*** harlowja has quit IRC | 00:02 | |
*** harlowja has joined #openstack-ironic | 00:02 | |
*** HenryG has quit IRC | 00:04 | |
*** achanda has quit IRC | 00:08 | |
*** shikui_ has joined #openstack-ironic | 00:16 | |
* NobodyCam must be far to tired.. | 00:17 | |
*** Kui has quit IRC | 00:19 | |
*** achanda has joined #openstack-ironic | 00:20 | |
*** foexle has joined #openstack-ironic | 00:29 | |
*** achanda has quit IRC | 00:34 | |
openstackgerrit | A change was merged to openstack/ironic: Adds instructions for deploying instances on real hardware https://review.openstack.org/118614 | 00:40 |
---|---|---|
dlaube | Can anyone shed some light on why "disk-image-create ubuntu stackuser -a i386 -o LocalTestImage" is throwing this failure -> http://pastie.org/private/qmcoiu9mengit8n4zktsw ? | 00:45 |
Haomeng|2 | dlaube: hi | 00:45 |
Haomeng|2 | your url can not be accessed now - This webpage is not available | 00:46 |
Haomeng|2 | dlaube: I use http proxy, now it works | 00:46 |
*** rwsu has quit IRC | 00:47 | |
Haomeng|2 | dlaube: what linux are you runing | 00:47 |
dlaube | ubuntu 14.04 | 00:48 |
Haomeng|2 | dlaube: notice "qemu-img: command not found" | 00:48 |
dlaube | Ubuntu 14.04.1 LTS | 00:48 |
Haomeng|2 | did you install qemu-img | 00:48 |
dlaube | ahh, let me try that and run again | 00:48 |
Haomeng|2 | ok, good luck | 00:48 |
*** foexle has quit IRC | 00:48 | |
dlaube | I was so caught up with the loop device error | 00:48 |
Haomeng|2 | line 25: /usr/local/bin/../share/diskimage-builder/lib/img-functions: line 117: qemu-img: command not found | 00:49 |
Haomeng|2 | this should be first error | 00:49 |
Haomeng|2 | maybe it is root cause | 00:49 |
dlaube | I've got it installed now | 00:49 |
Haomeng|2 | will be back for a while | 00:49 |
dlaube | rerunning the build | 00:50 |
Haomeng|2 | ok | 00:50 |
Haomeng|2 | good luck | 00:50 |
dlaube | thank you | 00:50 |
Haomeng|2 | welcome, any time:) | 00:50 |
dlaube | --------------------- END PROFILING --------------------- | 00:51 |
dlaube | Image file LocalTestImage.qcow2 created... | 00:51 |
dlaube | NICE! | 00:51 |
*** yongli has joined #openstack-ironic | 01:00 | |
*** marcoemorais1 has quit IRC | 01:05 | |
*** marcoemorais has joined #openstack-ironic | 01:06 | |
*** Haomeng has joined #openstack-ironic | 01:24 | |
*** Haomeng|2 has quit IRC | 01:25 | |
*** marcoemorais has quit IRC | 01:26 | |
*** nosnos has joined #openstack-ironic | 01:39 | |
Haomeng | dlaube: great, qemu-img is needed for image format conversion | 01:43 |
*** rainya has joined #openstack-ironic | 01:47 | |
*** rainya has quit IRC | 02:01 | |
*** rloo has quit IRC | 02:18 | |
*** lazy_prince has quit IRC | 02:23 | |
*** HenryG has joined #openstack-ironic | 02:26 | |
*** dlaube has quit IRC | 02:29 | |
*** dlaube has joined #openstack-ironic | 02:33 | |
*** killer_prince has joined #openstack-ironic | 02:33 | |
*** killer_prince is now known as lazy_prince | 02:33 | |
*** harlowja is now known as harlowja_away | 02:37 | |
*** achanda has joined #openstack-ironic | 02:42 | |
*** achanda has quit IRC | 02:58 | |
*** openstackgerrit has quit IRC | 02:58 | |
*** achanda has joined #openstack-ironic | 02:59 | |
*** kongfy has joined #openstack-ironic | 03:00 | |
*** kongfy has quit IRC | 03:03 | |
*** Poornima has joined #openstack-ironic | 03:04 | |
*** kongfy has joined #openstack-ironic | 03:05 | |
*** kongfy has left #openstack-ironic | 03:05 | |
*** kongfy has joined #openstack-ironic | 03:06 | |
*** kongfy has quit IRC | 03:08 | |
*** harlowja_away has quit IRC | 03:12 | |
*** kongfy has joined #openstack-ironic | 03:16 | |
*** kongfy has left #openstack-ironic | 03:16 | |
*** kongfy has joined #openstack-ironic | 03:17 | |
*** dlaube has quit IRC | 03:25 | |
*** nosnos has quit IRC | 03:26 | |
*** nosnos has joined #openstack-ironic | 03:26 | |
*** nosnos has quit IRC | 03:30 | |
*** antonym has quit IRC | 03:31 | |
*** antonym has joined #openstack-ironic | 03:33 | |
*** andreykurilin_ has joined #openstack-ironic | 03:53 | |
*** pcrews has quit IRC | 03:55 | |
*** dlaube has joined #openstack-ironic | 04:07 | |
*** andreykurilin_ has quit IRC | 04:10 | |
*** Poornima has quit IRC | 04:15 | |
*** nosnos has joined #openstack-ironic | 04:32 | |
*** yongli has quit IRC | 04:36 | |
*** yongli has joined #openstack-ironic | 04:37 | |
*** jgrimm is now known as zz_jgrimm | 04:39 | |
*** achanda has quit IRC | 04:58 | |
*** achanda_ has joined #openstack-ironic | 05:01 | |
*** lazy_prince has quit IRC | 05:12 | |
*** pensu has joined #openstack-ironic | 05:13 | |
*** Poornima has joined #openstack-ironic | 05:13 | |
*** pensu has left #openstack-ironic | 05:15 | |
*** pensu has joined #openstack-ironic | 05:16 | |
*** loki184 has joined #openstack-ironic | 05:18 | |
*** rakesh_hs has joined #openstack-ironic | 05:25 | |
*** rushiagr_away is now known as rushiagr | 05:26 | |
*** lazy_prince has joined #openstack-ironic | 05:30 | |
*** ramineni has joined #openstack-ironic | 05:34 | |
*** takadayuiko_away is now known as takadayuiko | 05:37 | |
*** shikui_ has quit IRC | 05:38 | |
*** praneshp has quit IRC | 05:41 | |
*** loki184 has quit IRC | 05:48 | |
*** loki184 has joined #openstack-ironic | 06:01 | |
*** achanda_ has quit IRC | 06:17 | |
*** d0m0reg00dthing has joined #openstack-ironic | 06:23 | |
*** achanda has joined #openstack-ironic | 06:29 | |
*** Kui has joined #openstack-ironic | 06:32 | |
*** k4n0 has joined #openstack-ironic | 06:34 | |
*** achanda has quit IRC | 06:38 | |
*** jcoufal has joined #openstack-ironic | 06:39 | |
*** d0m0reg00dthing has quit IRC | 06:47 | |
*** d0m0reg00dthing has joined #openstack-ironic | 06:47 | |
*** d0m0reg00dthing has quit IRC | 06:47 | |
*** d0m0reg00dthing has joined #openstack-ironic | 06:48 | |
*** kevinbenton has quit IRC | 06:50 | |
*** d0m0reg00dthing has quit IRC | 06:50 | |
*** d0m0reg00dthing has joined #openstack-ironic | 06:50 | |
*** d0m0reg00dthing has quit IRC | 06:53 | |
*** d0m0reg00dthing has joined #openstack-ironic | 06:53 | |
*** coolsvap|afk is now known as coolsvap | 07:02 | |
*** foexle has joined #openstack-ironic | 07:08 | |
*** pensu has quit IRC | 07:16 | |
*** pradipta_away is now known as pradipta | 07:16 | |
*** dtantsur|afk is now known as dtantsur | 07:22 | |
dtantsur | Morning Ironic | 07:22 |
*** ifarkas has joined #openstack-ironic | 07:24 | |
*** k4n0 has quit IRC | 07:24 | |
*** k4n0 has joined #openstack-ironic | 07:28 | |
*** killer_prince has joined #openstack-ironic | 07:30 | |
Haomeng | dtantsur: morning:) | 07:31 |
dtantsur | :) | 07:31 |
Haomeng | dtantsur: :) | 07:33 |
*** pensu has joined #openstack-ironic | 07:52 | |
*** dtantsur is now known as dtantsur|bbl | 07:54 | |
*** jistr has joined #openstack-ironic | 07:56 | |
*** athomas has joined #openstack-ironic | 08:00 | |
GheRivero | morning all | 08:07 |
Haomeng | GheRivero: morning:) | 08:12 |
*** derekh has joined #openstack-ironic | 08:15 | |
*** lucasagomes has joined #openstack-ironic | 08:27 | |
*** dlaube has quit IRC | 08:52 | |
*** foexle has quit IRC | 08:54 | |
ramineni | lucasagomes, dtantsur : hi | 09:01 |
*** igordcard has joined #openstack-ironic | 09:02 | |
ramineni | have one question , as ironic supports booting up uefi boot mode , want to propose adding get/set boot mode functionalities to management interface | 09:02 |
lucasagomes | ramineni, I think that's good yeah | 09:04 |
ramineni | but ref drivers (ipmi) doesnt support get/set boot mode currently , is it still reasonable to propose to management interface | 09:04 |
lucasagomes | hmm good point :/ | 09:04 |
lucasagomes | maybe starts as a vendor_passthru method for iLO and as other drivers adept we can promote the method to the management interface? | 09:05 |
ramineni | but its good to have it , as other drivers might make use of it :( .. is it not ok to propose , and throw not supported exception for ipmi? | 09:06 |
lucasagomes | oh it's totally OK to propose | 09:13 |
lucasagomes | yeah, sure, let's see what people find about it | 09:13 |
lucasagomes | I personally like it, tho it seems that it's only supported by iLO right now so that's why I think that the vendor interface might make some more sense for it | 09:13 |
lucasagomes | (side note, we should not be afraid of proposing anything! if you think it makes sense you can propose it no problem with that) | 09:14 |
ramineni | lucasagomes: :) | 09:15 |
ramineni | lucasagomes : thanks | 09:16 |
lucasagomes | yw | 09:19 |
*** rameshg87 has joined #openstack-ironic | 09:24 | |
*** killer_prince has quit IRC | 09:31 | |
*** nosnos has quit IRC | 09:39 | |
*** nosnos has joined #openstack-ironic | 09:40 | |
*** killer_prince has joined #openstack-ironic | 09:41 | |
*** takadayuiko has quit IRC | 09:42 | |
*** nosnos has quit IRC | 09:44 | |
*** killer_prince has quit IRC | 09:47 | |
*** lazy_prince has quit IRC | 09:47 | |
*** lazy_prince has joined #openstack-ironic | 09:48 | |
*** andreykurilin_ has joined #openstack-ironic | 09:54 | |
*** pelix has joined #openstack-ironic | 09:57 | |
*** d0m0reg00dthing has quit IRC | 09:58 | |
*** athomas has quit IRC | 10:00 | |
*** kevinbenton has joined #openstack-ironic | 10:03 | |
*** athomas has joined #openstack-ironic | 10:07 | |
*** andreykurilin_ has quit IRC | 10:21 | |
*** igordcard has quit IRC | 10:24 | |
*** igordcard has joined #openstack-ironic | 10:25 | |
*** loki184 has quit IRC | 10:40 | |
*** pradipta is now known as pradipta_away | 10:59 | |
*** rameshg87 has quit IRC | 11:00 | |
*** Haomeng|2 has joined #openstack-ironic | 11:06 | |
*** Haomeng has quit IRC | 11:07 | |
*** ramineni has quit IRC | 11:08 | |
*** dtantsur|bbl is now known as dtantsur | 11:08 | |
dtantsur | lucasagomes, we need to start to cope somehow with the fact that not every driver supports every feature | 11:09 |
dtantsur | lucasagomes, we may want to come up with some capability framework | 11:09 |
lucasagomes | dtantsur, yeah... well it's usually why the vendor_passthru is made for afaict | 11:09 |
lucasagomes | if only one driver supports certain feature it can expose it via vendor_passthru and then as others starts adopting it we can promote that interface | 11:10 |
dtantsur | lucasagomes, not sure. we need to talk more about what driver is and what vendor passthru is. | 11:10 |
lucasagomes | but I understand that vendor_passtru is kinda limited right now | 11:10 |
lucasagomes | yeah | 11:10 |
dtantsur | what if some drivers (e.g. ipmi) will never support some feature? | 11:10 |
lucasagomes | dtantsur, +1 | 11:10 |
lucasagomes | well the abstraction layer should abstract common features of the drivers | 11:11 |
lucasagomes | if it's not supported by one we don't abstract on the standard interfaces | 11:11 |
lucasagomes | but use the vendor_passthru... but yeah we should talk about it | 11:11 |
lucasagomes | I see benefits on both sides | 11:12 |
dtantsur | lucasagomes, see line 29 in https://docs.google.com/spreadsheets/d/1XBKdeDeGfaRYaThjIIoYRwe_zPensECnxsKUuqdoVmQ/edit#gid=0 | 11:18 |
dtantsur | what do you think? | 11:18 |
lucasagomes | looks good, yeah I'm interested on that | 11:19 |
dtantsur | lucasagomes, that aside, I think we should backport https://bugs.launchpad.net/ironic/+bug/1379705 to Juno | 11:19 |
lucasagomes | :/ yeah I thought about that | 11:19 |
lucasagomes | rc-3? | 11:19 |
lucasagomes | gosh if I saw that before :/ | 11:19 |
dtantsur | lucasagomes, well, we have a broken driver. we should backport it. but I don't remember which tag you have to assign :D | 11:21 |
lucasagomes | rc-candidate or something like that | 11:21 |
lucasagomes | yeah will tag it like that | 11:21 |
lucasagomes | juno-rc-potential | 11:21 |
*** lucasagomes is now known as lucas-hungry | 11:33 | |
dtantsur | lucas-hungry, do we have a spec for https://blueprints.launchpad.net/ironic/+spec/new-boot-interface ? I'd read some details actually? Or we decided to go without a spec? | 11:37 |
*** lazy_prince is now known as killer_prince | 11:37 | |
dtantsur | lucas-hungry, I assume you tested https://review.openstack.org/#/c/127454 so we can go ahead and approve? | 11:41 |
*** k4n0 has quit IRC | 11:42 | |
*** Haomeng|2 has quit IRC | 11:43 | |
*** Haomeng|2 has joined #openstack-ironic | 11:45 | |
*** jistr has quit IRC | 12:00 | |
*** kongfy has quit IRC | 12:03 | |
*** dprince has joined #openstack-ironic | 12:08 | |
*** Poornima has quit IRC | 12:12 | |
*** jistr has joined #openstack-ironic | 12:19 | |
*** foexle has joined #openstack-ironic | 12:21 | |
*** lucas-hungry is now known as lucasagomes | 12:25 | |
lucasagomes | dtantsur, I'm writting the spec cause it changes pretty much all drivers | 12:25 |
dtantsur | good | 12:25 |
lucasagomes | so I'm still having problems figuring out what I need to put on the spec | 12:26 |
lucasagomes | e.g | 12:26 |
*** coolsvap is now known as coolsvap|afk | 12:26 | |
lucasagomes | the pxe template had a lot of deploy information in it | 12:26 |
lucasagomes | yeah it's fine to approve the drac thing | 12:26 |
dtantsur | ack | 12:30 |
*** igordcard has quit IRC | 12:40 | |
*** kongfy has joined #openstack-ironic | 12:40 | |
dtantsur | lucasagomes, I know it's not quite Ironic, but could you find some time to do at least quick review of https://github.com/Divius/ironic-discoverd ? I feel strange without people -1ing me :D | 12:40 |
lucasagomes | hah dtantsur sure! | 12:40 |
dtantsur | not sure if github allows commenting, you can just open a bug with everything you spot | 12:41 |
*** igordcard has joined #openstack-ironic | 12:42 | |
lucasagomes | yeah I can't comment indeed, right a bug sounds good | 12:42 |
*** foexle has quit IRC | 12:46 | |
lucasagomes | the more I look at the boot interface :/ the harder I see it is to abstract it | 12:52 |
lucasagomes | cause things like virtual media depends on the deploy method to be able to build the right ISO to be booted and all | 12:52 |
lucasagomes | so they r not completely unrelated | 12:52 |
*** rushiagr is now known as rushiagr_away | 12:55 | |
dtantsur | lucasagomes, I actually don't particularly like our idea of loosely-coupled interfaces. I just does not work. I would prefer much bigger number of reusable functions and solid drivers that just call one or another. | 12:56 |
*** HenryG has quit IRC | 12:57 | |
lucasagomes | dtantsur, yeah I'm seem it as I dig into splitting the deploy :/ it seems to be better to just have common code shared between the drivers for the deploy part than splitting boot and deploy interfaces | 12:58 |
lucasagomes | also for e.g although the agent uses PXE to boot, the PXE part quite diff than the PXE ISCSI driver | 13:00 |
lucasagomes | diff PXE templates and args | 13:00 |
*** igordcard has quit IRC | 13:02 | |
*** jcoufal has quit IRC | 13:07 | |
*** rakesh_hs has quit IRC | 13:09 | |
*** jjohnson2 has joined #openstack-ironic | 13:15 | |
*** jistr is now known as jistr|bbl | 13:16 | |
*** kongfy has quit IRC | 13:23 | |
*** jcoufal has joined #openstack-ironic | 13:25 | |
*** rushiagr_away is now known as rushiagr | 13:44 | |
*** Kui has quit IRC | 13:45 | |
*** jjohnson2 has quit IRC | 13:59 | |
NobodyCam | Goom Morning Ironic....TGIF!!! | 14:05 |
dtantsur | NobodyCam, morning, really TGIF! | 14:07 |
lucasagomes | NobodyCam, morning | 14:07 |
NobodyCam | morning dtantsur and lucasagomes, you all ready for the weekend? | 14:09 |
dtantsur | NobodyCam, looking forward to it :) | 14:09 |
NobodyCam | oh ya :) | 14:10 |
*** jjohnson2 has joined #openstack-ironic | 14:15 | |
lucasagomes | NobodyCam, yeah :) going to the pycon ireland | 14:15 |
lucasagomes | let's see how is it, I've been in 2012 and it was ok | 14:15 |
NobodyCam | oh Nice | 14:15 |
*** openstackgerrit has joined #openstack-ironic | 14:16 | |
yuriyz | morning NobodyCam evening lucasagomes dtantsur | 14:17 |
lucasagomes | yuriyz, hey ya g'evening | 14:17 |
NobodyCam | hey hey yuriyz :) how are you doing | 14:18 |
dtantsur | yuriyz, g'evening | 14:18 |
yuriyz | all right I will go to Paris :) | 14:18 |
NobodyCam | woo hoo :) | 14:18 |
NobodyCam | brb ... gota run the store real quick for some somkes... brb | 14:20 |
*** igordcard has joined #openstack-ironic | 14:25 | |
*** pensu has left #openstack-ironic | 14:27 | |
*** ifarkas has quit IRC | 14:29 | |
*** zz_jgrimm is now known as jgrimm | 14:29 | |
lucasagomes | yuriyz, nice ! | 14:30 |
*** kongfy has joined #openstack-ironic | 14:30 | |
*** HenryG has joined #openstack-ironic | 14:31 | |
*** dprince has quit IRC | 14:38 | |
*** dhellmann is now known as dhellmann_ | 14:40 | |
*** dhellmann_ is now known as dhellmann | 14:41 | |
*** pcrews has joined #openstack-ironic | 14:42 | |
* NobodyCam is back | 14:42 | |
*** rwsu has joined #openstack-ironic | 15:05 | |
*** foexle has joined #openstack-ironic | 15:07 | |
openstackgerrit | David Shrewsbury proposed a change to openstack/ironic: Enable H904 PEP8 check https://review.openstack.org/127348 | 15:13 |
*** pensu has joined #openstack-ironic | 15:14 | |
*** foexle has quit IRC | 15:18 | |
*** pensu has quit IRC | 15:19 | |
*** pensu has joined #openstack-ironic | 15:35 | |
devananda | mornin, al | 15:35 |
NobodyCam | good morning devananda | 15:35 |
lucasagomes | devananda, morning and happy birthday :) | 15:38 |
devananda | lucasagomes: thanks! | 15:38 |
lucasagomes | devananda, have a min to talk about boot interface? I don't know if it worth breaking the deploy() interface into two anymore :/ | 15:39 |
*** pensu has quit IRC | 15:39 | |
lucasagomes | thing is, boot and deploy are related... for e.g, the virtualmedia needs to know what is the deploy method in order to build the iso | 15:39 |
lucasagomes | I'm starting to think that having only the deploy interface but abstracting what is possible and putting on common files is better than trying to split it in two interfaces | 15:40 |
NobodyCam | brb | 15:44 |
devananda | lucasagomes: humm | 15:45 |
lucasagomes | our PXE template for e.g have a lot of deploy parameters deploy_key etc... | 15:46 |
devananda | lucasagomes: so in principle, "run this temporary thing on that server" and "write this stuff to that server's disks" seem distinct | 15:46 |
lucasagomes | and a deploy entry on the template too :) | 15:46 |
devananda | but maybe they're not | 15:46 |
lucasagomes | yeah | 15:47 |
lucasagomes | because to run this thing, you need to know what is things thing | 15:47 |
devananda | in some cases, like PXE and IPA ,they are distinct | 15:47 |
lucasagomes | and what is it is usually the deploy method | 15:47 |
devananda | to run it, i think you need to know how to cause the machine to execute it | 15:47 |
devananda | (i)PXE is one method to "boot" | 15:48 |
devananda | vmedia is another | 15:48 |
devananda | the format of the binary data being "booted" is different, yes | 15:48 |
dtantsur | have a nice weekend, folks! | 15:49 |
*** dtantsur is now known as dtantsur|afk | 15:49 | |
devananda | dtantsur: you too! | 15:49 |
lucasagomes | dtantsur|afk, see ya | 15:50 |
*** jistr|bbl is now known as jistr | 15:50 | |
*** foexle has joined #openstack-ironic | 15:58 | |
*** dlaube has joined #openstack-ironic | 16:01 | |
NobodyCam | have a good weekend dtantsur|afk | 16:06 |
*** derekh has quit IRC | 16:11 | |
*** pensu has joined #openstack-ironic | 16:13 | |
*** foexle has quit IRC | 16:17 | |
lucasagomes | devananda, well kinda... e.g PXE IPA, if I try to use the same PXE as the ISCSI with IPA what happens after the machine ping the ironic API back? should the boot interface have control over the PXE template and switch the default entry from deploy to boot | 16:19 |
lucasagomes | but if it's IPA the pxe template won't even have that entry, cause IPA is a full disk image | 16:20 |
lucasagomes | and if I make the deploy method control the PXE template, well we already violated the layer there | 16:20 |
devananda | lucasagomes: so partition vs. whole disk is a huge bug | 16:29 |
devananda | we absolutely should not base API decisions on working around that | 16:29 |
lucasagomes | +1 | 16:30 |
lucasagomes | we had that bp for deploying full disk images before | 16:30 |
lucasagomes | but the code didn't land | 16:30 |
devananda | it required chagnes in Nova at a time we couldn't really make big changes there | 16:30 |
lucasagomes | right | 16:31 |
lucasagomes | we are talking about installing a boot loader in the disk in case we don't have the full disk image? | 16:31 |
lucasagomes | because that is a big problem to try to do via iscsi I think | 16:31 |
lucasagomes | many bootloader look at the host (for e.g to see if it's EFI or not) and install the bootloader accordingly | 16:31 |
*** ChuckC__ has joined #openstack-ironic | 16:32 | |
jroll | morning all, tgif :) | 16:32 |
jroll | and happy birthday devananda :D | 16:32 |
lucasagomes | jroll, morning | 16:32 |
devananda | jroll: o/^ | 16:32 |
JayF | devananda: where's your birthday hat? | 16:33 |
*** kongfy has quit IRC | 16:33 | |
lucasagomes | devananda, actually I was thinking about it some days ago, if we start fetching the image in the ramdisk and partitoning the image there | 16:33 |
JayF | shouldn't you be like, ^devananda for the day or smth? | 16:33 |
lucasagomes | would be cool, tho it would be kinda some code duplication with other tools (like anaconda/kickstart) | 16:33 |
lucasagomes | we would be kinda re implementing the wheel there :/ | 16:34 |
ChuckC__ | devananda: happy birthday \o/ | 16:34 |
jroll | cU <- birthday beer | 16:35 |
lucasagomes | heh "cu" means something pretty bad in portuguese :P | 16:35 |
devananda | lucasagomes: "partitioning the image" -- i'm not sure what this means | 16:38 |
lucasagomes | oh | 16:38 |
lucasagomes | the disk sorry | 16:38 |
devananda | lucasagomes: move the partitioning logic into IPA, so IPA woudl support both whole-disk and partition images? | 16:40 |
devananda | seems logical to me | 16:40 |
jroll | I agree with that, fwiw | 16:40 |
lucasagomes | devananda, yeah that works too | 16:40 |
devananda | JayF: you guys use Ansible at all? | 16:40 |
JayF | devananda: jroll: I've wondered if we should have some shared set of code between IPA and Ironic | 16:40 |
JayF | devananda: jroll: particularly around disk partitioning | 16:41 |
JayF | WTB oslo.disk? | 16:41 |
jroll | devananda: we use ansible as part of control plane deploys | 16:41 |
*** jcoufal has quit IRC | 16:41 | |
devananda | JayF: oslo is useful when we want to maintain that code across two projects | 16:41 |
jroll | also masterless puppet running locally | 16:41 |
jroll | but I think it's all moving to ansible eventually | 16:41 |
JayF | devananda: so oslo.disk then? heh | 16:41 |
devananda | JayF: not when we want to move it | 16:42 |
lucasagomes | tho how complex would be that partitioning stuff? cause couldn't we use part of anaconda for e.g? | 16:42 |
*** ChuckC__ has quit IRC | 16:42 | |
lucasagomes | I mean, kickstart does support installing image and tarballs now | 16:42 |
jroll | devananda: you'll still need partitioning in ironic for pxe driver, though | 16:42 |
devananda | jroll: this may sound crazy, but what about a ramdisk that just ran sshd, and an ansible driver in ironic? | 16:42 |
lucasagomes | it's not only os installation anymore | 16:42 |
lucasagomes | and it does partitioning, boot loader installation | 16:42 |
lucasagomes | etc | 16:42 |
*** krtaylor has quit IRC | 16:42 | |
lucasagomes | kinda sounds like we are reinventing the wheel here | 16:43 |
jroll | devananda: interesting... I like it, but to what benefit? | 16:43 |
JayF | devananda: sounds pretty razy | 16:43 |
JayF | *crazy | 16:43 |
JayF | devananda: you're reinventing the cloud-init wheel | 16:43 |
devananda | JayF: heh. exceut for provisioning. | 16:43 |
jroll | meanwhile, JayF is duct-taping that wheen | 16:43 |
jroll | wheel | 16:43 |
JayF | devananda: well I'd think the usage pattern becomes very un-cloudy when you embed config mangement in that deeply | 16:44 |
JayF | if Ironic isn't a CMDB, perhaps it shouldn't use a CMS system either :) | 16:44 |
devananda | heh | 16:44 |
*** slagle has quit IRC | 16:44 | |
jroll | JayF: ansible just runs commands... it happens to facilitate config management as well | 16:44 |
jroll | via other things | 16:44 |
devananda | JayF: except that is what we're doing, except with a limited scope, when ever we talk about declaratively asserting hardware state during provisioning / decom | 16:44 |
devananda | instead of a CMS asserting software state | 16:45 |
JayF | HMS :) | 16:45 |
JayF | The HMS Ironic? | 16:45 |
devananda | we're talking about asserting hardware state (RAID, BIOS, etc), partition table, | 16:45 |
devananda | oh, HMS Ironic. I like | 16:45 |
JayF | that's all hardware though, not software | 16:45 |
devananda | yes | 16:46 |
NobodyCam | good thing we didn't call our selfs titanic | 16:49 |
NobodyCam | oh wait that was RMS... :-p | 16:49 |
* NobodyCam will be afk for about an hour (or so-ish) so he can go visit the doctor ;-p | 16:56 | |
*** achanda has joined #openstack-ironic | 16:57 | |
lucasagomes | right I'll call it a day too | 16:59 |
lucasagomes | have a good night everyone! enjoy the weekend | 16:59 |
jroll | night lucas :) | 17:01 |
*** lucasagomes is now known as lucas-dinner | 17:01 | |
lucas-dinner | jroll, night :) | 17:01 |
jroll | devananda: you know offhand, for conf options, deprectaed_(group|name) means you can still use that group/name, right? | 17:04 |
devananda | it should | 17:04 |
jroll | ok, thanks | 17:05 |
devananda | that capability is there to allow upgrades while giving time for operators to update configs | 17:05 |
devananda | should log a warning when used, etc | 17:05 |
*** krtaylor has joined #openstack-ironic | 17:05 | |
jroll | right, that's what I figured, seeing some weirdness in nova/neutron things internally, wanted to verify | 17:06 |
*** viktors is now known as viktors|afk | 17:08 | |
*** igordcard has quit IRC | 17:08 | |
*** athomas has quit IRC | 17:16 | |
*** marcoemorais has joined #openstack-ironic | 17:25 | |
*** ryanpetrello has quit IRC | 17:25 | |
*** ryanpetrello has joined #openstack-ironic | 17:26 | |
*** killer_prince has quit IRC | 17:30 | |
*** rameshg87 has joined #openstack-ironic | 17:30 | |
*** rameshg871 has joined #openstack-ironic | 17:39 | |
*** rameshg871 has quit IRC | 17:39 | |
*** rameshg871 has joined #openstack-ironic | 17:40 | |
*** rameshg87 has quit IRC | 17:42 | |
*** chuckC_ has quit IRC | 17:43 | |
*** rameshg871 has left #openstack-ironic | 17:45 | |
*** ChuckC has quit IRC | 17:50 | |
*** harlowja has joined #openstack-ironic | 17:54 | |
openstackgerrit | David Shrewsbury proposed a change to openstack/ironic: Enable H904 PEP8 check https://review.openstack.org/127348 | 17:57 |
*** coolsvap|afk has quit IRC | 18:03 | |
openstackgerrit | David Shrewsbury proposed a change to openstack/ironic: Enable H702 PEP8 check https://review.openstack.org/127611 | 18:12 |
*** jistr has quit IRC | 18:13 | |
*** pelix has quit IRC | 18:20 | |
*** andreykurilin_ has joined #openstack-ironic | 18:25 | |
dlaube | Hey guys, when I attempt node-validate, I receive the following error/reason: "| deploy | False | Cannot validate iSCSI deploy. The following parameters were not passed to ironic: ['root_gb', 'image_source']" | 18:26 |
dlaube | but node-show has driver_info u'image_source': | | 18:27 |
dlaube | | | u'c969461f-a132-47f0-81e3-3db9c7053d6b', u'root_gb': u'10', | 18:27 |
dlaube | can anyone help me figure out why it's throwing that error when I try to validate the node? | 18:28 |
openstackgerrit | A change was merged to openstack/ironic: Add the PXE VendorPassthru interface to PXEDracDriver https://review.openstack.org/127454 | 18:29 |
dlaube | oh, maybe I need be defining those under node properties instead of under driver_info | 18:30 |
*** ChuckC has joined #openstack-ironic | 18:31 | |
*** chuckC_ has joined #openstack-ironic | 18:35 | |
*** ChuckC has quit IRC | 18:36 | |
*** ChuckC has joined #openstack-ironic | 18:36 | |
*** achanda has quit IRC | 18:37 | |
*** achanda_ has joined #openstack-ironic | 18:40 | |
*** openstackgerrit has quit IRC | 18:48 | |
*** Jmank has quit IRC | 18:52 | |
*** igordcard has joined #openstack-ironic | 18:53 | |
*** openstackgerrit has joined #openstack-ironic | 18:55 | |
openstackgerrit | Josh Gachnang proposed a change to openstack/ironic-python-agent: Add command metrics to IPA API https://review.openstack.org/119981 | 19:01 |
devananda | dlaube: yep. properties. | 19:05 |
devananda | hm. that isn't very clear. care to file a bug ? | 19:06 |
*** achanda_ has quit IRC | 19:11 | |
dlaube | ahh, I thought so. was just peeping the code to try to confirm | 19:17 |
dlaube | will do | 19:17 |
*** rushiagr is now known as rushiagr_away | 19:18 | |
dlaube | just filed bug 1379911 | 19:28 |
*** rainya has joined #openstack-ironic | 19:29 | |
* NobodyCam is back | 19:30 | |
*** achanda has joined #openstack-ironic | 19:36 | |
dlaube | huh, still seeing "| deploy | False | Cannot validate iSCSI deploy. The following parameters were not passed to ironic: ['root_gb', 'image_source']" on node-validate even though I've set them under properties | 19:39 |
dlaube | | properties | {u'root_gb': u'10', u'image_source': | | 19:39 |
dlaube | | | u'c969461f-a132-47f0-81e3-3db9c7053d6b'} | 19:39 |
*** ChuckC__ has joined #openstack-ironic | 19:43 | |
*** ChuckC__ has quit IRC | 19:47 | |
*** achanda has quit IRC | 19:48 | |
*** achanda has joined #openstack-ironic | 19:48 | |
*** pensu has quit IRC | 19:58 | |
NobodyCam | dlaube: https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/iscsi_deploy.py#L114-L120 | 19:59 |
NobodyCam | looks like that should be set in instance_info | 20:00 |
dlaube | ooh | 20:02 |
dlaube | ok i will try that | 20:02 |
dlaube | thanks | 20:02 |
jroll | dlaube: to be clear, nova should put those in instance_info | 20:03 |
NobodyCam | dlaube: are your setting doing all this manually ? | 20:04 |
dlaube | i thought i could get by with just interacting with ironic | 20:04 |
dlaube | yeah | 20:04 |
jroll | ah | 20:04 |
jroll | I see | 20:04 |
NobodyCam | :) | 20:04 |
dlaube | i will go install nova if thats the best way | 20:04 |
jroll | dlaube: so, what nova does is: set instance_info, set instance_uuid, hit provision state api with a target of active | 20:05 |
jroll | iirc | 20:05 |
NobodyCam | best is not the word I would use, is another way, | 20:05 |
NobodyCam | though our docs prob dont cover the manual process | 20:05 |
NobodyCam | its kinda out of scpoe for them, I feel | 20:06 |
jroll | it's the intended use right now, though we realize it's not the only use :) | 20:06 |
dlaube | gotcha | 20:07 |
NobodyCam | oh much better stated jroll :) | 20:07 |
jroll | :) | 20:07 |
devananda | dlaube: yea, it's possible to use ironic directly (no nova) or, in principle, alone (no keystone, neutron, or glance) but not recommended today | 20:14 |
devananda | some drivers have a hard dependency on swift, and all drivers depend on glance today | 20:15 |
dlaube | good to know | 20:15 |
Shrews | devananda: any opinions on whether there are any of the pep8 checks we want to continue ignoring in tox.ini? | 20:19 |
devananda | Shrews: quite possibly | 20:19 |
devananda | Shrews: have a link handy to which ones mean what? | 20:19 |
Shrews | devananda: http://legacy.python.org/dev/peps/ | 20:20 |
devananda | H are hacking rules, not pep8 rules | 20:21 |
Shrews | ugh | 20:21 |
devananda | yep | 20:21 |
Shrews | yeah... so, no :) | 20:21 |
Shrews | i've just been turning them on and seeing what they mean :) | 20:21 |
devananda | http://docs.openstack.org/developer/hacking/ | 20:22 |
NobodyCam | devananda: just checking as to why only a +1 on https://review.openstack.org/#/c/122453? | 20:22 |
devananda | Shrews: we should not ignore 702 | 20:24 |
devananda | I don't even see 904 in the docs tho :( | 20:24 |
devananda | 405 is worth cleaning up. should be easy | 20:24 |
devananda | ditto for 307 | 20:24 |
devananda | and 305 -- both of those are import orderings | 20:25 |
JayF | NobodyCam: IIRC only stable maintainers have Core on stable/ branches | 20:25 |
JayF | NobodyCam: but imbw | 20:25 |
devananda | NobodyCam, JayF: exactly. I can't +2 it | 20:26 |
Shrews | devananda: fyi, 405 was tedious | 20:27 |
openstackgerrit | David Shrewsbury proposed a change to openstack/ironic: Enable H405 PEP8 check https://review.openstack.org/127648 | 20:28 |
*** krtaylor has quit IRC | 20:28 | |
devananda | Shrews: ugh. thanks | 20:28 |
*** achanda has quit IRC | 20:36 | |
*** harlowja is now known as harlowja_away | 20:43 | |
*** harlowja_away is now known as harlowja | 20:47 | |
*** mitz_ has quit IRC | 20:47 | |
*** spandhe_ has joined #openstack-ironic | 20:47 | |
*** marcoemorais has quit IRC | 20:49 | |
*** marcoemorais has joined #openstack-ironic | 20:50 | |
NobodyCam | Shrews: slick use of (None,) * 5 in ipmitool :) | 20:51 |
*** jgrimm is now known as zz_jgrimm | 20:52 | |
Shrews | NobodyCam: not my idea | 20:53 |
NobodyCam | :p | 20:53 |
*** dhellmann has quit IRC | 20:56 | |
*** dhellmann has joined #openstack-ironic | 20:57 | |
*** marcoemorais has quit IRC | 21:02 | |
*** marcoemorais has joined #openstack-ironic | 21:02 | |
*** marcoemorais has quit IRC | 21:03 | |
*** marcoemorais has joined #openstack-ironic | 21:03 | |
NobodyCam | lol ... huummmmm /Shrews didn't call not that he also corrected capitalization on 127648 .. should i -1 this late on a friday... hummm... na ... more fun to just razz him about it | 21:05 |
*** achanda has joined #openstack-ironic | 21:06 | |
spandhe_ | JayF: hey! :) | 21:08 |
JayF | hi :) | 21:09 |
spandhe_ | JayF: do you use neutron or your own dhcp server? I have read about scaling problems with dnsmasq (neutrons default dhcp agent) | 21:09 |
*** achanda_ has joined #openstack-ironic | 21:10 | |
JayF | We wrote the code (we= ellenh in this case) for pluggable DHCP providers | 21:10 |
JayF | the options are None and neutron | 21:10 |
JayF | we use None :) | 21:10 |
JayF | and basically statically assign BMC and IPA IPs using isc-dhcp-server | 21:10 |
JayF | we do use neutron to swap networks around, some info about that is in the blogpost including link to our driver | 21:11 |
spandhe_ | JayF: ok.. thanks! Ill go through the blogpost.. | 21:11 |
*** achanda has quit IRC | 21:12 | |
*** achanda_ has quit IRC | 21:15 | |
*** jjohnson2 has quit IRC | 21:20 | |
*** praneshp has joined #openstack-ironic | 21:30 | |
openstackgerrit | A change was merged to openstack/ironic: Enable H904 PEP8 check https://review.openstack.org/127348 | 21:30 |
NobodyCam | +2'd 127648 did not +a as I also left a comment / question on it | 21:32 |
*** lucas-dinner has quit IRC | 21:33 | |
Shrews | NobodyCam: yeah, i removed that comment b/c the code totally didn't do what the comment said | 21:37 |
Shrews | NobodyCam: i was going to open a bug but i've been sidetracked with a family matter | 21:38 |
Shrews | NobodyCam: the option doesn't accept a list of servers, only a string, so no way we could get more than 1 anyway | 21:38 |
Shrews | i need to go afk now. feel free to open the bug for me. otherwise, i'll try to remember to do that later | 21:40 |
*** yjiang5 has joined #openstack-ironic | 21:40 | |
Shrews | (if a bug is even warranted... not sure about it yet) | 21:40 |
NobodyCam | enjoy the weekend Shrews :) | 21:43 |
*** achanda has joined #openstack-ironic | 21:44 | |
*** killer_prince has joined #openstack-ironic | 21:44 | |
*** killer_prince is now known as lazy_prince | 21:44 | |
openstackgerrit | Josh Gachnang proposed a change to openstack/ironic-specs: Support Decommissioning of Nodes https://review.openstack.org/102685 | 21:47 |
*** achanda has quit IRC | 21:48 | |
NobodyCam | here a quiestion for the spec review folk. say a spec calls for a nova driver change, and describes it in the nova driver section, would it be better to list the nova [spec|review] in that section or in Dependencies ??? | 21:50 |
NobodyCam | just asking.. | 21:50 |
praneshp | devananda: ping | 21:55 |
praneshp | devananda: is there a flow diagram somewhere that shows what happns when a node is rebooted using ironic | 21:55 |
praneshp | NobodyCam: JayF ^ | 21:55 |
devananda | praneshp: it depends on which driver is in use. also, i'm not aware of any recent flow diagrams for that | 21:55 |
openstackgerrit | Josh Gachnang proposed a change to openstack/ironic-specs: Support Decommissioning of Nodes https://review.openstack.org/102685 | 21:56 |
praneshp | devananda: ok. i am setting up an ironic cluster (for the first time), and was wondering how a node gets a ramdisk when rebooted, etc | 21:56 |
praneshp | i have the ironic (api and conductor) services running on an API node | 21:57 |
devananda | praneshp: using pxe_ipmi* drivers, it will get ramdisk via PXE boot | 21:57 |
praneshp | devananda: ok, is http://docs.openstack.org/developer/ironic/deploy/user-guide.html still valid documentation? | 21:58 |
praneshp | rloo seemed to think it might be a bit out of date | 21:58 |
JayF | praneshp: same with IPA driver, you pxe boot the agent up. Although I'd strongly reccomend enabling lucas's support for iPXE | 22:00 |
yjiang5 | praneshp: I'm setting up the environment also, and that user-guide seems still up to date. | 22:01 |
jroll | NobodyCam: I wonder if we should remove that section? idk | 22:01 |
*** achanda has joined #openstack-ironic | 22:01 | |
jroll | otherwise, yes, nova references should be there | 22:01 |
praneshp | yjiang5: great! did you start from scratch? I’ll ping you if I have questions. | 22:02 |
jroll | all these users :D | 22:02 |
yjiang5 | praneshp: In fact, I'm just using devstack to achieve it. | 22:03 |
yjiang5 | praneshp: We are trying to setup a mini openstack cloud (about 12 machines) for developer usage and we want to use ironic to manage the physical machines and then the ironic guest will be the real openstack cloud. | 22:06 |
NobodyCam | oh heres a question. "i want to deploy on to sdb and leave sda untouched. can Ironic support that layout" <- real question /me has been asked | 22:06 |
devananda | NobodyCam: not with the iscsi deploy, no. JayF: can IPA support that? | 22:07 |
* devananda files a bug in keystone ... https://bugs.launchpad.net/keystone/+bug/1379952 | 22:07 | |
jroll | devananda: "it depends" | 22:07 |
jroll | devananda: today, IPA writes to the smallest disk that is larger than 4GB | 22:07 |
praneshp | yjiang5: hmm, sounds like a fair use case. We’ve just been deploying openstack and use some of the VMs as management node. | 22:08 |
*** krtaylor has joined #openstack-ironic | 22:08 | |
*** andreykurilin_ has quit IRC | 22:08 | |
*** igordcard has quit IRC | 22:08 | |
jroll | devananda: so targeting it like that, no | 22:08 |
praneshp | yjiang5: by ironic guest, you mean whatver nodes you manage right (just to get the terminology right) | 22:08 |
yjiang5 | praneshp: yes. I'm new the ironic side, so possibly wrong term. | 22:08 |
NobodyCam | is that [system layout] something we should support? | 22:10 |
NobodyCam | s/system layout/disk config/ | 22:10 |
jroll | NobodyCam: I tend to think "don't touch this little snowflake disk" isn't something we should support, but we have preserve-ephemeral, so that's shot | 22:11 |
NobodyCam | no but I have use sda for system and sdb for data and do not think it out of thought to ask for the invers | 22:11 |
jroll | oh, I see | 22:12 |
jroll | idk, maybe? | 22:12 |
NobodyCam | lol and have you ever thought about deploying a data work load and not the system? | 22:13 |
jroll | I have not | 22:13 |
jroll | that's an interesting concept | 22:13 |
NobodyCam | think about loading a disk with a mysql database in to glance and then deploying it to the data driver powering up the node and the system disk boot and loads that database | 22:14 |
jroll | seems odd, but interesting | 22:15 |
NobodyCam | or dbl deploy... system and data payloads on to different drives... | 22:16 |
jroll | yeah | 22:16 |
* NobodyCam stops before he scares him self | 22:16 | |
jroll | lol | 22:16 |
NobodyCam | so I can see a use case for put this thing on this drive / lun / device ... | 22:18 |
* NobodyCam hides from devananda | 22:18 | |
JayF | NobodyCam: jroll: picking which device is used to write the image to is in the HardwareManager | 22:22 |
JayF | NobodyCam: jroll: So can be modified with some code-writing by the deployer :) | 22:22 |
jroll | oh, right | 22:22 |
JayF | And I wouldn't object to someone putting up more sane general-purpose logic into API | 22:22 |
JayF | er IPA | 22:22 |
NobodyCam | JayF: is that something we leave up to the user to do? | 22:22 |
JayF | wow IPA is inverse of API | 22:22 |
NobodyCam | lol | 22:22 |
* JayF is going to coin "from the API to the IPA" in a talk | 22:23 | |
NobodyCam | nice | 22:23 |
JayF | NobodyCam: Well we have a GenericHardwareManager which uses the logic jroll pointed out (smallest disk >4GB) | 22:23 |
JayF | NobodyCam: but I'm saying that logic was made when that was teeth-agent to meet onmetal needs. Given it's in the hw manager we could make the behavior more general purpose, then we'd just override it in our hardware manager (https://github.com/rackerlabs/onmetal-ironic-hardware-manager) | 22:24 |
jroll | JayF: nooooooooo, I wanted that phrase | 22:24 |
jroll | s/wanted/want/ I only just began wanting | 22:24 |
yjiang5 | JayF: Hi, I'm study the IPA, and a bit curiosity that why we have the call from ironic to IPA, instead of vice versa? | 22:24 |
JayF | jroll: we'll have to nerf swordfight to the death on Tuesday then | 22:24 |
jroll | hmm | 22:24 |
NobodyCam | oh that need a webcast | 22:24 |
jroll | yjiang5: the deploy call? | 22:25 |
ChuckC | first one to turn the whole phrase into a palindrome wins | 22:25 |
jroll | or prepare_image, technically | 22:25 |
JayF | yjiang5: I don't understand what you mean? Agents heartbeat to the API, and transmit information that it's completed commands that way | 22:25 |
JayF | yjiang5: the api only talks to api to tell it to do things :) | 22:25 |
JayF | it's just two rest services talking to one another :) | 22:25 |
JayF | except agent uses almost all vendor_passthru, which isn't very restful, but w/e | 22:26 |
yjiang5 | JayF: jroll, there was some discussion in Nova side before, that we want nova to talk to Ceilometer, and was rejected, because nova don't want to have such 3rd party dependency. I think the IPA is something similar (or we can argue that IPA is not 3rd party). | 22:27 |
JayF | IPA is tightly coupled to Ironic | 22:27 |
JayF | it's completely unusable without Ironic | 22:28 |
JayF | just because it's in a separate repo doesn't mean it's really a standalone project | 22:28 |
yjiang5 | JayF: But it's iin the ironic guest, right? So it basically means we have Ironic to depends on the guest? I'm quite new to ironic, so possibly my understanding/concern is wrong. | 22:29 |
JayF | yjiang5: Ironic, when configured to do so, manages everything about getting the ramdisk on the node, including powering the server on, telling it via BMC to boot from pxe, and telling Neutron to configure dhcp+pxe booting on the node | 22:31 |
yjiang5 | JayF: I'm just trying to imagine that nova will talk to some code running in a libvirt KVM guest. | 22:31 |
JayF | Think of IPA *as* KVM | 22:32 |
JayF | once we reboot into the instance, Ironic has no communication inside it whatsoever | 22:32 |
jroll | IPA does not run on a provisioned instance | 22:32 |
yjiang5 | JayF: Hmm, if considering the deploy kernel/ramdisk information is provided by admin, it seems make sense to take it as KVM, or , PV driver. Thanks for clarification. | 22:34 |
JayF | yjiang5: exactly; it's just a tool for doing the deployment; same as the pxe deploy ramdisk, although all that runs is a program to export the disk to image as an iscsi target (which is then mounted+imaged by the conductor) | 22:35 |
devananda | yjiang5: for analogy, ironic ~= libvirt && ipa|pxe|ilo|drac ~= kvm | 22:36 |
yjiang5 | JayF: thanks for answer. I will have a look on the code and then, maybe back to you later. | 22:36 |
yjiang5 | devananda: thanks. | 22:37 |
devananda | there is an additional separation of functionality, though. where KVM handles all aspects of starting a VM, within Ironic that is broken down by Driver Interface | 22:37 |
devananda | so we have power, deploy, console, management, etc... | 22:38 |
devananda | different interfaces for different functionality on the hardware | 22:38 |
yjiang5 | devananda: that make sense. KVM has multiple interface like different ioctl :) | 22:38 |
*** rainya has quit IRC | 22:40 | |
ChuckC | jroll: hey | 22:46 |
jroll | hiya ChuckC | 22:46 |
ChuckC | jroll: I'd like to get your thoughts on a spec I'm working on to make sure I'm representing ironics needs appropriately | 22:47 |
ChuckC | jroll: not out for review yet | 22:48 |
jroll | sure! | 22:48 |
jroll | you should put the short version on gerrit, to be honest, so everyone can chime in | 22:48 |
jroll | but happy to look before you do that, as well | 22:48 |
ChuckC | jroll: sure, maybe that's easiest anyway, I'll ping you when I have it out there (this is for 'resilient boot') | 22:49 |
jroll | cool | 22:50 |
jroll | ChuckC: what does 'resilient boot' mean? | 22:50 |
ChuckC | jroll: booting even though a nic failed | 22:51 |
jroll | aha, ok | 22:51 |
jroll | cool, looking forward to it :) | 22:51 |
ChuckC | jroll: thanks! | 22:51 |
NobodyCam | j* got a good chuckle from dtantsur|afk comment in the Developer impact section of https://review.openstack.org/#/c/97150/13/specs/kilo/whole-disk-image-pxe-driver.rst | 22:53 |
yjiang5 | hi, all, I have patches to devstack, trying to support hardware delopyment, https://review.openstack.org/#/q/status:open+project:openstack-dev/devstack+branch:master+topic:ironic-hw,n,z . It still have an issue that public interface's IP address is not restored cleanly when unstack.sh (Checking the code , I suspect baremetal has similar issue also), anyone has interests can have a look. | 22:53 |
*** openstackgerrit has quit IRC | 22:54 | |
jroll | NobodyCam: but we could still do those... | 22:55 |
jroll | tack on a swap partition after writing | 22:56 |
jroll | uses other disks for ephemeral | 22:56 |
jroll | etc | 22:56 |
NobodyCam | oh thats intrestring thought | 22:56 |
*** rwsu has quit IRC | 22:59 | |
*** praneshp has quit IRC | 23:02 | |
*** chuckC_ has quit IRC | 23:13 | |
*** achanda has quit IRC | 23:16 | |
NobodyCam | lol | 23:17 |
*** ChuckC has quit IRC | 23:22 | |
NobodyCam | anyone have a keystone v3 server up and running? | 23:28 |
jroll | lol I wish | 23:29 |
NobodyCam | lol | 23:29 |
jroll | NobodyCam: I want soft reboot :( | 23:30 |
NobodyCam | have you added your name to the list? | 23:30 |
jroll | maybe | 23:31 |
jroll | idk if we need a whole design session | 23:31 |
jroll | just, nova reboot should try to soft power off | 23:31 |
jroll | and kill it with fire if it's not responsive to that | 23:31 |
jroll | but I will add my name | 23:32 |
NobodyCam | :) and how long is that... a 4TB ram based data base takes awhile to write dout | 23:32 |
NobodyCam | dout=out | 23:32 |
NobodyCam | :-p | 23:32 |
jroll | ... | 23:32 |
jroll | yeah | 23:32 |
dlaube | hmm.. "Failed to deploy. Error: No VIFs found for node e9eff0a1-5971-4ddf-b0cb-e7d98f255fb6 when attempting to update DHCP BOOT options." | 23:33 |
dlaube | is this because I dont have any subnets configured in neutron? | 23:33 |
dlaube | heh | 23:33 |
jroll | dlaube: oh, right, nova puts VIF data in the ironic port objects | 23:33 |
jroll | NobodyCam: oh my name was there | 23:34 |
*** praneshp has joined #openstack-ironic | 23:38 | |
*** praneshp_ has joined #openstack-ironic | 23:39 | |
NobodyCam | doh | 23:39 |
NobodyCam | brb | 23:39 |
*** praneshp has quit IRC | 23:42 | |
*** praneshp_ is now known as praneshp | 23:42 | |
NobodyCam | dlaube: http://docs.openstack.org/developer/ironic/deploy/install-guide.html#configure-neutron-to-communicate-with-the-bare-metal-server | 23:48 |
dlaube | thanks NobodyCam | 23:51 |
dlaube | man, these docs have come a long way! | 23:51 |
*** Kui has joined #openstack-ironic | 23:51 | |
NobodyCam | :) | 23:52 |
dlaube | that's some really excellent work guys | 23:53 |
* NobodyCam points too all the great folks who have submitted the patches :) | 23:53 | |
NobodyCam | s/folks/folk/ | 23:54 |
NobodyCam | :-p | 23:54 |
NobodyCam | dlaube: if you run it in to areas that need improving toss up a patch :) | 23:56 |
NobodyCam | shameless plug! | 23:56 |
dlaube | sure thing ;) | 23:57 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!