at the first place was because that patch was broken | 13:06 |
| lucasagomes | so I wouldn't be able to test deploying an instance | 13:06 |
| lucasagomes | but there's a new review up right now | 13:06 |
| lucasagomes | I will give it another go | 13:06 |
| jroll | lucasagomes: hey, I'm wondering if deploy_kernel/ramdisk should stay in driver_info | 13:09 |
| jroll | they're more flavor-specific, yes? | 13:09 |
| *** athomas has quit IRC | 13:09 | |
| lucasagomes | jroll, yeah, they r part of the flavor extra_specs | 13:09 |
| jroll | right, not really instance-specific | 13:09 |
| jroll | I'm fine with it either way, but I think deva and I may have spoke about it | 13:10 |
| jroll | I'm assuming there's a reason those are not on the list of fields to move in the blueprint | 13:10 |
| *** matty_dubs|gone is now known as matty_dubs | 13:10 | |
| lucasagomes | right... yeah not that instance related indeed | 13:10 |
| lucasagomes | oh true yeah they r not | 13:10 |
| lucasagomes | I will move then back to the driver_info | 13:10 |
| lucasagomes | mark as WIP :) | 13:11 |
| *** rloo has quit IRC | 13:11 | |
| * lucasagomes mark as WIP* | 13:11 | |
| jroll | cool, thanks | 13:11 |
| *** rloo has joined #openstack-ironic | 13:11 | |
| jroll | I'll put a note on the review | 13:11 |
| lucasagomes | jroll, right thanks for pointing that out | 13:11 |
| jroll | np | 13:12 |
| *** rloo has quit IRC | 13:13 | |
| *** athomas has joined #openstack-ironic | 13:13 | |
| *** rloo has joined #openstack-ironic | 13:13 | |
| *** linggao has joined #openstack-ironic | 13:14 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/python-ironicclient: node-show to show the instance_info field https://review.openstack.org/94860 | 13:15 |
| lucasagomes | ^ helps to test it | 13:15 |
| jroll | ha, nice | 13:15 |
| openstackgerrit | Aleksandr Gordeev proposed a change to openstack/ironic-python-agent: Introduce MainlineHardwareManager https://review.openstack.org/92847 | 13:15 |
| jroll | I'm a fan of curl, so I tend to forget about the client :P | 13:15 |
| dtantsur | lucasagomes, do we really have any kind of checks on total size of root + swap + eph? I can't find Oo | 13:16 |
| lucasagomes | jroll, hah yeah | 13:16 |
| lucasagomes | dtantsur, hmm not that I recall | 13:17 |
| lucasagomes | dtantsur, I think the only sanity check we have is if the image fits into the root partition | 13:17 |
| dtantsur | lucasagomes, should I file a bug? I think we need to be clear to user on what we expect to fit into his disk :) | 13:18 |
| lucasagomes | dtantsur, sure yeah definitely | 13:18 |
| lucasagomes | dtantsur, afair, we need to fix some misunderstands with nova as well | 13:18 |
| lucasagomes | for e.g | 13:18 |
| lucasagomes | we have properties/local_gb, which represents the size of the _whole_ disk for ironic | 13:19 |
| lucasagomes | but nova uses it as the size of the root_partition, independent of the ephemeral size or swap size | 13:19 |
| lucasagomes | so what's passed for us is, size of the whole disk + eph + swap | 13:19 |
| dtantsur | O_O | 13:20 |
| lucasagomes | I think that is one reason why we don't have a sanity check within ironic | 13:20 |
| lucasagomes | dtantsur, yeah, there's something odd around that area | 13:20 |
| lucasagomes | is a bit grey area | 13:20 |
| lucasagomes | dtantsur, we may want to calculate it on our driver tho | 13:20 |
| dtantsur | lucasagomes, could you add your consideration about Nova here: https://bugs.launchpad.net/ironic/+bug/1322179 | 13:27 |
| dtantsur | just so that we have everything in one place :) | 13:27 |
| lucasagomes | dtantsur, sure, I gotta do some small investigation to see if it's still like that as well, I will copy the IRC for now as a note there | 13:31 |
| dtantsur | thnx! | 13:31 |
| NobodyCam | rloo: Shrews: I'm finding many conflicting ideas on how best to remove gpt tables | 13:31 |
| rloo | NobodyCam: :-( | 13:32 |
| NobodyCam | but i do think -35 is wrong | 13:32 |
| rloo | NobodyCam: from what I could tell, it seemed like the 'older' calculations from that link you gave me, should work. | 13:32 |
| dtantsur | NobodyCam, could you also have a look at https://bugs.launchpad.net/ironic/+bug/1322179 | 13:32 |
| rloo | NobodyCam: I think the 35 worked in that calculation, cuz they calculated some location in the disk, and wiped everything after that location. | 13:33 |
| Shrews | NobodyCam: rloo: I don't think wiping too much at the end of the disk is too big of a deal since the root partition is always last. It's just sloppy to not be exact. It's wiping the data at the front that's scary to me. | 13:34 |
| lucasagomes | NobodyCam, what about just removing the labels with tunefs instead of ereasing parts of the disk? | 13:34 |
| lucasagomes | I think that might be safer and quicker | 13:34 |
| NobodyCam | lucasagomes: that was where I started | 13:34 |
| lucasagomes | NobodyCam, right | 13:35 |
| rloo | Shrews: yes, sloppy is the other issue :-( | 13:35 |
| lucasagomes | NobodyCam, gotta see the comments on that patch | 13:35 |
| NobodyCam | lucasagomes: see adam's comment on rev 1: https://review.openstack.org/#/c/93133/1/ironic/drivers/modules/deploy_utils.py | 13:35 |
| rloo | Shrews: also, NobodyCam's version didn't wipe out the rest of the disk, only 4096 bytes from the seek position :-( | 13:35 |
| lucasagomes | btw people with +2 power: https://review.openstack.org/#/c/91062/ | 13:35 |
| lucasagomes | NobodyCam, will do | 13:36 |
| rloo | lucasagomes: I already +2'd it :-) | 13:36 |
| lucasagomes | rloo, yup, thanks :) | 13:36 |
| * lucasagomes wants to get that series merged cause it's hard to keep rebasing it | 13:36 | |
| rloo | where's max? He only +1 | 13:36 |
| rloo | lucasagomes: yeah, I want those merged too! | 13:37 |
| lucasagomes | max seems to be offline :( | 13:37 |
| *** r-daneel has joined #openstack-ironic | 13:38 | |
| Shrews | NobodyCam: rloo: Are we sure sgdisk will not work in our case? The bug that lucasagomes pointed out with it seems to indicate that only fdisk is affected by the bug, but parted may still work? | 13:38 |
| lucasagomes | Shrews, which canse? zap-all? | 13:38 |
| Shrews | lucasagomes: yes | 13:38 |
| rloo | Shrews: no idea. I only saw the comment about the bug. Didn't look into sgdisk at all. | 13:38 |
| lucasagomes | Shrews, I didn't test it | 13:39 |
| lucasagomes | Shrews, but it might worth to give it a go | 13:39 |
| Shrews | so maybe we should not count that out yet until we try it :) | 13:39 |
| lucasagomes | cause that options is meant to do exactly what is being done in that patch | 13:39 |
| lucasagomes | that option* | 13:40 |
| lucasagomes | remove the MBR and GPT data structures from the disk | 13:40 |
| Shrews | yup | 13:40 |
| lucasagomes | Shrews, do you have a ubuntu precise handy there? can you give it a go? | 13:40 |
| Shrews | lucasagomes: unfortunately no | 13:41 |
| lucasagomes | :( | 13:41 |
| Shrews | but vm's are easily spun up | 13:41 |
| lucasagomes | true | 13:42 |
| Shrews | brb. quick walk before heat/humidity prevents it | 13:42 |
| openstackgerrit | Chris Krelle proposed a change to openstack/ironic: Workaround for conflicting volume labels https://review.openstack.org/93133 | 13:42 |
| NobodyCam | rloo: Shrews: lucasagomes: ^^^^^ | 13:42 |
| NobodyCam | lol raining buckets here in NE | 13:42 |
| rloo | NobodyCam: 1 should work. BUT several questions. | 13:45 |
| rloo | NobodyCam: Why do we want to zap 4096 bytes? | 13:46 |
| rloo | NobodyCam: why 4096? I think we should add a comment about that. | 13:46 |
| rloo | NobodyCam: second thing. for the end of the disk, because of possible truncation, using count=1 may not guarantee that the end of the disk is wiped. | 13:47 |
| NobodyCam | 4096 was lifeless' sugestion | 13:48 |
| rloo | NobodyCam: ok, will ask lifeless why he suggested it. | 13:49 |
| NobodyCam | rloo: from what I read this morning 1024 (512 *2) would also work | 13:49 |
| rloo | NobodyCam: also, if it is MBR, we're erasing the end of the disk (in case it is GPT) and that should be OK - we should add a comment about that. | 13:50 |
| NobodyCam | I can do that | 13:50 |
| rloo | NobodyCam: yeah, I think 512 *2 should work, since for MBR, only need 512, and for GPT, 1024 = pseudo-whatever-MBR+GPTheader. | 13:51 |
| NobodyCam | this seems the "most" common approch | 13:51 |
| NobodyCam | https://snipt.net/raw/b37c7950d46cc83ec2fd1ec57138fac6/?nice | 13:51 |
| rloo | NobodyCam: yeah. I was also wondering what the code for sgdisk did but i was too lazy to look. | 13:52 |
| rloo | NobodyCam: if lines 8&9 work, I'd go with that because it seems the most 'precise'. | 13:53 |
| NobodyCam | rloo: https://bugs.launchpad.net/ubuntu/+source/gdisk/+bug/1303903 | 13:53 |
| rloo | NobodyCam: also, blockdev gives number in 512 chunks, so the math is cleaner | 13:54 |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Correct console behavior when open session fails https://review.openstack.org/94869 | 13:54 |
| NobodyCam | Shrews: lucasagomes: how does that sound. | 13:55 |
| NobodyCam | brb | 13:57 |
| *** blamar has quit IRC | 13:58 | |
| rloo | NobodyCam: interesting. in the sgdisk bug report, they zap 1M bytes from the beginning, and 100x512 bytes from the end. If I understand it. | 14:00 |
| NobodyCam | rloo: in my reading this mornig I have found every thing from 1024 to 16k (and one 8MiB) | 14:02 |
| rloo | NobodyCam: kind of scary. I think when you're unsure, more is better than less. Which is probably why the numbers range. | 14:03 |
| NobodyCam | i'm a less is more type person so I'm going to redo to use 512 *2 | 14:04 |
| *** geertj has quit IRC | 14:05 | |
| Shrews | i'm still a fan of trying sgdisk instead of dd'ing, but i will defer to the cores on this one | 14:06 |
| openstackgerrit | A change was merged to stackforge/pyghmi: Correct console behavior when open session fails https://review.openstack.org/94869 | 14:06 |
| NobodyCam | Shrews: did you see https://bugs.launchpad.net/ubuntu/+source/gdisk/+bug/1303903 | 14:07 |
| Shrews | NobodyCam: yes, but it only affects fdisk, not parted :) | 14:08 |
| lucasagomes | another more elegant way of doing it would be to drop parted, and use fdisk/sgdisk for partitioning... keep the DiskPartitioner interface and change the tool behind it | 14:08 |
| lucasagomes | it was something that lifeless suggested | 14:08 |
| lucasagomes | afair | 14:09 |
| Shrews | lucasagomes: the fdisk/sgdisk combination will not work, according to that bug. so that confuses me | 14:09 |
| lucasagomes | Shrews, I mean use the tools for partitioning | 14:09 |
| Shrews | ah | 14:09 |
| lucasagomes | the reason why we have to remove the labels is because parted doesn't work on ubuntu if two devices | 14:10 |
| lucasagomes | contains a partition with the same label | 14:10 |
| lucasagomes | and one of these partition is mounted | 14:10 |
| lucasagomes | there's a bunch of problems there | 14:10 |
| lucasagomes | first, we should not have two disks with the same label | 14:10 |
| lucasagomes | but that's a problem that involves other stuff, cloud-init expects the root partition to have a specific label | 14:12 |
| lucasagomes | :( | 14:12 |
| lucasagomes | all that sounds like a problem on top of another problem for me :( | 14:12 |
| NobodyCam | lets use dodwipe and be done with it | 14:13 |
| NobodyCam | :-p | 14:13 |
| lucasagomes | but anyway... aparrently, fdisk/sgdisk is more smart and they detect if the device you want to partition is the one mounted or not | 14:13 |
| lucasagomes | lol | 14:13 |
| lucasagomes | NobodyCam, once we wipe the tenant disk after tear down | 14:14 |
| lucasagomes | this problem will also be gone | 14:14 |
| lucasagomes | I suspect | 14:14 |
| NobodyCam | lucasagomes: nope. because we can not guarantee that we wiped the first time. | 14:15 |
| NobodyCam | ie a new node | 14:15 |
| lucasagomes | ah yeah | 14:15 |
| NobodyCam | we have no idea what was one it | 14:15 |
| NobodyCam | :-p | 14:15 |
| lucasagomes | urgh heh | 14:15 |
| NobodyCam | s/one/on/ | 14:15 |
| NobodyCam | :-p | 14:15 |
| lucasagomes | maybe the fdisk/sgdisk being used to partitioning the disk would be the most elegant solution for now | 14:15 |
| NobodyCam | i'm up for ideas. I just want something that has a chance to land :-p | 14:17 |
| NobodyCam | (oh and actually works) | 14:17 |
| NobodyCam | hehehe :-p | 14:17 |
| openstackgerrit | Chris Krelle proposed a change to openstack/ironic: Workaround for conflicting volume labels https://review.openstack.org/93133 | 14:23 |
| NobodyCam | ^^ that uses 512 * 2 | 14:23 |
| Shrews | ooh, HP is poised to overtake Red Hat as top contributor to Juno, according to Stackalytics. only 2 commits behind. | 14:28 |
| * Shrews -1's all of lucasagomes' reviews | 14:28 | |
| Shrews | :) | 14:28 |
| NobodyCam | lol | 14:28 |
| NobodyCam | bad Shrews bad | 14:28 |
| NobodyCam | lol | 14:28 |
| lucasagomes | Shrews, heh | 14:29 |
| lucasagomes | Shrews, I think it's all bs you know... as no of commits would implies something good | 14:29 |
| lucasagomes | I don't like when RH keep claiming that we are top contributor cause | 14:29 |
| Shrews | lucasagomes: yeah. just injecting some humor into a Thursday :) | 14:30 |
| lucasagomes | man at the moment we are not the one in the top anymore | 14:30 |
| lucasagomes | everyone would start talking about it | 14:30 |
| lucasagomes | Shrews, yeah I know :) | 14:30 |
| openstackgerrit | Ruby Loo proposed a change to openstack/ironic: Implement API to get driver properties https://review.openstack.org/73005 | 14:30 |
| rloo | until stackalytics takes into consideration 0/comments, I'm ignoring it. | 14:31 |
| lucasagomes | heh | 14:32 |
| lucasagomes | yeah, and there's a lot of "sync X from oslo" stuff as well | 14:32 |
| lucasagomes | I mean... this argument about top contri is all bullcrap | 14:32 |
| lucasagomes | people should not be allowed to spread such FUD | 14:32 |
| lucasagomes | not FUD... I mean crap :) | 14:33 |
| rloo | companies will do whatever. shrug. | 14:33 |
| lucasagomes | :( | 14:34 |
| *** geertj has joined #openstack-ironic | 14:34 | |
| *** Mikhail_D_wk has quit IRC | 14:41 | |
| *** athomas has quit IRC | 14:41 | |
| openstackgerrit | Imre Farkas proposed a change to openstack/ironic: Cleanup images after deployment https://review.openstack.org/92340 | 14:44 |
| *** athomas has joined #openstack-ironic | 14:46 | |
| lucasagomes | devananda, ping r u around? re instance_info | 14:54 |
| lucasagomes | NobodyCam, can you please take a look at https://review.openstack.org/#/c/91062? | 14:56 |
| NobodyCam | lucasagomes: sure... if you look at 93133 again.. just passed my rebuild test | 14:57 |
| NobodyCam | hehehe | 14:57 |
| dtantsur | review exchange in action ^^^ | 14:58 |
| dtantsur | :D | 14:58 |
| lucasagomes | hah yeah | 14:58 |
| NobodyCam | hehe | 15:01 |
| *** igor_ has quit IRC | 15:07 | |
| *** igor has joined #openstack-ironic | 15:07 | |
| devananda | lucasagomes: pong | 15:10 |
| lucasagomes | devananda, hey... so I was talking to jroll about the instance_info | 15:10 |
| *** hemna_ has joined #openstack-ironic | 15:10 | |
| lucasagomes | and he mentioned that you guys have talked before and kinda agreed to not have the deploy_{kernel, ramdisk} in the instance_info field | 15:10 |
| devananda | those are really a property of the driver | 15:11 |
| lucasagomes | devananda, I agree with that cause the deploy ramdisk/kernel is not instance related... but at the same time, I kinda like the idea of having a field in the nodes that can be totally discarded after the node tear_down | 15:11 |
| lucasagomes | devananda, yeah | 15:12 |
| *** igor has quit IRC | 15:12 | |
| lucasagomes | devananda, or... do you think they are part of the deployment information? | 15:12 |
| lucasagomes | devananda, I was thinking about not having an instance_info, but having a deploy_info field | 15:12 |
| devananda | lucasagomes: so we need instance_info for information about the requested instance -- both the disk root and swap sizes, as well as user image, ephemeral, etc | 15:12 |
| lucasagomes | where all the info about the deployment would go, and would be discarded after the deploy | 15:12 |
| devananda | hmm | 15:12 |
| lucasagomes | where the driver_info, properties etc... are more like long term information | 15:13 |
| devananda | right | 15:13 |
| lucasagomes | that is enrolled with the driver | 15:13 |
| devananda | we may be saying the same thing :) | 15:13 |
| lucasagomes | devananda, it sounds a bit more tidy to me | 15:13 |
| lucasagomes | devananda, heh yeah | 15:13 |
| devananda | can you sketch out everything you want in this field? | 15:13 |
| devananda | it may be exactly the same list | 15:13 |
| NobodyCam | lucasagomes: code LGTM. I'm going to issue a recheck because check-tripleo-ironic-seed-precise failed. | 15:13 |
| NobodyCam | and Good Morning devananda | 15:14 |
| lucasagomes | devananda, yes... basically the deploy_info would contain all the information required for the deployment, from both flavor and instance, right now this info is: deploy_kernel, deploy_ramdisk, image_source, root_gb, swap_mb, ephemeral_gb, ephemeral_format | 15:14 |
| devananda | lucasagomes: great. so, initially, yes, that's the same list I'd use. But to support different drivers, we will need to move deploy_* to driver_info, and remove it from nova flavor. | 15:15 |
| lucasagomes | devananda, ahh so we are removing it from flavor? | 15:15 |
| devananda | what happens if we use the PXE driver with the IPA deploy_ramdisk? | 15:16 |
| devananda | today it wont work. we can address taht in a few ways | 15:16 |
| lucasagomes | devananda, right I thought it would be 2 diff drivers | 15:17 |
| lucasagomes | PXE and IPA driver | 15:17 |
| devananda | right, it will be | 15:17 |
| devananda | but the flavor != driver today | 15:17 |
| devananda | nova scheduler isn't matching flavor to node driver | 15:17 |
| devananda | it's matching cpu_arch, num cpu, ram, disk | 15:17 |
| lucasagomes | correct | 15:18 |
| lucasagomes | that's why we have the nova ironic driver to set those fields before deploying the node? | 15:18 |
| lucasagomes | the nova driver checks the driver that tne node requires | 15:18 |
| lucasagomes | and get the deploy_ramdisk/kernel from the flavor and set it | 15:18 |
| devananda | well. it still pulls the deploy ramdisk info from the flavor | 15:18 |
| devananda | so if I specify the PXE ramdisk on flavor F, then the scheduler picks a node whose driver is IPA, the nova.virt.ironic driver will copy the PXE ramdisk to that node's driver_info ... and it'll fail | 15:19 |
| lucasagomes | yeah :( gotcha | 15:19 |
| lucasagomes | devananda, now I connected the dots heh | 15:20 |
| lucasagomes | thanks for the explanation | 15:20 |
| devananda | so three solutions there: IPA is compatible with PXE driver; scheduler matches flavor to node.driver; deploy ramdisk is a property of the node, not the flavor | 15:20 |
| devananda | jroll, russell_h: above conversation may be interesting for you guys | 15:20 |
| lucasagomes | so we should not set it via the driver/flavor, it should be part of the info when we enroll the node with a specific driver | 15:20 |
| devananda | lucasagomes: np :) | 15:20 |
| lucasagomes | devananda, cooleo, makes total sense to me now | 15:21 |
| lucasagomes | will modify my patch to cover this scenario | 15:21 |
| devananda | we should, you know, record taht somewhere besides IRC. like maybe a spec? | 15:21 |
| devananda | hehe | 15:21 |
| lucasagomes | devananda, can add to the bp there | 15:21 |
| lucasagomes | will do | 15:21 |
| lucasagomes | devananda, or, it will needed to create a spec for this blueprint? because it's already started | 15:22 |
| devananda | oh - and we have a specs repo now | 15:22 |
| devananda | https://github.com/openstack/baremetal-specs | 15:22 |
| lucasagomes | the instance_info is already there | 15:22 |
| *** geertj has quit IRC | 15:23 | |
| lucasagomes | ok I can create a spec for that | 15:23 |
| *** pbrooko has joined #openstack-ironic | 15:26 | |
| *** martyntaylor has left #openstack-ironic | 15:26 | |
| lucasagomes | devananda, are you working on add the basic structure to the baremetal-specs repo? or I can go ahead and do it? | 15:27 |
| *** igor_ has joined #openstack-ironic | 15:28 | |
| devananda | lucasagomes: i've got it -- just be a few minutes | 15:28 |
| devananda | i'm making coffee | 15:28 |
| lucasagomes | devananda, ok dokie, thank you | 15:28 |
| NobodyCam | lucasagomes: :( 91062 looks like its need another rebase :( | 15:30 |
| *** tteggel has quit IRC | 15:30 | |
| lucasagomes | NobodyCam, urgh damn... ack lemme rebase it | 15:30 |
| *** geertj has joined #openstack-ironic | 15:31 | |
| *** ramineni has quit IRC | 15:31 | |
| NobodyCam | lucasagomes: :-p sorry... rloo: TY | 15:31 |
| rloo | NobodyCam: YW. Sorry for giving you more grief. | 15:32 |
| lucasagomes | NobodyCam, heh not ur fault | 15:32 |
| *** igor_ has quit IRC | 15:32 | |
| *** tteggel has joined #openstack-ironic | 15:33 | |
| NobodyCam | rloo: not at all... that what the review process is for!!! | 15:34 |
| *** lcostantino has joined #openstack-ironic | 15:35 | |
| *** blamar has joined #openstack-ironic | 15:38 | |
| *** igor_ has joined #openstack-ironic | 15:38 | |
| *** uberj_ is now known as uberj | 15:38 | |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Remove 'node' parameter from the Console and Rescue interfaces https://review.openstack.org/91097 | 15:39 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Remove 'node' parameter from the Power interface https://review.openstack.org/91093 | 15:39 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Remove 'node' parameter from the Deploy interface https://review.openstack.org/91062 | 15:39 |
| openstackgerrit | Lucas Alvares Gomes proposed a change to openstack/ironic: Remove 'node' parameter from the validate() methods https://review.openstack.org/91104 | 15:39 |
| lucasagomes | rloo, NobodyCam ^ rebased | 15:39 |
| *** igor__ has joined #openstack-ironic | 15:40 | |
| NobodyCam | lucasagomes: ack.. will watch for jenkins :) | 15:42 |
| lucasagomes | NobodyCam, ta much! | 15:42 |
| lucasagomes | those patches r being hard to keep up-to-date cause they change various files | 15:42 |
| NobodyCam | ya | 15:42 |
| *** dkehn_ has joined #openstack-ironic | 15:42 | |
| *** igor_ has quit IRC | 15:43 | |
| rloo | lucasagomes: yeah. will see if NobodyCam or I get to it second, after jenkins! | 15:43 |
| *** igor__ has quit IRC | 15:44 | |
| *** dkehnx has quit IRC | 15:46 | |
| lucasagomes | rloo, ta much! | 15:46 |
| *** dkehn_ is now known as dkehnx | 15:55 | |
| *** pbrooko has quit IRC | 16:05 | |
| *** rloo has quit IRC | 16:05 | |
| *** rloo has joined #openstack-ironic | 16:06 | |
| NobodyCam | Shrews: got a free minute to take a look at the latest 93133? | 16:06 |
| NobodyCam | devananda: ^^^^ ??? | 16:06 |
| devananda | NobodyCam: later today, yes. I have a conf call in ~20 | 16:08 |
| NobodyCam | ack as do I | 16:09 |
| *** pbrooko has joined #openstack-ironic | 16:10 | |
| *** datajerk has quit IRC | 16:10 | |
| *** dkehn_ has joined #openstack-ironic | 16:12 | |
| *** eghobo has joined #openstack-ironic | 16:13 | |
| *** matty_dubs is now known as matty_dubs|lunch | 16:14 | |
| *** rameshg87 has joined #openstack-ironic | 16:15 | |
| *** dkehnx has quit IRC | 16:16 | |
| *** dkehn_ is now known as dkehnx | 16:16 | |
| Shrews | Nobodycam: will look a bit later. Meeting someone for lunch now | 16:19 |
| NobodyCam | :) TY | 16:21 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/baremetal-specs: Initial commit https://review.openstack.org/94922 | 16:23 |
| russell_h | devananda: lucasagomes: agree completely with where your discussion of deploy_{kernel, ramdisk} wound up | 16:23 |
| *** hemna_ has quit IRC | 16:23 | |
| devananda | russell_h: of those 3 options, which do you guys plan to support? | 16:24 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/baremetal-specs: Add spec for fully asynchronous API https://review.openstack.org/94923 | 16:25 |
| NobodyCam | nice :) | 16:25 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/baremetal-specs: Add spec to improve logging https://review.openstack.org/94924 | 16:25 |
| *** rloo has quit IRC | 16:25 | |
| russell_h | devananda: ah, actually missed that last little bit. I guess my opinion is that there must be a common specification of "instance_InfO | 16:26 |
| russell_h | er, instance_info across all drivers | 16:26 |
| *** rloo has joined #openstack-ironic | 16:26 | |
| russell_h | we don't have a real use case for specifying the other bits yet | 16:26 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/baremetal-specs: Add spec for improving the handling of bootloaders https://review.openstack.org/94925 | 16:26 |
| *** rloo has quit IRC | 16:26 | |
| russell_h | so, anything that we can find a way to expose via the Nova API that isn't support in _every_ Ironic driver, should probably still go into instance_info | 16:27 |
| russell_h | but be documented as not supported by all driivers | 16:27 |
| *** rloo has joined #openstack-ironic | 16:27 | |
| russell_h | for things that are static... I'd say put it on the flavor if possible? | 16:27 |
| russell_h | I'd have to think more on that | 16:27 |
| russell_h | like would you ever have two nodes of the same flavor with a different deploy_ramdisk? | 16:27 |
| *** Rikkol has joined #openstack-ironic | 16:28 | |
| devananda | russell_h: sure - if they use different drivers | 16:28 |
| russell_h | right, but why would you do that? | 16:28 |
| russell_h | I guess its a philosophical thing | 16:28 |
| devananda | russell_h: say an upgrade | 16:28 |
| devananda | like someone migrating from PXE to IPA | 16:28 |
| russell_h | ah, fair | 16:28 |
| russell_h | right | 16:28 |
| russell_h | yeah, so put it on the node | 16:28 |
| devananda | ok - i'm late for this call ... bbiah | 16:29 |
| russell_h | now.. how it gets there | 16:29 |
| * russell_h too | 16:29 | |
| *** ndipanov has quit IRC | 16:29 | |
| * NobodyCam book marks https://github.com/openstack/baremetal-specs :) | 16:29 | |
| *** Rikkol has left #openstack-ironic | 16:33 | |
| dtantsur | devananda, hi! Will have time to have a short look on https://bugs.launchpad.net/ironic/+bug/1322179 ? It's an outcome of our (with lucasagomes) 2-day-long discussion of current partitioning | 16:35 |
| lucasagomes | russell_h, I will see if I can come up with a spec for the instance_info work | 16:35 |
| *** blamar has quit IRC | 16:40 | |
| *** igor__ has joined #openstack-ironic | 16:40 | |
| openstackgerrit | Andreas Jaeger proposed a change to openstack/ironic: Remove all mostly untranslated PO files https://review.openstack.org/94932 | 16:45 |
| *** coolsvap|afk is now known as coolsvap | 16:45 | |
| *** igor__ has quit IRC | 16:45 | |
| *** ramineni has joined #openstack-ironic | 16:50 | |
| *** harlowja_away is now known as harlowja_ | 16:54 | |
| lucasagomes | alright I'm heading home... have a good night everybody | 16:54 |
| *** lucasagomes has quit IRC | 16:55 | |
| *** Ajaeger1 has joined #openstack-ironic | 16:56 | |
| NobodyCam | have a good night lucas | 16:56 |
| openstackgerrit | Dan Prince proposed a change to openstack/ironic: Port iBoot PDU driver from Nova https://review.openstack.org/50977 | 17:02 |
| *** jistr has quit IRC | 17:05 | |
| *** derekh_ is now known as derekh_afk | 17:06 | |
| *** rloo has quit IRC | 17:07 | |
| *** rloo has joined #openstack-ironic | 17:07 | |
| *** jcoufal has quit IRC | 17:11 | |
| *** matty_dubs|lunch is now known as matty_dubs | 17:13 | |
| *** BadCub01_ has joined #openstack-ironic | 17:19 | |
| *** blamar has joined #openstack-ironic | 17:25 | |
| *** geertj has quit IRC | 17:28 | |
| Ajaeger1 | I'd like to have some review of https://review.openstack.org/94932, please - this will significantly cut down the size and frequency of the daily translation import since you then will only have files that are mostly translated. | 17:30 |
| *** athomas has quit IRC | 17:30 | |
| *** ramineni1 has joined #openstack-ironic | 17:33 | |
| Ajaeger1 | would be great if the patch goes in before the next run of "Imported Translations from Transifex" | 17:33 |
| *** ramineni1 has left #openstack-ironic | 17:34 | |
| *** ramineni has quit IRC | 17:35 | |
| *** rameshg87 has left #openstack-ironic | 17:36 | |
| devananda | Ajaeger1: hi! why 75%? | 17:36 |
| *** igordcard has joined #openstack-ironic | 17:38 | |
| Ajaeger1 | devananda: that was the outcome of the discussion on the i18n list, see this thread: | 17:39 |
| devananda | reading now | 17:39 |
| openstackgerrit | Jarrod Johnson proposed a change to stackforge/pyghmi: Correct pool usage count on failed login https://review.openstack.org/94948 | 17:39 |
| Ajaeger1 | http://lists.openstack.org/pipermail/openstack-i18n/2014-May/000594.html | 17:39 |
| Ajaeger1 | 75 % is now done in the infra job that downloads the packages, see: https://review.openstack.org/92997 | 17:40 |
| devananda | Ajaeger1: cool. thanks! | 17:41 |
| Ajaeger1 | devananda: 75 % was the value that was considered best by all. We didn't want to include files that are not at all translated | 17:41 |
| Ajaeger1 | devananda: here're the stats for ironic: https://www.transifex.com/projects/p/ironic/resource/ironic-translations/ | 17:41 |
| *** igor_ has joined #openstack-ironic | 17:42 | |
| devananda | yea, I generally agree that un- or poorly-translated langauges shouldn't be included, and since 75% is what i18n team and infra agreed to, it's fine with me | 17:42 |
| Ajaeger1 | ok, I stop arguing ;) | 17:42 |
| devananda | yep. already looked at all that and +2'd the patch | 17:42 |
| Ajaeger1 | devananda: thanks a lot. | 17:42 |
| Ajaeger1 | Is any other core around that would like to review it as well? | 17:43 |
| devananda | lucas has done some work with our translations before, so I think he'll want to see it -- i just tagged him on the review, but he's not onlien atm | 17:44 |
| Ajaeger1 | devananda: ok, thanks! | 17:44 |
| devananda | Ajaeger1: for reference, is this new 75% limit documented anywhere, like on a wiki? | 17:45 |
| Ajaeger1 | devananda: not yet;) I'll do this later today - good idea | 17:45 |
| devananda | :) | 17:45 |
| openstackgerrit | A change was merged to stackforge/pyghmi: Correct pool usage count on failed login https://review.openstack.org/94948 | 17:45 |
| *** igor_ has quit IRC | 17:46 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/baremetal-specs: Initial commit https://review.openstack.org/94922 | 17:50 |
| *** geertj has joined #openstack-ironic | 17:50 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/baremetal-specs: Add spec for fully asynchronous API https://review.openstack.org/94923 | 17:52 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/baremetal-specs: Add spec to improve logging https://review.openstack.org/94924 | 17:53 |
| Ajaeger1 | devananda: what about https://wiki.openstack.org/wiki/Translations#Re-incorporating_Translations ? | 17:59 |
| devananda | looks good | 18:00 |
| devananda | Ajaeger1: actually - one thing | 18:00 |
| devananda | in "Only files that have a translation rate of 75 or more strings are downloaded" | 18:00 |
| devananda | I think you mean "a rate of 75% or more translated strings" | 18:01 |
| *** pelix has quit IRC | 18:02 | |
| Ajaeger1 | devananda: indeed, will fix. | 18:02 |
| *** zackf has joined #openstack-ironic | 18:03 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/baremetal-specs: Initial commit https://review.openstack.org/94922 | 18:06 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/baremetal-specs: Initial commit https://review.openstack.org/94922 | 18:07 |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/baremetal-specs: Add spec for fully asynchronous API https://review.openstack.org/94923 | 18:07 |
| devananda | anyone around and want to review the spec template? it can be seen rendered here: http://docs-draft.openstack.org/22/94922/2/check/gate-baremetal-specs-docs/390fe55/doc/build/html/specs/template.html | 18:11 |
| *** lcostantino has quit IRC | 18:13 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/baremetal-specs: Initial commit https://review.openstack.org/94922 | 18:16 |
| *** gilliard has quit IRC | 18:17 | |
| *** jcoufal has joined #openstack-ironic | 18:19 | |
| *** lcostantino has joined #openstack-ironic | 18:23 | |
| *** Ajaeger1 has left #openstack-ironic | 18:25 | |
| openstackgerrit | A change was merged to openstack/ironic: Remove all mostly untranslated PO files https://review.openstack.org/94932 | 18:26 |
| *** jcoufal has quit IRC | 18:29 | |
| *** jcoufal has joined #openstack-ironic | 18:30 | |
| *** gilliard has joined #openstack-ironic | 18:36 | |
| *** igor_ has joined #openstack-ironic | 18:42 | |
| *** igor_ has quit IRC | 18:47 | |
| *** jcoufal has quit IRC | 18:55 | |
| *** igordcard has quit IRC | 19:19 | |
| *** geertj has quit IRC | 19:21 | |
| *** geertj has joined #openstack-ironic | 19:22 | |
| *** GheRivero has quit IRC | 19:23 | |
| *** GheRivero has joined #openstack-ironic | 19:30 | |
| NobodyCam | brb | 19:42 |
| *** igor_ has joined #openstack-ironic | 19:43 | |
| *** geertj has quit IRC | 19:47 | |
| *** pbrooko has quit IRC | 19:47 | |
| *** igor_ has quit IRC | 19:47 | |
| devananda | NobodyCam, rloo: either of you want to review the specs repo template and some initial specs? | 19:54 |
| rloo | devananda: am reviewing the template now. | 19:55 |
| devananda | rloo: awesome, thanks | 19:55 |
| NobodyCam | I looked at it when you had it on your personal github.. has it changed? | 19:56 |
| devananda | a bit | 19:58 |
| *** zackf has left #openstack-ironic | 20:02 | |
| rloo | devananda: do you want grammatical feedback, or general feedback or both? wrt the template I mean. | 20:02 |
| *** pbrooko has joined #openstack-ironic | 20:03 | |
| *** rloo has quit IRC | 20:06 | |
| devananda | rloo both | 20:06 |
| *** rloo has joined #openstack-ironic | 20:06 | |
| *** pbrooko has quit IRC | 20:14 | |
| devananda | rloo: both | 20:21 |
| devananda | running an errand ... bbiah | 20:22 |
| rloo | devananda: yeah, i figured i might as well do both while it was fresh in my mind. | 20:22 |
| Shrews | NobodyCam: did you test --preserve-ephemeral with your change? | 20:22 |
| Shrews | i'm doing so now, fyi | 20:24 |
| *** igor_ has joined #openstack-ironic | 20:25 | |
| *** igor__ has joined #openstack-ironic | 20:27 | |
| *** foexle has quit IRC | 20:29 | |
| *** igor_ has quit IRC | 20:30 | |
| *** pbrooko has joined #openstack-ironic | 20:30 | |
| *** igor__ has quit IRC | 20:32 | |
| NobodyCam | Shrews: yeppers.. I use "nova rebuild --preserve-ephemeral epDeployTest basicNew" for my testing | 20:33 |
| Shrews | NobodyCam: yeah, wfm too | 20:36 |
| NobodyCam | :) | 20:48 |
| *** derekh_afk has quit IRC | 20:48 | |
| NobodyCam | I do test it | 20:48 |
| NobodyCam | s/do/did/ :-p | 20:48 |
| NobodyCam | hehehe | 20:48 |
| *** derekh_ has joined #openstack-ironic | 20:49 | |
| Shrews | tempest is driving me freaking nuts | 20:52 |
| Shrews | i think banging my head with a hammer would be more fun and less painful | 20:52 |
| *** lcostantino has left #openstack-ironic | 20:53 | |
| *** praefect_ has joined #openstack-ironic | 20:56 | |
| *** linggao has quit IRC | 20:59 | |
| NobodyCam | :( | 21:01 |
| *** jdob has quit IRC | 21:03 | |
| *** pbrooko has quit IRC | 21:04 | |
| NobodyCam | devananda: question. on the template | 21:04 |
| *** jbjohnso has quit IRC | 21:06 | |
| *** boris-42 has quit IRC | 21:06 | |
| NobodyCam | line 16 has "...that operators can understand" would "reviewers" be a better fit then "operators" as i do not expect many opertiors to be reading our sepc repo?? | 21:07 |
| *** coolsvap is now known as coolsvap|afk | 21:08 | |
| *** boris-42 has joined #openstack-ironic | 21:08 | |
| *** matty_dubs is now known as matty_dubs|gone | 21:13 | |
| *** mrda_away is now known as mrda | 21:19 | |
| mrda | Morning ironic! | 21:19 |
| Shrews | mrda: hey hey | 21:19 |
| NobodyCam | morning mrda | 21:20 |
| *** igor_ has joined #openstack-ironic | 21:28 | |
| mrda | \o | 21:29 |
| *** igor_ has quit IRC | 21:32 | |
| * devananda is back | 21:34 | |
| devananda | g'morning, mrda ! | 21:34 |
| devananda | NobodyCam: no - it should say "operators" because: they will probably be reading this; it'll form the basis for blueprints, whcih they read; it'll also end up in release notes with they definitely read. | 21:36 |
| NobodyCam | devananda: ack, added some comments | 21:36 |
| NobodyCam | :) | 21:36 |
| devananda | thanks | 21:37 |
| mrda | hey devananda. Just FYI, I might take another pass over disabling XML in the i-api today. | 21:37 |
| NobodyCam | and wow thats supper detailed | 21:38 |
| devananda | mrda: fantastic! | 21:38 |
| devananda | NobodyCam: yep :) | 21:38 |
| mrda | I gave up (knowing I wouldn't make I) but I'll circle back now | 21:38 |
| *** ifarkas has quit IRC | 21:47 | |
| NobodyCam | ok almost 5 o'clock here... time to remove a tick from one of the kids... happy happy joy joy | 21:59 |
| *** Penick has joined #openstack-ironic | 22:00 | |
| devananda | ewwwww | 22:06 |
| NobodyCam | came off pritty easy :) | 22:06 |
| NobodyCam | icky looking tihng | 22:06 |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic: Factoring out PXE and TFTP functions https://review.openstack.org/90233 | 22:15 |
| openstackgerrit | Josh Gachnang proposed a change to openstack/ironic: Adding a reference driver for the agent https://review.openstack.org/84795 | 22:15 |
| *** boris-42 has quit IRC | 22:17 | |
| devananda | rloo: have you seen what Nova has done? https://github.com/openstack/nova-specs/blob/master/specs/template.rst | 22:19 |
| *** boris-42 has joined #openstack-ironic | 22:20 | |
| devananda | rloo: not that "its right cause someone else did it" is a valid position, but I do think we should be somewhat consistent and other projects are following that template as well | 22:20 |
| mrda | I actually like that .rst | 22:21 |
| mrda | it's quite intuitive and helpful | 22:22 |
| *** BadCub01_ has quit IRC | 22:24 | |
| *** igor_ has joined #openstack-ironic | 22:28 | |
| *** igor_ has quit IRC | 22:33 | |
| *** praefect_ has quit IRC | 22:38 | |
| *** Penick has quit IRC | 22:39 | |
| openstackgerrit | Devananda van der Veen proposed a change to openstack/baremetal-specs: Initial commit https://review.openstack.org/94922 | 22:51 |
| *** radsy has joined #openstack-ironic | 22:56 | |
| *** radsy has joined #openstack-ironic | 22:56 | |
| *** blamar has quit IRC | 22:58 | |
| devananda | NobodyCam, rloo: updated with your feedback, thanks! | 23:01 |
| NobodyCam | devananda: Thank you awesome, thought pervoking way to submit a blue print | 23:05 |
| NobodyCam | provoking maybe even | 23:06 |
| NobodyCam | lol | 23:06 |
| *** derekh_ has quit IRC | 23:12 | |
| *** blamar has joined #openstack-ironic | 23:13 | |
| *** stevebaker has joined #openstack-ironic | 23:17 | |
| stevebaker | devananda: hey | 23:17 |
| *** openstackgerrit has quit IRC | 23:19 | |
| *** eguz has joined #openstack-ironic | 23:20 | |
| *** eguz has quit IRC | 23:20 | |
| *** openstackgerrit has joined #openstack-ironic | 23:21 | |
| devananda | stevebaker: hi | 23:21 |
| stevebaker | devananda: I'm just looking at your comment on https://review.openstack.org/#/c/92257/1 | 23:22 |
| stevebaker | devananda: I'm guessing the git-clone provided enough delay to allow ironic-api to start before n-cpu | 23:23 |
| *** eghobo has quit IRC | 23:24 | |
| devananda | stevebaker: is there a deterministic order to service startups? I know some of that was refactored | 23:24 |
| devananda | but i haven't followed all the changes in a while | 23:27 |
| stevebaker | I'm not sure. As an aside, SpamapS is in favour of delay-and-retry if $thing isn't there yet. Fixes of this type have gone into nova already, doing the same for ironic-api might be best | 23:28 |
| devananda | we already have delay-and-=retry in there | 23:28 |
| devananda | it retried 5 times befoer giving up | 23:28 |
| stevebaker | retry harder! | 23:28 |
| devananda | :p | 23:28 |
| SpamapS | I'm in favor of what? :) | 23:29 |
| *** igor_ has joined #openstack-ironic | 23:29 | |
| stevebaker | SpamapS: handling service dependencies by just waiting for the service to appear | 23:29 |
| devananda | oh | 23:29 |
| devananda | so there's another fix for this | 23:29 |
| devananda | that we landed bu tit looks like it's not being used | 23:29 |
| SpamapS | ah yes, do that | 23:29 |
| SpamapS | converge | 23:29 |
| SpamapS | do not workflow | 23:30 |
| devananda | errors: http://logs.openstack.org/57/92257/1/check/check-tempest-dsvm-virtual-ironic-nv/9ab97b4/logs/screen-n-cpu.txt.gz?level=WARNING | 23:30 |
| devananda | fix: https://github.com/openstack/ironic/blob/master/ironic/nova/compute/manager.py | 23:30 |
| devananda | now, why is devstack's nova not using that? ... | 23:30 |
| * devananda looks | 23:30 | |
| devananda | ah | 23:31 |
| devananda | because an out-of-tree compute manager is not supported by nova | 23:31 |
| devananda | we need to actually COPY that file into nova's tree to fix this | 23:31 |
| devananda | or land it in nova | 23:31 |
| devananda | heh | 23:31 |
| * devananda moves to -nova | 23:32 | |
| stevebaker | my work here is done | 23:32 |
| *** igor_ has quit IRC | 23:34 | |
| rloo | devananda: and here I was, impressed that you wrote up that template on your own :-) Not sure which of my comments you were referring to, but you got my first impressions. I leave it for you to decide what to do. (If there's room for improvement, we should continue to improve...) | 23:38 |
| devananda | rloo: I appreciate your vote of confidence in my word-smithing skills :) | 23:38 |
| rloo | devananda: definitely. I'll vote for you! | 23:39 |
| devananda | rloo: you caught some good things. and we can continue to improve if necessary. thanks! | 23:40 |
| devananda | if there's no significant objections left, I'll land it | 23:40 |
| rloo | devananda: I don't have time to review it again, just trying to wrap up some stuff and am gone til next week. the main stuff is there so I say land it. | 23:41 |
| devananda | cool | 23:41 |
| devananda | NobodyCam: ? | 23:41 |
| *** radsy has quit IRC | 23:49 | |
| *** r-daneel has quit IRC | 23:50 | |
| *** radsy has joined #openstack-ironic | 23:50 | |
| *** jgrimm has quit IRC | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!