Monday, 2014-10-20

« Sunday, 2014-10-19 Index Tuesday, 2014-10-21 »
*** Haomeng|2 has joined #openstack-ironic00:08
*** Haomeng has quit IRC00:08
*** mitz_ has joined #openstack-ironic00:33
Haomeng|2mrda: morning:)00:38
*** spandhe___ has joined #openstack-ironic01:32
*** spandhe_ has quit IRC01:34
spandhe___Hi everyone! I have a question. if dhcp_provider config option in ironic.conf is set to None, does that mean Neutron is not used at all?01:45
*** praneshp has joined #openstack-ironic01:50
*** spandhe___ has quit IRC01:53
*** nosnos has joined #openstack-ironic01:54
*** pcrews has joined #openstack-ironic01:55
*** spandhe has joined #openstack-ironic01:57
*** praneshp_ has joined #openstack-ironic01:59
*** praneshp has quit IRC02:02
*** praneshp_ is now known as praneshp02:02
*** kongfy has joined #openstack-ironic02:05
*** takadayuiko has joined #openstack-ironic02:26
*** takadayu_ has joined #openstack-ironic02:26
*** takadayu_ has quit IRC02:27
*** dlaube has joined #openstack-ironic02:37
*** ramineni has joined #openstack-ironic02:49
*** Poornima has joined #openstack-ironic03:05
*** Poornima has quit IRC03:05
*** kongfy has quit IRC03:09
*** pcrews has quit IRC03:12
*** dlaube has quit IRC03:12
*** nosnos has quit IRC03:24
*** nosnos has joined #openstack-ironic03:25
*** nosnos has quit IRC03:29
*** ryanpetrello has quit IRC03:40
*** ryanpetrello has joined #openstack-ironic03:41
openstackgerritAnusha Ramineni proposed a change to openstack/ironic: Update node-validate error messages  https://review.openstack.org/12886203:44
*** nosnos has joined #openstack-ironic03:48
*** HenryG has quit IRC03:49
*** rameshg87 has joined #openstack-ironic04:10
*** andreykurilin_ has joined #openstack-ironic04:17
*** lazy_prince is now known as killer_prince04:23
*** andreykurilin_ has quit IRC04:29
Haomeng|2spandhe___: I understand it is, but need to be confirmed from neutron guys04:35
*** dlaube has joined #openstack-ironic04:51
*** dlaube has quit IRC04:52
*** yarker has quit IRC04:57
*** pensu has joined #openstack-ironic04:59
*** yarker has joined #openstack-ironic05:00
*** Haomeng|2 has quit IRC05:03
*** Haomeng|2 has joined #openstack-ironic05:08
*** krtaylor has quit IRC05:09
*** krtaylor has joined #openstack-ironic05:13
*** kongfy has joined #openstack-ironic05:15
*** killer_prince has quit IRC05:31
*** spandhe has quit IRC05:42
*** pradipta_away is now known as pradipta05:43
*** loki184 has joined #openstack-ironic05:58
*** nosnos has quit IRC05:59
*** nosnos has joined #openstack-ironic06:00
*** killer_prince has joined #openstack-ironic06:01
*** killer_prince is now known as lazy_prince06:01
*** praneshp has quit IRC06:09
*** takadayuiko has quit IRC06:10
*** vinbs has joined #openstack-ironic06:10
*** k4n0 has joined #openstack-ironic06:11
*** nosnos has quit IRC06:39
*** nosnos has joined #openstack-ironic06:39
*** kongfy has quit IRC06:45
openstackgerritAnusha Ramineni proposed a change to openstack/ironic-specs: Add get/set boot mode to Management Interface  https://review.openstack.org/12952906:50
*** kongfy has joined #openstack-ironic07:09
*** rushiagr_away is now known as rushiagr07:10
*** pradipta is now known as pradipta_away07:10
*** pensu has quit IRC07:15
*** kongfy has quit IRC07:17
*** ndipanov has joined #openstack-ironic07:23
*** jcoufal has joined #openstack-ironic07:28
*** foexle has joined #openstack-ironic07:31
*** viktors|afk is now known as viktors07:57
*** leopoldj has joined #openstack-ironic08:01
*** lazy_prince is now known as killer_prince08:04
*** dtantsur has joined #openstack-ironic08:06
openstackgerritMotohiro/Yuanying Otsuka proposed a change to openstack/python-ironicclient: Switch to oslo.i18n  https://review.openstack.org/12954008:06
dtantsurMorning Ironic08:06
dtantsurMy IRC proxy was down since Fri evening, so if you wanted to tell me something - do it now please :)08:07
yuriyzmorning dtantsur and Ironic08:07
*** rameshg87_ has joined #openstack-ironic08:11
*** athomas has joined #openstack-ironic08:11
*** romcheg has joined #openstack-ironic08:12
*** jistr has joined #openstack-ironic08:12
*** derekh has joined #openstack-ironic08:12
*** lucasagomes has joined #openstack-ironic08:35
*** rushiagr is now known as rushiagr_away08:38
*** sambetts has joined #openstack-ironic08:48
*** Epistemic has joined #openstack-ironic08:49
*** kbyrne has joined #openstack-ironic08:51
openstackgerritA change was merged to openstack/ironic: Put a cap on our cyclomatic complexity  https://review.openstack.org/12913208:51
openstackgerritA change was merged to openstack/ironic: Continue heartbeating after DB connection failure  https://review.openstack.org/12930108:53
*** rushiagr_away is now known as rushiagr08:57
openstackgerritRoman Dashevsky proposed a change to openstack/ironic: Closes-Bug: #1383162  https://review.openstack.org/12955409:02
*** pelix has joined #openstack-ironic09:05
GheRiveromorning all09:08
*** killer_prince has quit IRC09:09
*** lazy_prince has joined #openstack-ironic09:09
*** rameshg87_ has quit IRC09:10
dtantsurGheRivero, yuriyz, morning!09:14
*** Epistemic has quit IRC09:29
*** k4n0 has quit IRC09:32
*** k4n0 has joined #openstack-ironic09:35
openstackgerritRoman Dashevsky proposed a change to openstack/ironic: Closes-Bug: #1383162  https://review.openstack.org/12955809:37
openstackgerritRoman Dashevsky proposed a change to openstack/ironic: Closes-Bug: #1383162  https://review.openstack.org/12917409:39
openstackgerritA change was merged to openstack/ironic: Update node-validate error messages  https://review.openstack.org/12886209:40
openstackgerritRoman Dashevsky proposed a change to openstack/ironic: Add SNMP driver for Aten PDU's. Closes-Bug: #1383162  https://review.openstack.org/12917409:41
*** ifarkas has joined #openstack-ironic09:43
*** Haomeng has joined #openstack-ironic10:02
*** Haomeng|2 has quit IRC10:03
openstackgerritRoman Dashevsky proposed a change to openstack/ironic: Add SNMP driver for Aten PDU's.  https://review.openstack.org/12917410:05
*** pensu has joined #openstack-ironic10:06
*** rushiagr is now known as rushiagr_away10:17
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Add a basic mechanism to route and validate vendor methods  https://review.openstack.org/12926110:28
*** ramineni has quit IRC11:04
*** lazy_prince has quit IRC11:09
*** slagle has joined #openstack-ironic11:09
*** lazy_prince has joined #openstack-ironic11:10
*** Poornima has joined #openstack-ironic11:11
*** Poornima has quit IRC11:13
*** Poornima has joined #openstack-ironic11:13
*** Poornima has quit IRC11:15
*** Poornima has joined #openstack-ironic11:15
*** Poornima has quit IRC11:17
*** rushiagr_away is now known as rushiagr11:21
*** loki184 has quit IRC11:24
*** romcheg has left #openstack-ironic11:24
*** killer_prince has joined #openstack-ironic11:32
*** Poornima has joined #openstack-ironic11:34
*** HenryG has joined #openstack-ironic11:39
*** lucasagomes is now known as lucas-hungry11:50
*** killer_prince has quit IRC11:52
*** vinbs has quit IRC11:52
*** lazy_prince has quit IRC11:52
*** lazy_pri- has joined #openstack-ironic11:52
*** lazy_pri- is now known as lazy_prince11:53
*** Poornima has quit IRC11:58
*** rameshg87 has quit IRC11:59
*** romcheg has joined #openstack-ironic12:03
*** romcheg has left #openstack-ironic12:03
*** romcheg has joined #openstack-ironic12:04
*** romcheg has left #openstack-ironic12:04
*** romcheg has joined #openstack-ironic12:04
*** marck has joined #openstack-ironic12:09
*** dprince has joined #openstack-ironic12:16
*** tatyana has joined #openstack-ironic12:30
*** rakesh_hs has joined #openstack-ironic12:30
openstackgerritRoman Dashevsky proposed a change to openstack/ironic: Add SNMP driver for Aten PDU's  https://review.openstack.org/12917412:32
*** Haomeng|2 has joined #openstack-ironic12:34
*** Haomeng has quit IRC12:35
*** lucas-hungry is now known as lucasagomes12:51
*** bradjones has joined #openstack-ironic12:52
*** pensu has quit IRC13:03
*** jjohnson2 has joined #openstack-ironic13:07
*** dprince has quit IRC13:13
*** rloo has joined #openstack-ironic13:13
*** lazy_prince has quit IRC13:18
*** dprince has joined #openstack-ironic13:19
*** romcheg has quit IRC13:23
*** romcheg has joined #openstack-ironic13:26
*** dprince has quit IRC13:29
*** dprince has joined #openstack-ironic13:30
jrollmorning ironic :)13:36
*** killer_prince has joined #openstack-ironic13:37
*** killer_prince is now known as lazy_prince13:37
jrolllucasagomes: I'm going to write that spec now13:38
lucasagomesjroll, oh crap! I forgot that13:39
lucasagomesjroll, right I cna write the API part if u want13:39
lucasagomesjroll, morning...13:40
jrollheh no it's fine :)13:41
jrollI didn't get to it friday13:41
jrollI'm just going to knock it out real quick before I get distracted :P13:41
lucasagomesright, yeah I woke up and started working on the vendor_passthru stuff13:41
lucasagomesI'm trying to add support of running methods in sync and async13:42
lucasagomeson both driver_vendor_passthru and vendor_passthru13:42
lucasagomeslet's see, it seems to work13:42
jrolllucasagomes: you think this should be an internal attribute, so only the new API can modify it?13:42
jrollI'm thinking so13:43
lucasagomesjroll, yeah... I think so too13:43
lucasagomesthe thing is... right now we are only changing the db values13:44
lucasagomesso it's not a big problem, we may want to support both13:44
lucasagomesbut once we start doing more things like talking to other services we def need to do that13:44
jrollhmm13:46
jrollI mean, other services can use the new endpoint, I don't see the issue13:46
jrollunless I'm misunderstanding13:46
lucasagomesjroll, right, no I think it's fine13:47
jrollok :)13:47
*** k4n0 has quit IRC13:55
*** rushiagr is now known as rushiagr_away13:58
*** lucasagomes has quit IRC14:01
*** tatyana has quit IRC14:06
*** tatyana has joined #openstack-ironic14:08
*** nosnos has quit IRC14:09
*** nosnos has joined #openstack-ironic14:09
openstackgerritJim Rollenhagen proposed a change to openstack/ironic-specs: Add maintenance reason field  https://review.openstack.org/12962914:13
jrollluca^ :)14:14
jrollugh14:14
*** nosnos has quit IRC14:14
jrolloh we lost him :(14:14
*** lucasagomes has joined #openstack-ironic14:16
openstackgerritJim Rollenhagen proposed a change to openstack/ironic: Add Node.maintenance_reason  https://review.openstack.org/12864514:17
lucasagomesand back14:18
openstackgerritJim Rollenhagen proposed a change to openstack/ironic: Add maintenance_reason when setting maintenance mode  https://review.openstack.org/12864614:18
jrolllucasagomes: https://review.openstack.org/#/q/status:open+branch:master+topic:bp/maintenance-reason,n,z14:18
jrollspec is in there too14:19
jrollshould tag your api thing :)14:19
lucasagomesjroll, awesome will take a look14:19
lucasagomesjroll, will do, ta much!14:19
NobodyCamgood morning Ironic14:19
NobodyCammorning jroll and lucasagomes :)14:20
lucasagomesNobodyCam, morning14:20
NobodyCamhave a good weekend?14:20
jrollheya NobodyCam :)14:20
NobodyCam:)14:20
rloomorning NobodyCam, jroll, lucasagomes :)14:21
lucasagomesrloo, hey ya! morning14:21
NobodyCammorning rloo :)14:21
*** pcrews has joined #openstack-ironic14:24
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Add API endpoint to set/unset the node maintenance mode  https://review.openstack.org/12892514:24
openstackgerritMathieu Mitchell proposed a change to openstack/ironic-python-agent: Write image directly to disk  https://review.openstack.org/12904614:27
*** jgrimm is now known as zz_jgrimm14:27
*** rwsu has joined #openstack-ironic14:30
jrolllucasagomes: did you catch my comments on patchset 5? https://review.openstack.org/#/c/128925/14:34
lucasagomesjroll, oh no, I just rebased, lemme mark it as WIP and address that14:34
jrollcool14:34
jrollnot sure if they're valid comments, but I think so :P14:35
lucasagomesjroll, sorry, it's because I'm looking at the vendor thing so I don't wanna stop and start fixing other things14:35
lucasagomesor I will lose the track14:35
jrollno worries, I can wait :P14:36
*** pensu has joined #openstack-ironic14:41
NobodyCambrb14:46
*** ndipanov has quit IRC14:54
*** jistr has quit IRC14:59
dtantsurjroll, NobodyCam, morning14:59
dtantsurand rloo :)14:59
jrollgood afternoon dtantsur :)14:59
jrolland rloo :)14:59
rloohi, afternoon dtantsur15:00
jrolldtantsur, rloo, I has a spec for you :) https://review.openstack.org/#/c/129629/15:00
dtantsurjroll, already reading :)15:01
*** jistr has joined #openstack-ironic15:01
jrollperfect15:01
rloothx jroll. Why did I ask for it? Now I'll have to read it :-)15:01
jrollhehe, you did ask for it15:02
NobodyCammorning dtantsur15:02
romchegHey hey everyone!15:02
dtantsurromcheg, hi!15:03
NobodyCammorning romcheg :)15:03
rlooafternoon romcheg!15:03
jrollhiya romcheg :)15:04
* NobodyCam makes his standard monday morning call for agenda Items15:09
devanandamorning, all15:11
jrollmorning devananda15:12
lucasagomesdevananda, morning15:12
*** jjohnson2 has quit IRC15:13
NobodyCammorning devananda15:15
yuriyzmorning/evening all15:15
*** zz_jgrimm is now known as jgrimm15:19
rloomorning devananda, evening yuriyz15:22
rloodevananda, now that juno has been released, were you/we (not me) going to put together release notes ala https://wiki.openstack.org/wiki/Ironic/ReleaseNotes/Icehouse?15:23
rlooNobodyCam: I was going to say that I sometimes think of agenda items but then I forget them when Monday rolls around.15:25
NobodyCam:-p15:25
rlooNobodyCam: but I just remembered something I had asked on IRC a while ago. what s/w to use for generating diagrams etc for our docn. Was that ever resolved/decided?15:26
NobodyCamI thouhgt we came up with (thinks) asciichat.net15:27
rlooNobodyCam: that doesn't exist? I see asciichart.net?15:28
NobodyCamahh thats it I think..15:28
*** mitz has quit IRC15:29
*** mitz has joined #openstack-ironic15:29
devanandarloo: yep, we need to do that15:30
rloodevananda: did you mention http://asciiflow.com/?15:30
rlooNobodyCam, devananda: basically, the source needs to be ascii?15:31
devanandaI believe I had suggested asciiflow -- that's in our spec template as a requirement, in fact15:31
*** leopoldj has quit IRC15:31
rloodevananda: ok, asciiflow it is I guess ;)15:32
NobodyCamthats the onE!!!! Ty devananda for correcting /me15:33
openstackgerritJim Rollenhagen proposed a change to openstack/ironic-specs: Add maintenance reason field  https://review.openstack.org/12962915:35
jrollJayF, dtantsur, updated ^15:35
openstackgerritA change was merged to openstack/ironic: Add SNMP driver for Aten PDU's  https://review.openstack.org/12917415:35
dtantsurjroll, still not clear, what do to with clearing maintenance mode via old API15:36
*** ndipanov has joined #openstack-ironic15:37
JayFdtantsur: it's supported; but doesn't touch maintenance_reason15:37
JayFdtantsur: I thought he was fairly explicit about it :x15:37
dtantsurJayF, that's bad IMO15:37
NobodyCammorning JayF :)15:37
dtantsurJayF, so we'll easily leave outdated maintenance reason, if someone uses old API, right?15:37
JayFI disagree; if you're calilng node.update you want to update fields of the node directly, and that shouldn't clear out potentially valuable information15:37
JayFlast_error, for instance, isn't cleared just because you reset a provision state15:38
*** dlaube has joined #openstack-ironic15:38
JayFmorning NobodyCam15:38
jrolldtantsur: if you want to clear it, use the new api15:38
JayFjroll: or call node.update maintenance-reason=null, right?15:38
jrolleh15:38
jrollI was thinking on making that internal15:38
JayFjroll: or something like that; since you said maintenance_reason should be added to node update api15:38
jrollso you can't edit directly15:38
dtantsurjroll, we're talking about backward compatibility. thus you can't trust the value, because someone might have used the old API15:38
jrolldtantsur: I mean, do you want to make it not backward compatible?15:39
jrolldtantsur: I prefer my tooling to keep working15:39
JayFdtantsur: I think it's OK as long as jroll keeps the support for clearing maintenance_reason in the node update api15:39
dtantsurjroll, yes with cleaning automagically on old API15:39
NobodyCamhey JayF just checking 103105, do you happen to have a chance to take a look at that one?15:39
jrolldtantsur: hmm, maybe15:39
rlooupdated to mention using asciiflow: https://wiki.openstack.org/wiki/Ironic/Developer_guidelines#diagrams15:39
jrollI mean this is why we have specs15:39
JayFNobodyCam: I don't have time to test it today; I can look at the changes and help troubleshoot any problems not sure I can give a great review on DIB anything15:40
dtantsurjroll, JayF, my point is: if we're in environment, where some tools use old API, some - new, we can't trust this value, if it's not cleaned on unsetting maintenance mode whichever way15:40
JayFjroll: dtantsur: I'm OK with either method; if we have maintenance_reason clear if maintenance=false via node-update api, lets not expose maintenance_reason via node update api then15:40
jrollok15:41
JayFdtantsur: I guess so; but I feel like we already have that bucket of pain with last_error15:41
JayFdtantsur: honestly regardless of how it works it's better than what we have now; which is a bucket of nothing :)15:41
dtantsurJayF, last_error is at least cleared on any operation touching last_error IIRC :) but yeah, anything is better15:42
Shrewsdtantsur: when you have a moment, i can't seem to get your mock suggestion to work  :(15:42
JayFdtantsur: except an oper calling node-set-provision-state to unbreak his last_error :P15:42
JayFjroll: http://logs.openstack.org/46/129046/2/check/check-tempest-dsvm-ironic-agent_ssh-src-nv/d853790/logs/devstacklog.txt.gz looks like we have to set SWIFT_TEMPURL_KEY :(15:42
dtantsurShrews, what happens?15:42
Shrewsdtantsur: http://paste.openstack.org/show/122238/15:43
jrollJayF: gogogo15:43
JayFjroll: what should I set it to? Anything? lol15:43
jrollyeah15:43
jrollset it to 'namesarehard'15:43
openstackgerritJim Rollenhagen proposed a change to openstack/ironic-specs: Add maintenance reason field  https://review.openstack.org/12962915:44
jrolldtantsur: updated again, I think this should handle the concerns here15:44
dtantsurShrews, lemme see15:44
*** todd_dsm has joined #openstack-ironic15:45
*** jistr has quit IRC15:45
dtantsurShrews, set CONF.hash_partition_exponent to 0 for this test. anyway we mock md5.15:46
dtantsurbrb15:47
Shrewsdtantsur: ah, yep. thx. that was driving me nuts15:48
openstackgerritDavid Shrewsbury proposed a change to openstack/ironic: Improve hash ring value conversion  https://review.openstack.org/12903115:53
*** eghobo has joined #openstack-ironic15:58
*** ndipanov has quit IRC16:00
JayFjroll: https://review.openstack.org/#/c/129653/16:00
rloodevananda: if you have a minute to discuss 118733. Seems like we should -2 or give more direction, cuz I think more time is being spent here by reviewers than is warranted.16:01
devanandarloo: /me looks16:01
jrollJayF: so on this change... I'd almost rather just set it globally like ADMIN_PASSWORD etc16:03
jroll"globally"16:03
jrolllike... for everything. can't hurt.16:03
JayFjroll: Eh. I would rather my change by scoped to only modify things in agent runs16:03
jrollidk, I'd see what reviewers think I guess16:04
devanandarloo: so, I left some detailed comments there almost a month ago, and there has been no activity since -- I'm not sure what your concern is?16:04
jrollJayF: seems like it should work16:04
rloodevananda: so if they update the commit and do whatever other suggestions people are making, it is ok to approve it? there's already a +2 and lucasagomes awhile ago almost approved it.16:05
NobodyCamdevananda: just a question hve the votes been counted for our sessions?16:06
devanandaNobodyCam: yes16:06
NobodyCam:)16:07
devanandaI've got a proposed schedule to go over16:07
devanandarloo: let me answer that by asking a question - how do you feel about us abandoning patches that have -1's with constructive feedback that go ignored by their original authors after, say, a month?16:08
devanandagerrit used to auto-abandon such things, but for reasons, that feature was disabled a while back16:09
rloodevananda: something that is abandoned can be 'resumed' or whatever again, so yes, after some period of time, I think that makes sense.16:09
devanandagreat. I'll do a round of abandoning things16:10
rloodevananda: thx. Is 4 weeks a good time frame? How long was mrda on vacation?16:10
rloothose australians take long vacations ;)16:10
devanandalucasagomes, dtantsur: since you had +A'd 118733, but didn't respond to my comments on it, I'd like to check that we agree on whether to approve or abandon it16:11
lucasagomesdevananda, lemme check16:11
*** jgrimm is now known as zz_jgrimm16:12
*** Hefeweizen has joined #openstack-ironic16:12
dtantsur118733?16:12
*** ndipanov has joined #openstack-ironic16:12
devanandayes16:12
dtantsuroh, I barely remembered16:13
devanandarloo: I think 4 weeks is a good length of time. after that, the codebase has probably changed enough that it would have to be rebased anyway16:13
rloodevananda: true enough. At least, I would hope that it would have to be rebased by then ;)16:13
dtantsurdevananda, I'm neutral here. And as I guess now it will require rebasing, I think we have to abandon it16:14
lucasagomesdevananda, I agree that fixing typos on comments or private methods is not a bug. Tho I'm fine with that change when he improves the commit message and remove the bug tag16:15
JayFThis https://review.openstack.org/#/c/129629/3 has 2x +2s. I'm going to land it before the meeting (in an hour or two) if nobody wants a look at it first.16:15
devanandalucasagomes: exactly16:15
lucasagomesdevananda, sorry for not answering before16:15
lucasagomesI will add it there16:15
devanandaJayF: /me looks16:16
rlooJayF: the pressure is on :-). looking since I asked for it...16:16
rlooJayF: isn't there jenkins for the specs? I wanted to read the beautifully formatted version.16:17
JayFrloo: looks like CI hasn't run on the last patchset yet16:18
rlooJayF: yeah. looking at the previous times, it could be sooner or later. oh well.16:19
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Add a basic mechanism to route and validate vendor methods  https://review.openstack.org/12926116:19
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Add sync and async support for [driver_]vendor_passthru  https://review.openstack.org/12966216:19
devanandaJayF: approved16:19
lucasagomesdevananda, ^ do you think I need a spec for those vendor passthru refactoring/enhancements?16:19
* lucasagomes wants to add support for diff HTTP methods as well16:20
*** igordcard has joined #openstack-ironic16:20
devanandalucasagomes: I think that's something I'd like to talk about at the summit16:21
NobodyCamlol .. devananda beets /me to the click again16:21
sambettslucasagomes: Did you have a chance to look at the patch that I put up for the passthru logging??16:21
devanandalucasagomes: namely, how and how much we expand support for vendor passthru16:21
lucasagomessambetts, yeah I did :/ but the logs are the same I was adding there already :/16:21
lucasagomesdevananda, yeah, right... these 2 patchs I put up now is mostly for making it consistent, instead of having driver_vendor_passthru to always runs sync and node vendor_passthru to always runs async16:22
lucasagomesdevananda, I also need GET support for the iPXE driver16:22
lucasagomesthat's why I was working on that16:22
lucasagomesI'm*16:22
*** eghobo has quit IRC16:23
sambettslucasagomes: didn't know if my function could be extended to reduce the duplication of the try catches used in your patch?16:24
lucasagomessambetts, oh, hmm lemme take a look I guess it could16:25
openstackgerritA change was merged to openstack/ironic-specs: Add maintenance reason field  https://review.openstack.org/12962916:25
sambettslucasagomes: Its ok if you don't think its nessessary its just a suggestion :-)16:26
NobodyCamwas that our first K spec?16:26
NobodyCamcongratz!!16:26
NobodyCam:)16:26
lucasagomessambetts, aight, yeah but worth taking a look anyway, removing duplication is good :)16:27
*** enikanorov has quit IRC16:32
openstackgerritRuby Loo proposed a change to openstack/ironic-specs: Delete the placeholder file: example.rst  https://review.openstack.org/12966516:33
dlaubeg'morning guys16:35
openstackgerritLucas Alvares Gomes proposed a change to openstack/ironic: Add sync and async support for [driver_]vendor_passthru  https://review.openstack.org/12966216:35
*** marcoemorais has joined #openstack-ironic16:36
NobodyCammorning dlaube16:38
*** andreykurilin_ has joined #openstack-ironic16:41
*** todd_dsm has quit IRC16:42
dlaubeanyone using contrail plugin with neutron/ironic?16:46
NobodyCamruby landing 12966516:47
NobodyCams/ruby/rloo/16:48
rloothx NobodyCam ;)16:48
openstackgerritJim Rollenhagen proposed a change to openstack/ironic: Add maintenance_reason when setting maintenance mode  https://review.openstack.org/12864616:48
jrollwhee, thanks for approving that spec quickly folks :)16:48
jrollnow let's get the code through :)16:49
* jroll heads out to commute16:49
*** spandhe has joined #openstack-ironic16:49
lucasagomesjroll, w00t16:51
lucasagomesdevananda, on the scheduling official slots tab, "exposing capabilities to Nova" we already do that no?16:52
lucasagomesohh or it's about driver capabilities?16:52
devanandalucasagomes: right. both.16:52
lucasagomesright ok16:53
*** tatyana has quit IRC16:57
*** igordcard has quit IRC16:58
openstackgerritA change was merged to openstack/ironic-specs: Delete the placeholder file: example.rst  https://review.openstack.org/12966516:59
*** marcoemorais has quit IRC17:00
*** igordcard has joined #openstack-ironic17:00
devanandalucasagomes: there are also some things we should talk about with Nova... adding those now17:02
lucasagomesdevananda, will take a look17:02
*** marcoemorais has joined #openstack-ironic17:03
*** athomas has quit IRC17:03
*** derekh has quit IRC17:09
*** ifarkas has quit IRC17:10
dlaubecan anyone tell me if official juno packages for ubuntu have dropped yet?17:12
NobodyCamShrews: wow I've never seen a pep8 error like that one F81117:12
NobodyCamwait thats in a file your not modifying17:13
NobodyCam./ironic/tests/conductor/test_rpcapi.py:131:5: F811 redefinition of unused 'test_get_topic_for_driver_doesnt_cache' from line 11617:13
*** bradjones has quit IRC17:14
*** harlowja_away is now known as harlowja17:14
*** zz_jgrimm is now known as jgrimm17:14
*** dtantsur is now known as dtantsur|afk17:15
*** bradjones has joined #openstack-ironic17:16
devanandadlaube: http://packages.ubuntu.com/search?suite=utopic&searchon=names&keywords=ironic17:19
devanandadlaube: they are not (and probably will not be) in Trusty, but they are in utopic17:19
*** Haomeng|2 has quit IRC17:21
NobodyCambrb17:22
*** Haomeng|2 has joined #openstack-ironic17:23
lucasagomesI will call it a day... see you guys at the meeting :)17:27
*** lucasagomes is now known as lucas-dinner17:27
NobodyCamhave a good night lucas-dinner see at the meeting17:29
NobodyCamgah... see *you* at the ...17:29
*** yjiang5 has quit IRC17:32
*** yjiang5 has joined #openstack-ironic17:32
*** marcoemorais has quit IRC17:32
*** yjiang5 has left #openstack-ironic17:34
*** yjiang5 has joined #openstack-ironic17:37
*** tonycoffman has left #openstack-ironic17:37
NobodyCamdevananda: I recall you haveing questions on https://review.openstack.org/#/c/111175 but don't see them in the comments. am I just wrong?17:37
*** yjiang5 has quit IRC17:38
dlaubedevananda: ok, thanks for confirming17:38
*** marcoemorais has joined #openstack-ironic17:38
devanandaNobodyCam: hm. I remember discussing that, but apparently I didn't comment on it17:38
devanandaI think my objection was that ironic doesn't support associating resources to tenants/projects17:39
*** yjiang5 has joined #openstack-ironic17:40
NobodyCamahh17:40
*** igordcard has quit IRC17:49
*** ndipanov is now known as ndipanov_gone17:58
*** sambetts has quit IRC18:02
ShrewsNobodyCam: umm... jenkins hates me? dunno where that came from18:03
*** todd_dsm has joined #openstack-ironic18:05
openstackgerritJim Rollenhagen proposed a change to openstack/python-ironicclient: Add node-set-maintenance command  https://review.openstack.org/12969318:10
openstackgerritJim Rollenhagen proposed a change to openstack/ironic: Clear maintenance_reason when setting maintenance=False  https://review.openstack.org/12969418:11
*** ifarkas has joined #openstack-ironic18:13
jrolllucas-dinner: ^ you should poke at those18:17
devanandajroll: just tossing this out there -- right now, using a newer version of our client against an older version of the API service results in 404 NOT FOUND18:17
devanandajroll: which is the same error it generates if the node doesn't exist18:17
devanandaour client does not support endpoint capability discovery -- in short, the API is not versioned18:18
NobodyCamjroll: fyi on 129693 : TypeError: set_maintenance() takes exactly 4 arguments (3 given)18:19
jrollNobodyCam: yeah, couldn't run tests offline, fixing things now :)18:20
jrolldevananda: hmm18:20
NobodyCam:-p18:20
jrolldevananda: wtb micro versions :)18:20
*** tatyana has joined #openstack-ironic18:22
*** pelix has quit IRC18:22
*** pelix has joined #openstack-ironic18:26
openstackgerritJim Rollenhagen proposed a change to openstack/python-ironicclient: Add node-set-maintenance command  https://review.openstack.org/12969318:33
jrollNobodyCam: fixed :)18:33
NobodyCam:)18:33
jrollum. http://logs.openstack.org/46/128646/7/check/gate-ironic-pep8/78c4a82/console.html18:34
*** tatyana has quit IRC18:34
jrollI didn't touch that18:34
jrollis our pep8 suddenly broken? :)18:34
NobodyCamjroll: thats the eact same Shrews encountered18:34
jrollall I did there was rebase 128646 on 12864518:35
jrollyeah... not good18:35
NobodyCamShrews: seem Mr. Jenkins dislikes more then you18:35
NobodyCams/seem/seems/18:35
jrollI remember the tag juno final review was weird18:35
jrollwith that method18:35
Shrewshrm18:35
jrollwhy can't I find this18:37
jrollugh.18:37
NobodyCambrb befor meeting18:38
jroll^18:39
jrollhere it is: https://github.com/openstack/ironic/blob/master/ironic/tests/conductor/test_rpcapi.py#L116-14418:40
jrollthis broke it: https://review.openstack.org/#/c/129174/18:41
jrollnot sure how18:41
jrollwait wtf18:41
jrollso this merge commit added it https://github.com/openstack/ironic/commit/deb21ef7067a157a632799a5783fe69e6de468da18:41
*** Guest52830 is now known as mgagne18:42
*** mgagne has joined #openstack-ironic18:42
jrollbut the parent did not18:42
jrollthis ain't good18:42
devananda??18:42
jrollright.18:42
*** pelix has quit IRC18:42
jrolldevananda: you want to poke infra folks, I'll make the fix commit?18:42
devanandaF811 error, yes?18:43
devanandawow18:43
jrollyes, F81118:43
devanandajroll: I'm not quite sure I understand the cause yet. feel free to poke -infra and I'll follow / jump in when I get it18:43
jrolldevananda: I don't understand it either, I was going to delegate that to infra18:43
jrollall I know is a merge commit added code that was not in the commit it was merging18:44
openstackgerritJim Rollenhagen proposed a change to openstack/ironic: Clear maintenance_reason when setting maintenance=False  https://review.openstack.org/12969418:44
*** andreykurilin_ has quit IRC18:45
lucas-dinnerjroll, awesome! will do18:46
jrolldevananda: making a bug at least18:46
devanandaserious wtf18:46
devanandathe git log attributes the COMMIT to jenkins18:46
Shrewsjroll: so, that parent was for proposed/juno, and the snmp one was on master18:46
Shrewsnow, i'm not sure how that happened, or what that means, but doesn't seem good18:46
jrollhuh18:46
devanandagit blame ironic/tests/conductor/test_rpcapi.py | grep "131)"18:46
devanandadeb21ef7 ironic/tests/conductor/test_rpcapi.py (Jenkins                2014-10-20 15:35:46 +0000 131)     def test_get_topic_for_driver_doesnt_cache(self):18:46
jrollShrews: ttx added a default branch of proposed/juno18:46
jrollI thought18:47
jrollmaybe not18:47
Shrewsi mean, it seems like 129174 was branched off a non-master branch18:47
Shrewsbut proposed (and merged) to master18:48
jrolloh18:48
* Shrews throwing out possibilites18:48
jrollstrange that unit test didn't make proposed/juno18:48
jrollbut I remember a wtf proposal for stable juno that removed that18:48
NobodyCamjroll: you happen to have a link for that?18:49
jrollNobodyCam: for what?18:49
NobodyCamthe "wtf proposal for stable juno"18:50
NobodyCam:0p18:50
jrollNobodyCam: no, I can't find it :(18:50
openstackgerritJim Rollenhagen proposed a change to openstack/ironic: Fix F811 error in pep8  https://review.openstack.org/12970218:50
*** BertieFulton has joined #openstack-ironic18:50
jrolloh wait18:50
jrollhttps://review.openstack.org/#/c/128965/18:50
jrollbut that didn't merge18:51
jrollmight be related though18:51
jrollanyway, I have to brb18:51
jrollhttps://review.openstack.org/129702 will fix the gate at least18:51
Shrews+2'd18:51
Shrewsoh, so did lucas-dinner18:51
lucas-dinneryeah just bumped inot it18:52
lucas-dinnerinto it*18:52
*** lucas-dinner is now known as lucasagomes18:52
Shrewslucasagomes: i didn't +A it. doing so now18:52
*** jgrimm is now known as zz_jgrimm18:52
lucasagomes:)18:52
NobodyCamShrews: not waiting for jenkins?18:52
ShrewsNobodyCam: of course i am18:53
Shrewsnow...18:53
Shrews:-P18:53
NobodyCamlol18:53
Shrewshe wouldn't merge it anyway if was still broke, but i don't trust him now18:54
*** pensu has quit IRC18:54
ShrewsMr. J, that is18:54
*** jcoufal has quit IRC18:55
*** rushiagr_away is now known as rushiagr18:57
*** praneshp has joined #openstack-ironic18:58
mrdaMorning Ironic18:58
NobodyCammorning mrda18:58
mrda\o18:58
*** dtantsur|afk is now known as dtantsur19:00
*** linggao has joined #openstack-ironic19:00
NobodyCammeet'n time19:01
mrdarloo: This year was an especially long holiday.  It isn't normal :)19:01
rloomrda: ohh, too bad :-)19:02
rloomrda: that it isn't normal ;)19:02
mrda:)19:02
*** zz_jgrimm is now known as jgrimm19:08
*** rushiagr is now known as rushiagr_away19:13
*** andreykurilin_ has joined #openstack-ironic19:23
*** ifarkas has quit IRC19:24
*** pensu has joined #openstack-ironic19:26
*** rakesh_hs has quit IRC19:44
*** igordcard has joined #openstack-ironic19:50
*** dprince has quit IRC19:58
lucasagomesdevananda, I will def do the spec for the ipxe... but I was thinking about extending the vendor_passthru as a separated feature, like a plumbing work to pave the way to do the iPXE stuff20:00
*** linggao has quit IRC20:00
lucasagomesso 2 specs?20:00
NobodyCamlucasagomes: I think that would be a great help20:00
jrolllucasagomes: yeah, I think there's going to be enough change/contention to get a spec up20:01
mrdarloo: fwiw, I'm also interested in async api20:01
lucasagomesNobodyCam, alright :)20:01
NobodyCamlucasagomes: I'm ok with one spec and maybe two patches20:01
jrollif you had to convince deva it's reasonable, probably needs a spec :P20:01
lucasagomesjroll, NobodyCam right, thanks... i will put 2 specs one for each feature20:01
jrollcool, thanks man20:01
lucasagomesheh true20:01
* jroll bbiab20:01
NobodyCam:) sweet20:01
devanandajroll: lol, now i'm your litmus test for specs? :p20:01
jrolldevananda: one of many20:01
jroll:D20:01
lucasagomesI mean, I know it can be misused, but I don't want to also not allow it because people can do good things with that20:02
rloomrda: i can't even remember why that stalled, but if things need to be discussed, we should try to do at the summit.20:02
devanandalucasagomes: so in general, I have a lot of concerns for the future of vendor passthru20:02
lucasagomesdevananda, right... for the v2 api u mean?20:02
lucasagomesor right now in v1?20:02
devanandathey might be ungrounded - after all, we can't be responsible for what a downstream driver does, but we are responsible for maintaining backwards compatibility20:03
NobodyCamoh gawd v2 api... that needs a slot :-p20:03
devanandaand that API endpoint is opaque20:03
lucasagomesdevananda, yeah, but it was designed like that20:03
devanandalike, there's currently no way to discover what methods a driver supports20:03
devanandaexactly20:03
devanandathe intent was: let drivers add new things that should become part of the standard20:03
mrdaNobodyCam: we have a pod right? There can be more discussions20:04
devanandawhat you're proposing here is never going to be part of the standard AIUI20:04
lucasagomesdevananda, yeah... right now, if we add that basic mechanism that exposes the routes for the vendor methods we could make it discoverable20:04
devanandamrda: yes20:04
Shrewsrloo: it stalled b/c the need for it was somewhat alleviated since we added the node lock retry logic. i honestly don't know if we still want it or not.20:04
lucasagomeshttps://review.openstack.org/#/c/129261/20:04
NobodyCammrda: YES!!!!!20:04
devanandalucasagomes: i mean, any sort of "get the ipxe config" request is, by definition, not generalizable20:04
devanandalucasagomes: it's specific to that driver and always will be20:04
lucasagomesdevananda, yes, that's why it goes to the driver vendor passthru20:04
lucasagomescause it's a unique capability20:05
lucasagomesit's not part of the general abstraction20:05
lucasagomesjust like the pass_deploy_info for pxe right now20:05
NobodyCamdevananda: the goal was to consolidate the vender pass thru based on a majority of the drivers having the "same" function in VPT20:05
devanandalucasagomes: pass deploy info is a node vendor-passthru, not driver20:06
lucasagomesdevananda, right yeah, that's correct... but it comes with the pxe driver20:06
lucasagomesthis is another thing that I have in mind, this separation of driver vendor passthru or node vendor passthru makes little sense20:06
devanandaactually yes, it falls into the same category of never going to becmoe standardized20:06
lucasagomesmaybe "node" should be only a parameter for the driver vendor passthru20:06
lucasagomesbut the idea is, if a driver needs something that is not possible to abstract, it goes to one of those vendor endpoints, which sounds like a good approach to me20:07
NobodyCambrb20:08
*** BertieFulton has quit IRC20:09
*** harlowja is now known as harlowja_away20:10
devanandalucasagomes: looking at ilo and drac, which is where I'd expect to see vendor_passthru do something truly unique, I don't20:11
lucasagomesdevananda, right, well for drac at least we wanted to have things like RAID config etc20:11
lucasagomeswe had specs for it but wasn't the right time20:11
devanandaright20:11
lucasagomeswe also have ideas like the ready state to do using the vendor passthru20:11
lucasagomeswhich is talking to the BMC and populating things in Ironic20:12
devanandaI believe HP plans to do more unique things there with iLO as well20:12
lucasagomesbut last cycle we only could add the base driver20:12
lucasagomesI'm sure things will come20:12
devanandaand seamicro did a while back (attach volume, for example)20:12
lucasagomesyeah, exactly20:12
devanandaso from an API perspective, what differentiates these?20:12
lucasagomesdevananda, these are the things that I think that, once more than 1 driver starts implementing it20:12
lucasagomeswe should promote to the normal api20:13
lucasagomescause they are common across multiple driver as I see it20:13
lucasagomesbut as a first stage, I think that having it on the vendor_passthru is fine20:13
devanandasure20:13
lucasagomeswhen no one else expose that capability20:13
devanandasorry, i'm not doing a good job explaining my POV20:14
devanandait seems like two categories of methods have emerged within vendor_passthru: 1. cool hardware stuff that we want to become standardized, 2. driver idiosyncracies20:16
lucasagomesright, yeah20:17
jrollI mean, reality is that drivers may need API endpoints that are not generalizable20:19
jrollPXE driver has one20:19
jrollIPA has one20:19
jrolliPXE driver will have one20:19
jrollIPA actually has multiple now that I think about it, but my point still stands20:19
devanandajroll: indeed, but should those endpoints be lumped together with generalizable ones?20:20
lucasagomesjroll, right, and things like lookup and heartbeat is very specific for the IPA right?20:21
devanandawhat makes POST /v1/node/NNNN/vendor_passthru?method=attach_volume20:21
devanandadifferent from /v1/node/NNNN/vendor_passthru?method=continue_deploy20:21
dtantsurso we have "technical" endpoints that are never exposed to the end user20:21
jrolllucasagomes: yes20:21
dtantsurand "public" that just provide additional functionality, that is not general20:21
devanandadtantsur: right20:21
jrolldevananda: I don't see your point20:22
dtantsurdevananda, btw all these "technical" endpoints are node-less now, right?20:22
devanandajroll: attach_volume is a hardware capability, enabled by one driver today, which more drivers will hopefully soon also support, at which point it completes the intended purpose of vendor_passthru and migrates to a common interface20:22
*** HenryG has quit IRC20:23
devanandajroll: continue_deploy exists for a different purpose. it has nothing to do with vendor support20:23
dtantsurhell, I feel like I'm close to some idea, but I'm too tired to do it :( It would be really nice to discuss f2f. and g'night for now.20:23
*** dtantsur is now known as dtantsur|afk20:23
jrolldevananda: ok, sure. what's your point? should we add /v1/node/NNNN/driver_weirdness?20:23
lucasagomesdtantsur, no, heartbeat and continue_deploy is node vendor passthru20:23
devanandajroll: ditto for agent heartbeat and so on. the reason those endpoints exist has nothing to do with enabling unique hardware or standardizing common functions across drivers20:23
NobodyCamnight dtantsur|afk20:23
lucasagomesdtantsur|afk, night20:24
jrolldevananda: right, but we need them. what do you suggest?20:24
jrollnight dtantsur|afk20:24
devanandadtantsur|afk: cheers. I think you're close to it, but I dont quite hve it either yet. looking forward to Paris!20:24
lucasagomesdevananda, oh I think I see where it's leading... maybe we should have another endpoint for it?20:24
devanandathese are essentially remote procedure calls, right?20:25
lucasagomesor extending the scope of the vendor_passthru20:25
devanandanot a user or operator facing API endpoint20:25
lucasagomesmaybe even renaming it to only "passthru"20:25
lucasagomesright20:25
devananda"continue deploy" is not a verb in our API20:25
devanandait's an RPC call20:25
devanandaheartbeat is much closer to a verb, fwiw, so I"m not sure where that falls in my thinking yet20:26
jroll"lookup" is not an rpc call, fwiw20:26
jroll(imo)20:26
devanandajroll: indeed. and I think I wanted that to be done by searching20:26
jrollright, but then we're going to end up with some endpoint20:26
devanandajroll: but ya'll had a reason not to use GET /v1/nodes for that20:26
devanandaI dont remember why20:27
jroll"lookup" is a little overloaded in that it returns some configuration20:27
jrollheartbeat timeout configs, etc20:27
devanandaoh, right20:27
jrollI'm going to +A https://review.openstack.org/#/c/129702/ btw20:27
devanandaretrieve some driver-specific config bits20:27
jrollright, which is the same thing ipxe is doing20:27
devanandajroll: /me looks20:27
jrolladam_g: https://review.openstack.org/#/c/129702/ failed parallel tests btw20:27
adam_gjroll, thanks20:28
jrolladam_g: all I did was remove a duplicate unit test, so that's not a real failure20:28
devananda+A'd20:28
jrolloh I already did20:28
lucasagomesyeah I'm good +a that20:28
jrollbut thanks :P20:28
adam_gjroll, yeah.. one common thing ive found with failures is one of the VMs never reboots out of the deploy ramdisk: http://logs.openstack.org/02/129702/1/check/check-tempest-dsvm-ironic-parallel-nv/3e35154/logs/ironic-bm-logs/baremetalbrbm_0_console.txt.gz20:28
jrolladam_g: odd20:29
adam_ghard to say if thats a libvirt/host issue or something in our ramdisk20:29
NobodyCamstrange20:29
adam_gi have a stack of devstack patches up to better monitor console logs20:29
adam_gright now they get truncated with every power cycle20:29
NobodyCamadam_g: are they by change running exec ?20:29
adam_gNobodyCam, huh? kexec?20:30
NobodyCamyea that too :-p20:30
jrolladam_g: wait...20:30
jrolladam_g: aren't we just calling libvirt power off/on20:30
adam_gNobodyCam, nah, just a regular reboot20:30
jrollor libvirt reboot20:31
* jroll looks20:31
lucasagomesdevananda, jroll right it's a bit late here, I'll call it a day. Thanks for the discussion20:31
adam_gjroll, not there, IIRC the deploy ramdisk is rebooting itself20:31
NobodyCamhttps://github.com/openstack/diskimage-builder/tree/master/elements/deploy-kexec20:31
jrolllucasagomes: cheers, have a good one20:31
devanandalucasagomes: cheers, g'night20:31
NobodyCamnight lucasagomes20:31
jrolladam_g: I don't think that's a thing?20:31
jrolladam_g: how would it know when to reboot?20:31
jrollI think it's https://github.com/openstack/ironic/blob/master/ironic/drivers/modules/ssh.py#L15720:31
lucasagomesdevananda, I see ur point that we may be missusing the vendor_passthru, maybe we should think about extending it's scope for the idiosyncracies part too20:32
lucasagomesbut we can talk about it later20:32
adam_gjroll, ah, right20:32
jrollI think I'd blame libvirt but idk20:33
jrolloh20:33
jrollmaybe it's like an ACPI reboot20:33
*** lucasagomes is now known as lucas-zZz20:33
jrolland disks aren't unmounting cleanly or something20:33
adam_gi dont think there are any disks mounted other than the ramfs20:35
adam_gand it looks liek the ACPI reboot is servied completely, [   52.315117] reboot: machine restart20:35
adam_glibvirt wonkiness isn't unusual so its possible we're just tickling something occasionally20:36
jrollI see20:36
adam_gwish there was something we could match on for an elastick-recheck query20:37
*** pensu has quit IRC20:39
*** foexle has quit IRC20:52
NobodyCamhow specific should our operator questions be? I am attempting to keep them mostly general (see: https://etherpad.openstack.org/p/Ironic_Questions)21:00
NobodyCamj* do you have any IPA feed back you looking to get for the Operators?21:01
NobodyCamgah21:01
*** andreykurilin_ has quit IRC21:01
NobodyCams/you/your/ : s/for/from21:02
*** andreykurilin_ has joined #openstack-ironic21:04
jrollNobodyCam: nothing specific21:07
NobodyCam:)21:08
*** andreykurilin_ has quit IRC21:13
*** andreykurilin_ has joined #openstack-ironic21:14
-openstackstatus- NOTICE: Zuul erroneously marked some changes as having merge conflicts. Those changes have been added to the check queue to be rechecked and will be automatically updated when complete.21:21
NobodyCamok dumb question.. this is our offical mission statment? http://git.openstack.org/cgit/openstack/governance/tree/reference/programs.yaml#n14621:26
NobodyCamdevananda: are those along the lines of the types of questions your thinking of? /me notes they not ordered at this point their just a breain dump21:32
NobodyCambrain dump even :-p21:33
devanandaNobodyCam: yes, that's the official mission statement21:33
devanandaNobodyCam: sort of. I'll add a few21:33
NobodyCam:) fire away21:33
*** spandhe has quit IRC21:34
*** spandhe has joined #openstack-ironic21:36
*** openstackgerrit has quit IRC21:40
NobodyCambrb... quick walkies :)21:46
*** todd_dsm has quit IRC21:46
*** jmank has joined #openstack-ironic21:48
*** marcoemorais has quit IRC21:48
*** marcoemorais has joined #openstack-ironic21:49
*** todd_dsm has joined #openstack-ironic21:52
* NobodyCam is back21:55
*** marcoemorais has quit IRC21:58
*** marcoemorais has joined #openstack-ironic21:59
*** marcoemorais has quit IRC21:59
*** marcoemorais has joined #openstack-ironic21:59
devanandaJayF: any followup to our chat last week // https://etherpad.openstack.org/p/hardware-capabilities-spec ?22:04
JayFdevananda: it's on my list for today, but tbh haven't had a productive day. It's close to the top of my list right now though22:05
devanandaJayF: np, just checkin22:05
* devananda heads out to a cafe for a bit22:06
NobodyCamenjoy22:09
*** marcoemorais has quit IRC22:10
*** igordcard has quit IRC22:15
*** HenryG has joined #openstack-ironic22:16
*** lucas-zZz has quit IRC22:20
*** eghobo has joined #openstack-ironic22:22
*** eghobo has quit IRC22:29
*** eghobo has joined #openstack-ironic22:32
*** jgrimm is now known as zz_jgrimm22:37
*** jmank has quit IRC22:40
*** openstackgerrit has joined #openstack-ironic22:43
*** marcoemorais has joined #openstack-ironic22:43
*** eghobo has quit IRC22:44
openstackgerritA change was merged to openstack/ironic: Fix F811 error in pep8  https://review.openstack.org/12970222:46
*** harlowja_away is now known as harlowja22:48
*** eghobo has joined #openstack-ironic22:48
*** andreykurilin_ has quit IRC22:49
*** eghobo has quit IRC22:49
yjiang5NobodyCam: hi22:51
NobodyCamhi yjiang5 :)22:51
*** eghobo has joined #openstack-ironic22:53
yjiang5I'm checking https://bugs.launchpad.net/ironic/+bug/1382164 and seems Dmitry is right that the check_image_size() is using the converted_image to compare, thus should be ok. According to your log, seems the error happens in converting to raw image?22:53
NobodyCamyea22:53
NobodyCamso we had the space when the image was downloaded but then lost it22:54
*** marcoemorais has quit IRC22:55
NobodyCamchecking with the folks who actually had the issue. it seems that another service (sphinx, or another indxer thing) started and sucked up the disk space22:56
yjiang5NobodyCam: Yes, is your machine (or the partition holding the image cache) already very crowded already (not sure if 'crowd' is correct, not well on english). Is it possible that during the check cache size and the real convert, something wrong to occupy the disk.22:56
yjiang5NobodyCam: yes, currently the image_cache does not gurantee/reserve any disk space.22:56
NobodyCamya22:57
yjiang5NobodyCam: Just wondering if force_raw_images is rquired to be True if we are using iscsi deploy?22:58
NobodyCammight be worth adding a check for diskspace right before we actually do the convert to raw. but even that is no gurantee22:59
NobodyCamyjiang5: I'm not sure there, I think there was a reason for that.23:00
yjiang5NobodyCam: The window still exists, unless we can reserve the space , I will try to check if such tools on linux side. Another questioin is, would it be possible to avoid the convert at all. Convert the image is really not so good because it's assumed that the ironic-conductor machine is sure to have bigger disk size than the converted image, or the compute node disk in some situation.23:01
NobodyCamyjiang5: yes. I think attempting to reserve the space is the wrong way to go, I would love to remove that convert or do the convert on the fly when we are writting the image to the node23:04
yjiang5NobodyCam: Yes, I will try to check if that's possible.23:04
NobodyCamawesome Thank you yjiang5 :)23:05
*** marcoemorais has joined #openstack-ironic23:05
yjiang5NobodyCam: :)23:07
*** praneshp has quit IRC23:08
*** todd_dsm has quit IRC23:24
devanandadtantsur|afk: I'm reading the code for discoverd and I'm pretty excited that you were able to do this in a separate service23:27
NobodyCamdevananda: On the question you added is "it" in the last part in refference to Ironic or the pxe deploy ?23:30
*** praneshp has joined #openstack-ironic23:47
*** todd_dsm has joined #openstack-ironic23:48
« Sunday, 2014-10-19 Index Tuesday, 2014-10-21 »

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