Wednesday, 2014-04-30

« Tuesday, 2014-04-29 Index Thursday, 2014-05-01 »
*** morgabra has quit IRC00:01
*** dwalleck__ has joined #openstack-ironic00:09
*** morgabra has joined #openstack-ironic00:10
openstackgerritJosh Gachnang proposed a change to openstack/ironic: Adding a reference driver for the agent  https://review.openstack.org/8479500:13
*** dwalleck_ has joined #openstack-ironic00:19
openstackgerritJim Rollenhagen proposed a change to openstack/ironic-python-agent: Use processutils properly  https://review.openstack.org/9120200:20
*** dwalleck__ has quit IRC00:23
openstackgerritA change was merged to openstack/ironic: Synced jsonutils from oslo-incubator  https://review.openstack.org/9105500:26
*** dwalleck_ has quit IRC00:30
*** Shrews has joined #openstack-ironic00:42
*** tteggel has quit IRC00:52
*** tteggel_ has joined #openstack-ironic00:54
*** zdin0bot has quit IRC01:03
*** zdin0bot has joined #openstack-ironic01:05
*** derekh has quit IRC01:13
*** newell_ has quit IRC01:20
*** eguz has quit IRC01:22
openstackgerritJosh Gachnang proposed a change to openstack/ironic: Adding a reference driver for the agent  https://review.openstack.org/8479501:29
*** nosnos has joined #openstack-ironic01:35
*** krtaylor has joined #openstack-ironic01:43
*** harlowja is now known as harlowja_away02:27
*** Shrews has quit IRC02:27
*** coolsvap|afk is now known as coolsvap02:30
*** zdin0bot has quit IRC02:32
*** zdin0bot has joined #openstack-ironic02:37
openstackgerritOpenStack Proposal Bot proposed a change to openstack/ironic: Updated from global requirements  https://review.openstack.org/9122402:40
openstackgerritOpenStack Proposal Bot proposed a change to openstack/ironic-python-agent: Updated from global requirements  https://review.openstack.org/8872202:40
*** zdin0bot has quit IRC02:41
*** Shrews has joined #openstack-ironic02:46
openstackgerritOpenStack Proposal Bot proposed a change to openstack/python-ironicclient: Updated from global requirements  https://review.openstack.org/9123902:46
*** matsuhashi has quit IRC02:56
*** matsuhashi has joined #openstack-ironic02:57
*** JayF has quit IRC03:20
*** JayF has joined #openstack-ironic03:20
*** rameshg87 has joined #openstack-ironic03:27
openstackgerritA change was merged to openstack/ironic: Add ManagementInterface  https://review.openstack.org/8606303:33
*** zdin0bot has joined #openstack-ironic03:35
*** coolsvap is now known as coolsvap|afk03:42
*** rameshg87 has left #openstack-ironic03:46
*** martyntaylor1 has quit IRC03:47
*** eghobo has joined #openstack-ironic03:49
*** datajerk has quit IRC03:53
*** coolsvap|afk is now known as coolsvap03:54
*** shakamunyi has quit IRC04:05
*** matsuhashi has quit IRC04:10
*** Nisha has joined #openstack-ironic04:14
*** hemna has quit IRC04:23
*** nosnos has quit IRC04:26
*** eghobo has quit IRC04:32
*** rameshg87 has joined #openstack-ironic04:35
*** zdin0bot has quit IRC04:38
*** zdin0bot has joined #openstack-ironic04:39
*** eghobo has joined #openstack-ironic04:55
*** radsy has quit IRC05:03
*** matsuhashi has joined #openstack-ironic05:04
*** dwalleck_ has joined #openstack-ironic05:09
*** vkozhukalov has joined #openstack-ironic05:10
*** nosnos has joined #openstack-ironic05:16
*** zdin0bot has quit IRC05:22
*** nosnos has quit IRC05:22
*** nosnos has joined #openstack-ironic05:23
*** zdin0bot has joined #openstack-ironic05:23
openstackgerritOpenStack Proposal Bot proposed a change to openstack/ironic: Updated from global requirements  https://review.openstack.org/9122405:25
openstackgerritOpenStack Proposal Bot proposed a change to openstack/ironic-python-agent: Updated from global requirements  https://review.openstack.org/8872205:25
*** matsuhashi has quit IRC05:30
*** matsuhashi has joined #openstack-ironic05:30
*** matsuhashi has quit IRC05:36
*** matsuhashi has joined #openstack-ironic05:36
*** zdin0bot has quit IRC05:42
*** zdin0bot has joined #openstack-ironic05:42
*** vkozhukalov has quit IRC05:55
*** zul has quit IRC05:59
*** zdin0bot has quit IRC06:00
*** vkozhukalov has joined #openstack-ironic06:04
openstackgerritOpenStack Proposal Bot proposed a change to openstack/ironic: Imported Translations from Transifex  https://review.openstack.org/8850806:07
*** vkozhukalov has quit IRC06:11
*** zul has joined #openstack-ironic06:14
*** viktors|afk is now known as viktors06:19
*** max_lobur has joined #openstack-ironic06:41
*** killer_prince has quit IRC06:44
*** lazy_prince has joined #openstack-ironic06:45
*** lazy_prince is now known as killer_prince06:45
Mikhail_D_wkMorning Ironic! :)06:45
*** dwalleck__ has joined #openstack-ironic06:49
*** dwalleck_ has quit IRC06:52
*** romcheg1 has joined #openstack-ironic06:53
romcheg1Morning guys06:56
romcheg1Good news06:56
romcheg1Or I think I will wait until there is someone to tell the good news :)06:57
*** matsuhashi has quit IRC07:08
*** dwalleck__ has quit IRC07:08
*** matsuhashi has joined #openstack-ironic07:11
*** ifarkas has joined #openstack-ironic07:16
*** matsuhashi has quit IRC07:17
*** ekarlso has joined #openstack-ironic07:17
*** matsuhashi has joined #openstack-ironic07:17
*** datajerk has joined #openstack-ironic07:18
dtantsurMorning Ironic07:19
ifarkasmorning dtantsur!07:23
*** mrda is now known as mrda_away07:23
romcheg1Morning dtantsur ifarkas!07:29
* dtantsur brb07:29
ifarkasmorning romcheg1!07:29
dtantsurmorning romcheg1 ifarkas Mikhail_D_wk :)07:29
Mikhail_D_wkdtantsur g'morning :)07:30
*** eghobo has quit IRC07:30
Mikhail_D_wk romcheg1: What good news???? :)07:32
*** Nisha has quit IRC07:33
romcheg1The good news is that the first one of the patche that enable benchmarking Ironic is finally merged07:38
Mikhail_D_wkReally good news :)07:39
*** vkozhukalov has joined #openstack-ironic07:40
*** coolsvap is now known as coolsvap|afk07:44
*** datajerk has quit IRC07:45
*** coolsvap|afk is now known as coolsvap07:51
*** zigo_ is now known as zigo07:55
*** jistr has joined #openstack-ironic07:56
*** ndipanov_gone is now known as ndipanov07:58
*** ifarkas has quit IRC08:15
yuriyzmorning all08:19
*** ifarkas has joined #openstack-ironic08:19
*** derekh has joined #openstack-ironic08:21
*** lucasagomes has joined #openstack-ironic08:23
Mikhail_D_wkyuriyz: g'morning :)08:23
*** foexle has joined #openstack-ironic08:31
*** romcheg1 has quit IRC08:49
*** coolsvap is now known as coolsvap|afk08:54
*** zdin0bot has joined #openstack-ironic08:56
*** max_lobur has quit IRC08:59
*** datajerk has joined #openstack-ironic09:03
*** zdin0bot has quit IRC09:17
romchegMorning yuriyz09:20
*** rameshg87 has quit IRC09:25
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Remove 'node' parameter from the VendorPassthru interface  https://review.openstack.org/9036309:50
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Remove 'node' parameter from the Console and Rescue interfaces  https://review.openstack.org/9109709:50
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Remove 'node' parameter from the Power interface  https://review.openstack.org/9109309:50
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Remove 'node' parameter from the Deploy interface  https://review.openstack.org/9106209:50
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Remove 'node' parameter from the validate() methods  https://review.openstack.org/9110409:50
*** max_lobur has joined #openstack-ironic09:54
*** coolsvap|afk is now known as coolsvap09:55
*** rameshg87 has joined #openstack-ironic10:00
*** athomas has quit IRC10:00
*** datajerk has quit IRC10:00
*** athomas has joined #openstack-ironic10:08
dtantsurromcheg, re https://review.openstack.org/#/c/85387/ : I left inline 2 questions about your review, could you have a look?10:10
*** nosnos has quit IRC10:21
*** nosnos has joined #openstack-ironic10:29
*** coolsvap is now known as coolsvap|afk10:33
*** datajerk has joined #openstack-ironic10:33
*** datajerk has quit IRC10:46
romchegdtantsur: I'm looking10:47
*** datajerk has joined #openstack-ironic10:51
*** openstackgerrit has quit IRC10:51
*** andreykurilin has quit IRC10:52
*** max_lobur has quit IRC10:54
*** Mikhail_D_wk has quit IRC10:54
*** Mikhail_D_wk has joined #openstack-ironic10:56
*** andreykurilin has joined #openstack-ironic10:57
*** max_lobur has joined #openstack-ironic10:58
*** lucasagomes is now known as lucas-hungry11:39
*** rameshg87 has left #openstack-ironic11:54
*** yuriyz has quit IRC12:02
*** matsuhashi has quit IRC12:02
*** matsuhashi has joined #openstack-ironic12:02
*** matsuhashi has quit IRC12:03
*** nosnos has quit IRC12:09
*** jdob has joined #openstack-ironic12:17
*** coolsvap|afk is now known as coolsvap12:20
NobodyCamgood morning Ironic12:24
*** vkozhukalov has quit IRC12:32
*** coolsvap is now known as coolsvap|afk12:34
*** lucas-hungry is now known as lucasagomes12:41
lucasagomesmorning NobodyCam :)12:41
lucasagomesNobodyCam, is the weather a bit better over there?12:41
*** datajerk has quit IRC12:41
*** linggao has joined #openstack-ironic12:42
NobodyCamwell the sun is out12:42
NobodyCammorning lucasagomes12:42
NobodyCamI have the the weather channel on atm12:42
lucasagomesNobodyCam, ^^12:43
NobodyCamour next goal is / was https://www.google.com/search?client=safari&rls=en&q=new+orleans&ie=UTF-8&oe=UTF-812:44
NobodyCamgahh12:44
lucasagomesNobodyCam, after ur cup of joe and when u get some free time, can you take a look at this series here: https://review.openstack.org/#/c/90363/?12:44
lucasagomesnew orleans?12:45
NobodyCamNew Orleans12:45
NobodyCamL-p12:45
* NobodyCam needs cofffee12:45
lucasagomes:D12:46
romchegMorning lucasagomes, NobodyCam12:48
lucasagomesromcheg, hey, g'morning12:48
NobodyCammorning romcheg12:48
*** datajerk has joined #openstack-ironic12:50
*** datajerk has quit IRC12:55
*** rloo has joined #openstack-ironic12:59
*** matty_dubs|gone is now known as matty_dubs13:07
*** datajerk has joined #openstack-ironic13:13
*** zdin0bot has joined #openstack-ironic13:20
*** jbjohnso has joined #openstack-ironic13:22
NobodyCamlucasagomes: 90363 LGTM13:32
lucasagomesNobodyCam, :D! thanks13:32
lucasagomesyeah I want to clean the driver interfaces before we start adding more stuff to it13:32
NobodyCamlucasagomes: +++13:34
boris-42romcheg so what do you think about make a new rally-dsvm job13:36
*** mkerrin has quit IRC13:37
boris-42romcheg rally-dsvm-ironic13:37
boris-42or i even don't know how to write it properly13:37
romchegboris-42: Lemme fix the client first13:37
boris-42romcheg to be able to run benchmarks at least in gates of rally13:37
Shrewslucasagomes: any reason you didn't use CONF.ironic.api_retry_interval here instead of the hardcoded value? https://github.com/openstack/ironic/blob/master/ironic/nova/virt/ironic/driver.py#L43613:39
*** Mikhail_D_wk has left #openstack-ironic13:39
lucasagomesShrews, hmmm not that I can remember, I left a TODO there to make it configurable13:39
romchegboris-42: There are several cases we need to consider when benchmarking Ironic13:39
lucasagomesso it was part of the plans13:39
*** zdin0bot has quit IRC13:40
Shrewslucasagomes: ok. i'll throw up a quick change to use that value instead, then13:40
romcheg1. Benchmarking API service, 2. benchmarking conductor 3. Benchmarking agent13:40
lucasagomesShrews, ack, thanks!13:40
boris-42romcheg I don't think that we should split them all13:40
boris-42romcheg everything should be benchmarked through API13:41
romchegboris-42: Those cases require different Ironic set up13:41
romchegIPA has its own API13:42
romchegvkozhukalov is working on that13:42
*** dkehn_ has joined #openstack-ironic13:43
NobodyCammorning boris-42 & Shrews13:44
boris-42NobodyCam hi13:44
boris-42NobodyCam almost evening=)13:45
ShrewsNobodyCam: morning. lots of tornados here. was woken up at 5am by a weather service alert13:45
romchegNobodyCam, lucasagomes: you might have missed the news I posted in the morning. I managed to merge the first of the patches that enablem benchmarks for Ironic.13:45
romchegIt contains a critical bug, however, but I'm fixing it right now :)13:45
lucasagomesromcheg, w00t!!!13:45
lucasagomesgood stuff13:46
*** dkehn_ has quit IRC13:46
NobodyCamShrews: our next stop is New Orleans ... if its still there13:46
*** dkehn_ has joined #openstack-ironic13:46
NobodyCamromcheg: w00t!!13:46
*** dkehnx has quit IRC13:46
NobodyCamShrews: where are you.. north east?13:47
ShrewsNobodyCam: NC13:48
*** datajerk has quit IRC13:48
NobodyCamShrews: yep you are all over the news.13:48
NobodyCamShrews: Stay Safe!!!13:49
*** jgrimm has joined #openstack-ironic13:51
* Shrews notes that openstackgerrit bot is missing13:56
*** openstackgerrit has joined #openstack-ironic13:56
romchegNobodyCam, lucasagomes, boris-42: http://paste.openstack.org/show/77978/13:57
romchegEt voilà!13:57
boris-42romcheg nice13:57
boris-42romcheg now we should make the same in gates like keystone and glance does13:57
NobodyCamromcheg: :) nice !!14:00
lucasagomesnice!!!14:00
*** zdin0bot has joined #openstack-ironic14:01
boris-42lucasagomes NobodyCam fyi14:02
boris-42lucasagomes NobodyCam https://review.openstack.org/#/c/89798/14:03
NobodyCamboris-42: :) two +2's :)14:04
boris-42lucasagomes NobodyCam  this is how it looks14:05
boris-42lucasagomes NobodyCam  integration of rally gates in porject14:05
boris-42lucasagomes NobodyCam  here is the task (set of benchmarks) that will be run in gates14:05
boris-42https://review.openstack.org/#/c/89798/5/rally-scenarios/glance.yaml14:05
boris-42lucasagomes NobodyCam  and this is the result http://logs.openstack.org/98/89798/4/check/check-rally-dsvm-glance/8ffcaf9/14:05
boris-42lucasagomes NobodyCam  you can take a look in rally-plot14:06
NobodyCamram:0 disk:0 io_ops:0 instances:1 does not have 2048 MB usable ram, it only has 0.0 MB usable ram humm14:07
*** zdin0bot has quit IRC14:15
*** ndipanov has quit IRC14:16
NobodyCamany one else seeing this: NovaException: Error contacting Ironic server for 'node.set_provision_state'. Attempt 5 of 514:22
NobodyCambrb14:24
ShrewsNobodyCam: indicates timeouts contacting ironic-api. Is it running, or down during that attempt?14:25
NobodyCamShrews: odd bit is that I see the node power on ans start to post via the console14:26
ShrewsNobodyCam: hrm, odd. would be interesting to look at the activity timestamps14:28
* romcheg hates new Gerrit14:28
NobodyCamohhh... http://paste.openstack.org/show/WrbbwLM9Y3olIrjY0i5M/14:31
NobodyCamlucasagomes: Shrews: ^^^^^14:31
* lucasagomes clicks14:31
lucasagomesNobodyCam, ah :)14:31
Shrewseek14:32
openstackgerritImre Farkas proposed a change to openstack/ironic: Fix workaround for the "device is busy" problem  https://review.openstack.org/9012614:34
NobodyCamlucasagomes: have you seen this? the vms disk is 1 gb larger then the requested size.. ie... disk 31gb .. flavor 30 gb14:37
lucasagomesNobodyCam, larger or smaller? looking at that error they seem to be smaller14:38
lucasagomesI mean the vm disks14:38
NobodyCam(seed)nobodycam@nobodycam-HP-EliteBook-8460p:~/tripleo$ sudo ls -alh /var/lib/libvirt/images/baremetal_0.qcow214:41
NobodyCam-rw------- 1 root root 32G Apr 30 07:20 /var/lib/libvirt/images/baremetal_0.qcow214:41
NobodyCamvirsh reports 31 GB14:41
lucasagomesit might be the GB vs GiB thing14:41
lucasagomeswe divide it to 1024, so we use MiB and GiB14:41
lucasagomes1 GB = 1000 MB, 1 GiB = 1024 MiB14:42
*** vkozhukalov has joined #openstack-ironic14:42
NobodyCamlet me set it larger14:42
lucasagomesNobodyCam, e.g14:43
lucasagomes[stack@localhost ~]$ sudo fdisk -l14:43
NobodyCamjust fyi that is from a default devtest ru14:43
lucasagomesDisk /dev/vda: 15 GiB, 16106127360 bytes, 31457280 sectors14:43
lucasagomesGiB14:43
*** tteggel_ has quit IRC14:43
lucasagomes[stack@localhost ~]$ sudo parted /dev/vda unit GiB print14:44
lucasagomesModel: Virtio Block Device (virtblk)14:44
lucasagomesDisk /dev/vda: 15.0GiB14:44
*** tteggel_ has joined #openstack-ironic14:44
lucasagomes[stack@localhost ~]$ sudo parted /dev/vda unit GB print14:44
lucasagomesModel: Virtio Block Device (virtblk)14:44
lucasagomesDisk /dev/vda: 16.1GB14:44
lucasagomesthese units are very misleading/confusing... but for disks we should use MiB, GiB14:44
*** tteggel_ is now known as tteggel14:44
*** zdin0bot has joined #openstack-ironic14:45
openstackgerritAleksandr Gordeev proposed a change to openstack/ironic-python-agent: Fix unexpected stevedore traceback in tests  https://review.openstack.org/9075614:46
openstackgerritRuby Loo proposed a change to openstack/ironic: Implement API to get driver properties  https://review.openstack.org/7300514:49
*** zdin0bot has quit IRC14:50
*** dshulyak has quit IRC15:05
openstackgerritJarrod Johnson proposed a change to stackforge/pyghmi: Correct check for _sdr in command  https://review.openstack.org/9138915:06
openstackgerritJarrod Johnson proposed a change to stackforge/pyghmi: Correct check for _sdr in command  https://review.openstack.org/9139015:08
jbjohnsodevananda, btw, once I have time, I do have a scheme that wouldn't require BMCs to test against, at least not persistently.  If I put in a hook to bypass a very small amount of code, it would be hypothetically possible for a conversation to be replayed.15:10
NobodyCambrb15:15
openstackgerritA change was merged to stackforge/pyghmi: Correct check for _sdr in command  https://review.openstack.org/9139015:22
*** andreykurilin has quit IRC15:26
*** dwalleck__ has joined #openstack-ironic15:30
*** dwalleck_ has joined #openstack-ironic15:31
*** dwalleck__ has quit IRC15:35
agordeevNobodyCam: 1 gb larger size could be produced by this line https://github.com/openstack/tripleo-incubator/blob/master/scripts/create-nodes#L6-L915:37
agordeevNobodyCam: the devstack has a copy ot that script for ironic and does the same thing as well15:37
agordeevNobodyCam: https://github.com/openstack-dev/devstack/blob/master/tools/ironic/scripts/create-nodes#L14-1615:38
NobodyCamagordeev: good morning ... yep my disk was 31 while I requested a 30 gb disk15:39
agordeevNobodyCam: morning :)15:39
NobodyCamI am testing now by changing https://github.com/openstack/tripleo-incubator/blob/master/scripts/create-nodes#L9 to +515:39
*** tatyana has joined #openstack-ironic15:43
*** zdiN0bot has joined #openstack-ironic15:44
NobodyCamodd still getting same error :( humm15:45
NobodyCamlucasagomes: http://paste.openstack.org/show/A5r1XsvWKA3zkcEs06zi/15:50
NobodyCamany thoughts?15:50
lucasagomesNobodyCam, lemme see15:51
lucasagomesNobodyCam, hmm right15:52
lucasagomesthe disk is 35GiB Apr 30 15:42:57 ubuntu kernel: [  507.915095] sd 8:0:0:1: [sdb] 73400320 512-byte logical blocks: (37.5 GB/35.0 GiB)15:52
lucasagomes'Error: The location 36866 is outside of the device /dev/sdb.\n'15:52
lucasagomes36866 = 36GiB15:53
*** dwalleck_ has quit IRC15:53
lucasagomesNobodyCam, when you created the disk, you inpute 37 or 35?15:54
lucasagomesinputted*15:54
NobodyCamvirsh vol-create-as default baremetal_0.qcow2 35G --format qcow2 --prealloc-metadata15:55
*** eghobo has joined #openstack-ironic15:55
lucasagomesright, and the flavor?15:55
NobodyCamahh c2e55f11-244a-4861-abe6-ef2bf40c6e75 | baremetal | 2048      | 10   | 30        |      | 1     | 1.0         | True15:56
NobodyCam10 is disk 30 ephemeral15:57
NobodyCamlooks like dib devtest may be doing some bad math..15:57
NobodyCamlooking now... but seems like ephermeral should be 20 not 3015:58
lucasagomesyeah15:58
lucasagomesit's bigger than the disk15:59
lucasagomes2048 is ram?15:59
NobodyCamsi15:59
*** KurtMartin is now known as kmartin15:59
lucasagomesso that gives u 42G for that flavor15:59
*** dwalleck_ has joined #openstack-ironic16:00
NobodyCamtrying with a 45 gb disk16:00
NobodyCamwander what changed in devtest16:01
lucasagomesyeah16:01
lucasagomesidk :/ but that's def doing the write math there16:01
lucasagomesthe wrong*16:02
*** zdiN0bot1 has joined #openstack-ironic16:02
*** zdiN0bot has quit IRC16:02
*** matty_dubs is now known as matty_dubs|lunch16:04
NobodyCamwill this time it looks like it did good math.. the code looks ok too... this is strange (https://github.com/openstack/tripleo-incubator/blob/master/scripts/setup-baremetal#L108)16:06
*** ifarkas has quit IRC16:08
NobodyCamappers to be working now.. :-/16:12
lucasagomeshmm odd16:12
lucasagomesyou didn't change anything in that script?16:12
*** viktors has quit IRC16:12
NobodyCamnope.. I changed https://github.com/openstack/tripleo-incubator/blob/master/scripts/create-nodes#L9 to +1516:12
*** romcheg has quit IRC16:14
devanandamorning, all16:15
* NobodyCam sobs 2014-04-30 16:14:22.241 4318 ERROR ironic.drivers.modules.deploy_utils [-] Failed to detect root device UUID.16:15
NobodyCammorning devananda :)16:15
devanandawow, it's sunny here16:17
NobodyCamheheh your setting high temp records today according to weather channel16:17
lucasagomesmorning devananda16:18
jrollmorning ironic16:18
lucasagomesmorning jroll16:18
jrollit's gorgeous here in sf as well16:18
NobodyCammorning jroll16:18
devanandaadam_g: is https://review.openstack.org/#/c/87377/ still necessary?16:19
devanandaNobodyCam: I'd believe it. supposed to be 30C tomorrow16:19
devanandaI need sunscreen just to sit on the porch for a few minutes, and it's not even 9:30am16:20
NobodyCam:-p16:20
*** dwalleck_ has quit IRC16:22
*** dwalleck_ has joined #openstack-ironic16:22
*** hemna__ is now known as hemna16:24
*** vkozhukalov has quit IRC16:26
*** krtaylor has quit IRC16:35
*** martyntaylor has joined #openstack-ironic16:35
*** martyntaylor has left #openstack-ironic16:35
*** jistr has quit IRC16:36
*** romcheg has joined #openstack-ironic16:43
lucasagomes:( it's raining here16:47
NobodyCam:( brb16:48
*** derekh has quit IRC16:48
*** matty_dubs|lunch is now known as matty_dubs16:48
rloolucasagomes: qq for you. https://review.openstack.org/#/c/90964/2/ironic/api/controllers/v1/node.py16:49
lucasagomesrloo, yup16:49
rloolucasagomes: so line 373 sets chassis_uuid = args chassis_id16:49
rloolucasagomes: I have forgotten how this is supposed to work. but it seems a bit confusing that you aren't doing setattr(self, 'chassis_id', kwargs.get('chassis_id') .16:50
devanandalucasagomes: suggesting a wording clarification on the swap partition change sets, otherwise LGTM16:50
lucasagomesrloo, yeah, when the chassis_uuid is set, it triggers a _set_chassis_id method in the api16:50
rloolucasagomes:  i wonder if it needs more comments about the diff between the chassis_id's.16:50
lucasagomesdevananda, ack thank you will take a look16:51
lucasagomesrloo, yeah... it's because chassis_id doesn't exist in the API until chassis_uuid is set16:51
lucasagomesthen it's create chassis_id16:51
lucasagomeswe do it, so we don't expose chassis_id via the api with wsme16:52
devanandagoing to change my +2 on the first one actually, now that i see the second one16:52
lucasagomes(wsme would expose all the attributes, so we can't have it there when the class is evaluated)16:52
NobodyCammorning rloo :)16:52
lucasagomesrloo, and yes, it's confusing16:52
rloohi NobodyCam!16:52
NobodyCam:)16:53
lucasagomesdevananda, on the get rid of swap?16:53
rloolucasagomes: it seems like there ought to be a way to make it less confusing but I'm too lazy to think of how ;)16:53
lucasagomesrloo, yeah, it worth trying indeed16:54
*** harlowja_away is now known as harlowja16:54
lucasagomesrloo, having a way to convert api object to rpc object nicely would be enough16:54
rloolucasagomes: i suppose we could open a bug for it.16:55
lucasagomesdevananda, thanks for that suggestion :D I will test editing the commit message via gerrit16:55
NobodyCammmmmm bagel :-p16:56
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Place root partition last so that it can always be expanded  https://review.openstack.org/9067516:57
devanandalucasagomes: ditto on the prior patch16:57
lucasagomesdevananda, thank you16:58
devanandalucasagomes: np! thanks for fixing those :)16:58
*** dwalleck_ has quit IRC16:59
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Stop creating a swap partition when none was specified  https://review.openstack.org/8372617:00
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Place root partition last so that it can always be expanded  https://review.openstack.org/9067517:02
devanandalucasagomes: oh, hum... i just realized 83726 failed: check-tripleo-ironic-seed-precise FAILURE in 30m 05s17:05
lucasagomesdevananda, oh lemme see the reason17:05
devanandaloks like the same patch set passed the day before17:05
lucasagomeshttp://logs.openstack.org/26/83726/13/check-tripleo/check-tripleo-ironic-seed-precise/5004b7c/console.html17:07
lucasagomeslooks like something when setting up the ssh keys17:07
lucasagomesit's going to run again now that I updated the patch, but the error doesn't seem to be related to the change17:08
devanandalucasagomes: yep. let's see17:10
openstackgerritChris Behrens proposed a change to openstack/ironic: TaskManager: Only support single node locking  https://review.openstack.org/9141917:13
NobodyCamhummmm17:14
comstudboom17:15
NobodyCamlucasagomes: http://paste.openstack.org/show/w8ifztSUiLwIUhEgEuz4/17:15
lucasagomesNobodyCam, hmmm ur deploying a cirros image?17:16
lucasagomesit's not finding the UUID of the root fs17:17
lucasagomeswhich comes with the image17:18
NobodyCamya checking image17:18
NobodyCamahh17:19
*** krtaylor has joined #openstack-ironic17:20
*** athomas has quit IRC17:20
*** dwalleck_ has joined #openstack-ironic17:23
*** zdiN0bot1 has quit IRC17:24
lucasagomesNobodyCam, found the reasons?17:26
NobodyCamyes17:27
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Use GB instead of MB for swap  https://review.openstack.org/8378817:27
openstackgerritChris Behrens proposed a change to openstack/ironic: Add create() and destroy() to Node  https://review.openstack.org/8482317:28
ShrewsHrm. The LP bug hook doesn't appear to be working17:29
NobodyCamhttp://paste.openstack.org/show/rfLL1Vsyba4vyn03FDyd/17:29
NobodyCamlucasagomes: ^^^17:29
lucasagomes:)17:30
openstackgerritChris Behrens proposed a change to openstack/ironic: Add create() and destroy() to Node  https://review.openstack.org/8482317:31
NobodyCamso I either need to remove the partiton from my test image or teach ironic about partitions17:31
NobodyCamwhich I don't really want to do17:31
lucasagomescomstud, ^ use Partial-Bug: #<num> tag17:31
lucasagomescomstud, https://wiki.openstack.org/wiki/GitCommitMessages#Including_external_references17:31
lucasagomescomstud, you can edit the commit message directly on gerrit now :) neat!17:32
openstackgerritChris Behrens proposed a change to openstack/ironic: TaskManager: Only support single node locking  https://review.openstack.org/9141917:32
NobodyCamlucasagomes: I think you can even edit code too17:32
comstudlucasagomes: Yeah, I just fixed it17:32
comstud:)17:32
lucasagomescomstud, :d17:32
comstudright before your comment!17:32
comstudhehe17:32
lucasagomes:D*17:32
lucasagomesheh ack17:32
lucasagomesNobodyCam, really? wow17:33
comstudOh, nice.. re: editing commit msg on gerrit17:33
lucasagomesI like when ur comparing the code with the previous reviews... now u've a tab with all the patch-sets so u just click to compare the current with the previous ones17:33
JoshNangdevananda: for the tftp refactor, I was leaving some config options in the individual drivers to allow heterogenous environments. pxe can have its pxe template, and the agent will need a slightly different one. is that a reasonable way to do it?17:33
NobodyCamlucasagomes: nope I was wrong17:34
NobodyCamjust commit messge17:34
comstudlucasagomes: that was there before, also17:34
comstudoh17:34
comstudIt was up at the top under 'Patch Sets'17:34
comstudbut now it's also included below, also17:35
comstudthat's cool17:35
lucasagomes:)17:35
comstudlucasagomes: I tagged my 'refactor task-manager to only support single node' review as partially fixing that 'remove node argument' bug.17:36
openstackgerritA change was merged to openstack/ironic: Updated from global requirements  https://review.openstack.org/9122417:36
comstud1) since they're related... but 2) since I did remove 'node' from 1 method17:36
comstudheh17:36
lucasagomescomstud, heh, so it's probably going to conflict with my changes heh17:36
comstudit might17:37
comstudwell17:37
lucasagomesur patch is going to get merged first probably so I will rebase mine on top of urs later17:37
comstudthe tests might17:37
lucasagomesyeah, well it's grand I can rebase it17:37
comstudThey might not... I only fixed driver/utils/ get_mac_addresses or whatever17:37
comstudThe rest of my chagnes are limited to task_mangager.py itself17:37
lucasagomesI hope the remove node change doesn't hang for a long time in the review queue because that's the type of patch that takes needs constant maintenance17:37
comstudbut I did have to modify a lot of calls to task_manager.acquire() in the tests17:38
lucasagomesyeah17:38
lucasagomeslet's see :)17:38
comstudthat was passing [self.node.uuid] as the node to lock17:38
comstudie, list with single node17:38
comstudheh17:38
comstudi wanted to chagne all of those to mock task17:38
comstudbut that would make this patch much larger17:38
comstudso anyway.17:38
lucasagomesdevananda, btw, https://review.openstack.org/#/c/90363/ this series removes the 'node' from all the driver interfaces17:39
lucasagomescomstud, yeah, it's better to tackle one thing per time17:39
lucasagomescomstud, I +2'ed that patch but I want to wait for the gerrit vote first17:39
lucasagomesit's late here so I'm going to have a dinner and can't wait for jenkins17:40
lucasagomesjenkins vote*17:40
*** lucasagomes is now known as lucas-dinner17:42
lucas-dinnerhave a good night everybody!17:42
NobodyCamnight lucas17:43
NobodyCamthanks for your help17:43
lucas-dinnerNobodyCam, g'night, np :)17:43
NobodyCambrb17:43
*** romcheg has quit IRC17:46
openstackgerritJarrod Johnson proposed a change to stackforge/pyghmi: General cleanup of session  https://review.openstack.org/9142517:56
*** max_lobur has quit IRC17:56
*** dwalleck_ has quit IRC18:04
*** rustlebee has left #openstack-ironic18:29
*** jistr has joined #openstack-ironic18:32
*** vkozhukalov has joined #openstack-ironic18:41
*** zdiN0bot has joined #openstack-ironic18:46
*** zdiN0bot has quit IRC18:56
*** max_lobur has joined #openstack-ironic18:56
*** Mikhail_D_ltp has joined #openstack-ironic18:58
*** athomas has joined #openstack-ironic18:59
NobodyCamstarts a new build and will brb19:01
*** eghobo has quit IRC19:01
*** Manishanker has joined #openstack-ironic19:01
*** zdiN0bot has joined #openstack-ironic19:03
*** datajerk has joined #openstack-ironic19:06
*** zdiN0bot has quit IRC19:10
*** zdiN0bot has joined #openstack-ironic19:13
*** athomas has quit IRC19:15
devanandaJoshNang: ah, good point, so then it would be worth pointing that out, either in the commit msg or with a comment19:25
*** zdiN0bot has quit IRC19:26
*** zdiN0bot has joined #openstack-ironic19:27
JoshNangdevananda: good point. i'll update that19:33
rloodevananda: wrt developer doc for python-ironicclient. I was wondering why they didn't show up on http://docs.openstack.org/developer/python-ironicclient.19:33
JayFRFC: Name for a skeleton repo for IPA hardware manager extensions -- I was thinking something like ironic-python-agent-extension-skel ... but that's not particularly terse :)19:37
JoshNangwhy not put it directly in the ipa?'19:37
JayF(Basically this would be the barebones bits needed to write your own extension to IPA to do hardware things that maybe are out of scope of the agent or under NDA and cannot be open sourced)19:38
jrolleh?19:38
NobodyCamIPA-ES19:38
jrollthere's an interface19:38
jrollyou need a single line in setup.cfg19:38
jrolloh and a setup.py19:38
jrollok ignore me19:38
JoshNangadd an examples directory to ipa maybe?19:39
JayFjroll: ?19:39
*** datajerk has quit IRC19:39
jrollJayF: I was going to make a point to not do that19:40
jrollbut it was a bad point19:40
JayFJoshNang: hmm. maybe an examples directory, but I think a separate repo is preferable because the person can just clone that, then use it as the base for whatever extension they're doing19:40
jroll^19:41
JayFNobodyCam: I like ipa-es, but are we 100% sure we want to codify ipa as an acronym for ironic-python-agent? Also it's not very google-friendly since the abbreviation appears nowhere in the existing repo19:41
NobodyCam:) I like a good ipa19:42
JayFI like ironic-python-agent more than I like IPAs :)19:42
JayFhops are for bunnies ;)19:42
jrollwasn't that abbreviation half of the reason we chose this name?19:42
NobodyCamlol19:43
JayFI think it's more why we called it ironic-python-agent instead of python-ironic-agent19:43
jrollhmm19:43
jrollthat was my reason19:43
jrollI want to make ironic-ruby-agent and get rich :P19:43
NobodyCamlol19:44
jrollanyway, I think ipa-es is fine19:44
jrolland then like19:44
jrolllink from the docs19:44
jrolland link from a readme in the hardware manager19:44
jrollit'll be fine19:44
openstackgerritlinggao proposed a change to openstack/ironic: Support serial console access  https://review.openstack.org/6410019:45
NobodyCamIPA-es: goolged to: IP address exchanging, http://en.wikipedia.org/wiki/Template:IPA-es and some other minor refs19:47
jrollside note: we've now successfully provisioned a node using the agent :)19:47
JayFwfm, I'm going to work on getting something up. I'll probably start it and get it on github first, then work with -infra to get it into stackforge19:47
NobodyCamjroll: \o/19:47
jrolljust a couple small things to add to our in-flight patches19:48
*** dwalleck_ has joined #openstack-ironic19:50
*** foexle has quit IRC19:54
*** zdiN0bot has quit IRC19:55
devanandarloo: client 0.1.4 just tagged19:56
NobodyCam:)19:56
rloothx devananda. I was just thinking that maybe we should have updated the readme, but we can do that another time. i think i'll open a bug about it; maybe it'll be magically updated ;)19:57
devanandaJayF: jroll: have you guys looked at os-*-config tools?19:57
jrolldevananda: not since the summit19:57
jrollwhy?19:57
devanandajroll: extensions19:59
*** zdiN0bot has joined #openstack-ironic19:59
jrollah19:59
jrollhrm19:59
devanandamaybe20:01
jrollyeah20:01
jrollidk what that would gain us20:01
*** rloo has quit IRC20:03
*** rloo has joined #openstack-ironic20:03
NobodyCamdoh.... after rebooting remember to re-export USE_IRONIC :-p20:04
JayFI'd say either way, that it's useful to have extensions work as separate installable things, using OS tools (like pip, or $package_manager), then if you want to use something like os-*-config to package that up, you can, if you don't, you can :)20:05
devanandawhat sorts of things do you envision as extensions?20:06
devanandaand what do you see the interface to load / invoke them as?20:06
jrolldevananda: there's two pieces here20:06
jrollthere are 'agent extensions', which are modules that expose commands, loaded by stevedore20:06
devanandastevedore is already quite good at loading python extensions20:06
devananda:)20:06
jrollthen there is the hardware manager, which is pluggable and also loaded by stevedore20:06
*** zdiN0bot has quit IRC20:07
devanandaagent <=> hardware manager?20:07
jrolldevananda: I'm guessing you didn't mean os-*-config to *load* extensions, right?20:07
*** eghobo has joined #openstack-ironic20:07
jrolldevananda: hardware manager: https://github.com/openstack/ironic-python-agent/blob/master/ironic_python_agent/hardware.py#L10120:07
devanandajroll: you're right. I was thinking of something else -- forget it for the moment20:08
jrollheh, ok20:08
devanandajroll: ok, HM isn't a separate tool/project/executable. it's the class // API definition for interrogating hardware?20:10
jrollcorrect20:11
jrollso, if you have some weird hardware requiring proprietary tools or whatever, you can load up a hardware manager that handles that20:11
devanandaso stevedore may be used to load classes which derive from HardwareManager20:11
jrollcorrect20:11
devanandaeg t oexpose some fun(ky) hardware20:11
devanandacool20:11
jrollhttps://github.com/openstack/ironic-python-agent/blob/master/setup.cfg#L2620:12
jrollyeah20:12
jrollit only loads the most specific manager, which I don't love20:12
jrollI'd rather be able to plug in multiple managers, one that handles my exotic disks and one that handles my exotic cpu or something20:12
devanandayea. it'll need to be modular at some point20:13
devanandaif you want to audotedect what to load20:13
jrollright20:14
devanandawhich means a generic detector and standard way for plugins to register which hw they support20:14
jrollyeah20:15
devanandaand iterate lspci/lshw/... and load specific plugins where necessary20:15
jrollI'm almost thinking load it like a tree20:15
devanandammm, fun20:15
devanandayes20:15
jrolliterate most specific to least specific20:15
jrollcall the function until one succeeds20:15
devanandawell20:16
jrolle.g. the rest would raise NotImplementedError or something20:16
russell_honce there was some tree-like thing going on20:16
devanandaagain, stevedore can do stuff like this20:16
devanandausing python namespaces20:16
jrollby succeeds I mean it doesn't raise NotImplemented20:16
russell_hwell, I think I iterated the inheritance tree doing something20:16
jrollyeah20:16
russell_hbut it sucked, so I stopped20:16
russell_hnow it just loads everything from stevedore, and picks the class which reports the highest priority20:16
russell_hthere is still... plenty of room for improvement20:17
jrollrussell_h: right, so I want two 'highest-priority' plugins, one for disks, one for nics20:17
jrollor something20:17
russell_hjroll: right, you basically want some mixable thing20:17
jrollyeah20:18
devanandajroll: possibly https://github.com/openstack/stevedore/blob/master/stevedore/driver.py20:18
russell_hdevananda: https://github.com/openstack/ironic-python-agent/blob/master/ironic_python_agent/hardware.py#L20220:18
devanandaor possibly https://github.com/openstack/stevedore/blob/master/stevedore/hook.py -- if you need to invoke multiple extensions20:19
jrolldevananda: thinking a namespace for each kind of HW or something?20:19
jrollmmm20:19
devanandajroll: a namespace for each type of device20:19
russell_hthat might be the right thing to do20:19
russell_hyeah20:19
jrollright20:19
devanandawith a common check func20:19
devanandausing a stevedore hook could call >1, which may not be the right thing here20:20
devanandaactually EnabledExtensionmanager may be the right one -- https://github.com/openstack/stevedore/blob/master/stevedore/enabled.py20:21
devanandawhere the check_func is basically: can you manage hardware devie X20:21
russell_hdevananda: the thing is, you really want prioritization, not a boolean20:21
devanandarussell_h: but you want only one result, not a series20:22
russell_hdevananda: so, I want to have a "generic" one that we think will support most use cases20:22
russell_hand I don't want to have to whitelist hardware on that or something20:22
russell_hso, we can special case that20:23
russell_hif nothing else says it is enabled, use the generic one20:23
russell_hbut having each implementation report how suitable it is, then choosing the most suitable implementation, seemed more flexible20:23
devanandadhellmann: hi! any thoughts on how to do ^ with stevedore?20:24
russell_hthis way we can add support for special hardware in-tree if we want, but if someone wants to do something really wild, they can report a really high number and not have to fight the in-tree driver20:24
russell_hdevananda: look at that link I sent, I already have this just for the whole HardwareManager20:24
-openstackstatus- NOTICE: the gate is backed up due to broken nodepool images, fix in progress (eta 22:00 utc)20:25
*** ChanServ changes topic to "the gate is backed up due to broken nodepool images, fix in progress (eta 22:00 utc)"20:25
russell_hdevananda: I can refactor it to do it with a namespace for each kind of hardware20:25
russell_hand have the HardwareManager just load a bunch of children20:25
russell_hhmmm20:25
russell_hI guess another case though, is how we enumerate hardware in the first place20:26
russell_hlike do we instantiate a driver for each NIC20:26
NobodyCamreally changing our channel title20:26
devanandaNobodyCam: yes. the bot does that for all official channels now. it'll switch back when the broadcast/notice is done20:26
NobodyCam:) cool20:26
devanandarussell_h: you should be able to enumerate hardware prior to loading any hw-specific drivers20:27
russell_hI think so20:27
devanandarussell_h: assuming the kernel has the right modules loaded20:27
russell_hI'm just thinking about if people are doing something really nuts, like fundamentally different hardware20:27
russell_ha bitcoin mining ASIC or something20:27
devanandarussell_h: so you can pass the sig for each device to all the drivers/extensions for that type of device to acertain their suitability20:28
russell_hnot sure exactly what the agent should be doing to that20:28
russell_hI mean, lets san an FPGA that we want to deprogram20:28
russell_hbefore recycling the node20:28
russell_hs/san/say/20:28
russell_hshould there be a way to extend the agent to do that?20:29
devanandaif it's not supported by lshw/lspci/lsusb/etc ... yea, then it's challenging20:29
*** dwalleck_ has quit IRC20:31
*** Manishanker has quit IRC20:31
russell_hdevananda: are you saying to iterate lshw and try to find a driver suitable for every device, without worrying about what type of hardware it is?20:31
*** dwalleck_ has joined #openstack-ironic20:31
russell_hie, we don't say "oh this is a NIC, lets find a NIC driver for it", we just try to find any driver that reports itself as suitable for that device20:31
devanandarussell_h: i see two approaches20:32
devananda1. load all drivers. ask each one to determine their own usefulness. each driver looks at the system (whether via /dev or by looking for kernel modules or invoking utilities like MegaCli)20:33
devananda2. get list of hardware devices (eg, via lshw). pass device signatures to an extension manager which determines the appropriat driver to laod for eachd evice20:34
*** dwalleck__ has joined #openstack-ironic20:34
devanandafirst is bottom-up. second is top-down20:35
russell_hright20:36
devanandaIIUC, you want to have extension points for each type of device - at a minimum, storage, network, cpu, etc20:36
*** romcheg has joined #openstack-ironic20:37
*** zdiN0bot has joined #openstack-ironic20:37
*** dwalleck_ has quit IRC20:38
russell_hthats how I see it20:38
russell_hthough iterating the entire list of hardware, picking the most suitable driver for each, then having a totally standard interface to those drivers ("describe yourself for inventory", "deprovision yourself", etc) is... interesting20:40
russell_hbut things like.. where do we write out an image to20:40
russell_hseem to get complicated20:40
russell_hI had originally envisioned this as more of a "one class per flavor" type thing20:41
russell_hwhere we continually try to promote things to a generic driver in mainline which will autodetect where necessary20:41
*** dwalleck__ has quit IRC20:42
russell_hbut, the "autodetect where necessary" sounds a lot like your option #220:42
russell_her... your #120:42
devanandaI dont immediately see the connection to "where do we write out an image"20:43
devanandaassuming the hardware has attached storage device(s), pick the first, write it there20:43
jrollfirst storage device is pretty fragile20:44
jrollall depends on bios order, if I'm not mistaken20:44
devanandawhich we (should be) control(ing)20:45
JayFit can also depend on things like kernel version and image20:45
jrollcan that be controlled programatically?20:45
devanandafor things that support ipmi - yes20:45
devanandaset boot device20:45
jrollmmm20:45
devanandawell - that isn't the whole picture20:45
JayFNot all ipmi supports specifying boot order within a category though, right?20:45
*** eghobo has quit IRC20:46
jrollbut then we need code to make the correct device the first one20:46
JayFIf I have a box with 4 disks, I can't say 'boot off the third disk'20:46
jrollso either way, we need code to choose the correct device20:46
*** vkozhukalov has quit IRC20:46
jrollright?20:46
devanandaok, well20:46
devanandaJayF: no. but why does that matter for a cloud environment?20:46
JayFIt's to jroll 's point, that ordering of disks inside the OS can depend on bios boot order in some bios/firmwares20:46
devanandalet's keep in mind, this doesn't need to do everything that i might do on my desktop20:46
JayFbut hell I've even seen servers buggy enough that sda would swap depending on if the onboard sata or raid card driver was loaded first by the kernel20:47
JayFa lot of these bugs are gone now, but that doesn't mean our agent shouldn't be able to make a more intellegent decision than 'plop an image down on the first drive' as to what device needs imaging20:47
devanandait needs to do enough to provision, boot, and wipe a machine in a cloud. changnig the boot disk from hdc to hda isn't really all that relevant in this situation, IMO20:47
devanandaJayF: how is the agent goign to determine which disk to write the image to? -- programatically, based on input from Ironic.20:48
jrolldevananda: which is different from 'the first disk' :)20:49
devanandawell20:49
devanandahow is Ironic goign to know no tto use the first disk?20:49
devanandaif I"m booting 1000 nodes, what makes node 376 so unique that its image shouldn't be written to the first disk?20:50
JayFI wouldn't see it varying within a flavor20:50
JayFbut it's not unreasonable to have a hardware profile that, say, has separate OS and data disks20:50
jrolldevananda: I'm not talking about one unique node, i'm talking about 'those 1000 boxes shipped with a misconfigured bios where the intended boot disk is disk two'20:50
JayFand you would want to ensure the image gets written to the disk that's your data disk20:50
devanandaJayF: it's easy enough to d othat today -- just have a server with >1 disk, and ironic will put the OS on the first one -- and the tenant can use the rest as they see fit20:51
jrollthat would really suck for the tenant if the first disk happened to be 2 TB, and the second was a 32gb ssd20:51
devanandajroll: right - so reconfigure them. use a unility to do that. make Irnoic able to do *that*. but programatically detect-and-determine which disk to put the image on at deploy tiem? taht seems unnecessary20:52
jrolldevananda: I want to avoid ironic operator intervention20:52
jrollI want to plug in a rack, put those nodes in ironic, and push deploy20:53
devanandajroll: "fix the broken BIOS that these servers came from the factory with" is a one-time operation20:53
jrollright20:53
jrollbut I don't want to test bios disk orders every time I get a new server20:53
*** eghobo has joined #openstack-ironic20:54
devanandajroll: when a server has 12 disks, and #4 is somehow exposed as the first bootable disk, that should be fixed once -- not rechecked every time I deploy to that node20:54
*** eghobo has joined #openstack-ironic20:54
devanandajroll: as for the 2TB HDD + 32GB SSD -- why on earth would you put your OS on the SSD? that's far better as a cache volume or FS journal space20:55
devanandajroll: but your point (do interesting things with interesting disks) isn't lost -- but getting there is goign to require changes in several projects I suspect, including nova20:56
jrollsure20:57
jrollthere's also the kernel issues etc JayF pointed out20:57
jrollanyway, we have code to do this today20:58
jrollit's not the best but it works20:58
jrollthe algorithm is smallest disk that is greater than 4gb20:58
*** jdob has quit IRC20:59
devanandaShrews: hi! what do you think of adding rebuild test to tempest?21:01
Shrewsdevananda: hi. i'm afraid i know zilch about tempest, so it would probably take me a bit21:04
* Shrews running on battery power and cell phone (power outage) so likely not around much longer today21:05
NobodyCamShrews: stay safe...21:06
Shrewswicked weather. can't wait for a boring, sunny day21:06
devanandaShrews: ack. well,i think it'd be neat. I bet adam_g could help. also - stay safe. there's too much bad weather out there right now ...21:06
*** tatyana has quit IRC21:06
*** jistr has quit IRC21:07
Shrewsdevananda: i'll start looking into it21:07
devanandaShrews: places to start: https://etherpad.openstack.org/p/IronicCI and http://lists.openstack.org/pipermail/openstack-dev/2014-April/033917.html21:08
Shrewsack21:09
devanandaFWI folks, swapping the timing on two of our sessions so the Rally folks can attend our performance session21:09
*** max_lobur has quit IRC21:09
devananda*FYI21:09
*** Mikhail_D_ltp has quit IRC21:11
*** jbjohnso has left #openstack-ironic21:13
*** zdiN0bot has quit IRC21:15
*** max_lobur has joined #openstack-ironic21:16
*** linggao has quit IRC21:17
*** zdiN0bot has joined #openstack-ironic21:22
*** zdiN0bot has quit IRC21:22
*** max_lobur has quit IRC21:23
Shrewsalso FYI, the gerrit LP bug hook doesn't seem to be working. you'll need to update the bug yourselves. i notified -infra earlier today21:23
*** vkozhukalov has joined #openstack-ironic21:24
*** max_lobur has joined #openstack-ironic21:26
*** jgrimm has quit IRC21:31
rlooHi, does anyone know. with docstrings, if it is more than one line, does the first have to be short, followed by blank, followed by details?21:33
boris-42rloo yep21:35
boris-42rloo and first line should end with "."21:35
rloothx boris-42. Oh, not sure I knew about the first line. Seems inconsistent with the commit msg :-)21:35
boris-42rloo yep =)21:36
rlooand no, I'm NOT going to -1 anyone if they are missing the period! ;)21:36
boris-42rloo in commit message avoid dots21:36
boris-42rloo why not?)21:36
rlooboris-42: not important to me. cuz I feel like I -1 too much anyway. cuz cuz ...21:37
boris-42rloo http://stackalytics.com/report/contribution/rally/90 56 % in my case21:38
boris-42rloo seems ok=)21:38
boris-42rloo where are you here http://stackalytics.com/report/contribution/ironic/90 ?)21:38
rlooboris-42: ha ha. (Ruby Loo) Percentage wise we are close. I hope you win :-)21:39
boris-42rloo lol=)21:39
*** max_lobur1 has joined #openstack-ironic21:46
*** max_lobur has quit IRC21:49
*** matty_dubs is now known as matty_dubs|gone21:54
*** zdiN0bot has joined #openstack-ironic21:58
*** mrda_away is now known as mrda22:02
*** zdiN0bot has quit IRC22:09
*** romcheg has quit IRC22:11
*** zdiN0bot has joined #openstack-ironic22:16
*** zdiN0bot has quit IRC22:30
*** rloo has quit IRC22:30
*** rloo has joined #openstack-ironic22:31
*** lucas-dinner has quit IRC22:31
openstackgerritJosh Gachnang proposed a change to openstack/ironic: Adding a reference driver for the agent  https://review.openstack.org/8479522:43
*** eguz has joined #openstack-ironic22:48
openstackgerritA change was merged to openstack/ironic: Return error immediately if set_console_mode is not supported  https://review.openstack.org/9037622:49
openstackgerritA change was merged to openstack/ironic: Fix chassis-node relationship  https://review.openstack.org/9096422:49
openstackgerritA change was merged to openstack/ironic: Remove 'node' parameter from the VendorPassthru interface  https://review.openstack.org/9036322:49
*** eghobo has quit IRC22:52
*** zdiN0bot has joined #openstack-ironic22:58
*** zdiN0bot has quit IRC22:59
*** zdiN0bot has joined #openstack-ironic23:06
*** hemna is now known as hemna__23:12
*** killer_prince has quit IRC23:13
*** zdiN0bot has quit IRC23:17
*** zdiN0bot has joined #openstack-ironic23:23
*** zdiN0bot has quit IRC23:30
openstackgerritJosh Gachnang proposed a change to openstack/ironic: Factoring out PXE and TFTP functions  https://review.openstack.org/9023323:31
openstackgerritJosh Gachnang proposed a change to openstack/ironic: Adding a reference driver for the agent  https://review.openstack.org/8479523:32
openstackgerritJosh Gachnang proposed a change to openstack/ironic: Adding a reference driver for the agent  https://review.openstack.org/8479523:50
openstackgerritJim Rollenhagen proposed a change to openstack/ironic: Drivers determine acceptable power states  https://review.openstack.org/8674423:54
« Tuesday, 2014-04-29 Index Thursday, 2014-05-01 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!