Monday, 2018-12-17

« Sunday, 2018-12-16 Index Tuesday, 2018-12-18 »
*** dsneddon has joined #openstack-ironic00:16
*** dsneddon has quit IRC00:21
*** dsneddon has joined #openstack-ironic00:52
*** tiendc has joined #openstack-ironic01:33
*** oanson has quit IRC01:59
*** bfournie has joined #openstack-ironic02:09
openstackgerritMerged openstack/networking-generic-switch master: Add pycharm .idea project dir to gitignore  https://review.openstack.org/62521403:16
openstackgerritKaifeng Wang proposed openstack/ironic master: Expose conductors: api-ref  https://review.openstack.org/62286203:32
*** bfournie has quit IRC03:36
*** baha has joined #openstack-ironic04:32
*** coreywright has quit IRC04:35
*** baha has quit IRC04:37
*** coreywright has joined #openstack-ironic05:02
*** rcernin has joined #openstack-ironic05:46
*** rcernin has quit IRC05:47
*** rcernin has joined #openstack-ironic06:09
*** rcernin has quit IRC06:09
*** rcernin has joined #openstack-ironic06:09
*** rcernin has quit IRC06:09
*** dnuka has joined #openstack-ironic07:14
dnukagood morning ironic07:14
*** jtomasek has joined #openstack-ironic07:33
*** jtomasek has quit IRC07:33
*** Emine has joined #openstack-ironic07:33
*** jtomasek has joined #openstack-ironic07:33
*** oanson has joined #openstack-ironic07:34
*** e0ne has joined #openstack-ironic07:35
*** e0ne has quit IRC07:39
*** dnuka has quit IRC07:43
*** dnuka has joined #openstack-ironic07:45
*** yolanda has joined #openstack-ironic07:50
*** rpittau has joined #openstack-ironic07:58
rpittaugood morning ironic o/07:58
dnukamorning rpittau :)08:00
rpittauhi dnuka :)08:00
*** pcaruana has joined #openstack-ironic08:18
openstackgerritRiccardo Pittau proposed openstack/networking-baremetal master: [ DNM/TEST ] testing zuulv3 jobs  https://review.openstack.org/62521108:19
*** gkadam has joined #openstack-ironic08:35
*** gkadam has quit IRC08:35
*** iurygregory has joined #openstack-ironic08:43
iurygregorygood morning o/08:43
dnukamorning iurygregory o/08:44
rpittauhi iurygregory :)08:56
*** dougsz has joined #openstack-ironic09:04
*** etingof has joined #openstack-ironic09:24
*** derekh has joined #openstack-ironic09:58
*** lucasagomes has joined #openstack-ironic09:59
*** e0ne has joined #openstack-ironic10:02
*** dnuka has quit IRC10:15
*** dnuka has joined #openstack-ironic10:21
*** sambetts_ has joined #openstack-ironic10:29
*** derekh has quit IRC10:46
*** derekh has joined #openstack-ironic10:47
openstackgerritMark Goddard proposed openstack/ironic-python-agent master: Applying the HTTPS protocal in project.yaml  https://review.openstack.org/61780010:50
*** dsneddon has quit IRC10:52
*** iurygregory is now known as iury|lunch11:00
openstackgerritMerged openstack/ironic-specs master: Update tox minversion to 2.0  https://review.openstack.org/61328111:00
openstackgerritMerged openstack/tenks master: Changes following becoming an official project  https://review.openstack.org/61958511:06
openstackgerritparesh sao proposed openstack/ironic master: OOB RAID implementation for ilo5 based HPE Proliant servers.  https://review.openstack.org/55566411:08
openstackgerrityolanda.robla proposed openstack/ironic master: WIP: Add pxe template per node  https://review.openstack.org/62555311:16
openstackgerritMerged openstack/tenks master: Fix deploy and teardown without node enrolment  https://review.openstack.org/61962611:28
*** dsneddon has joined #openstack-ironic11:28
openstackgerritMerged openstack/tenks master: Fail nicely if ironic port lookup fails  https://review.openstack.org/62004811:29
openstackgerritMerged openstack/tenks master: Pass node and volume name prefixes to tenks_update_state  https://review.openstack.org/62004911:30
openstackgerritMerged openstack/tenks master: Fix skipping glance check if deploy images not defined  https://review.openstack.org/62005011:30
openstackgerritMerged openstack/tenks master: Use only the first 6 characters of the node name in link names  https://review.openstack.org/62005111:30
*** e0ne has quit IRC11:31
*** jaypipes has joined #openstack-ironic11:33
*** dougsz has quit IRC11:36
openstackgerritRiccardo Pittau proposed openstack/networking-baremetal master: [ DNM/TEST ] testing zuulv3 jobs  https://review.openstack.org/62521111:37
*** dougsz has joined #openstack-ironic11:38
*** iury|lunch is now known as iurygregory11:45
*** dsneddon has quit IRC11:45
*** tiendc has quit IRC11:45
*** sri_ has joined #openstack-ironic11:51
*** dsneddon has joined #openstack-ironic12:00
*** rpittau is now known as rpittau|lunch12:09
openstackgerrityolanda.robla proposed openstack/ironic master: WIP: Add pxe template per node  https://review.openstack.org/62555312:09
*** diga has joined #openstack-ironic12:20
*** dnuka has quit IRC12:28
openstackgerrityolanda.robla proposed openstack/ironic master: WIP: Add pxe template per node  https://review.openstack.org/62555312:37
*** dnuka has joined #openstack-ironic12:38
*** e0ne has joined #openstack-ironic12:47
*** e0ne has quit IRC13:11
*** rpittau|lunch is now known as rpittau13:11
*** e0ne has joined #openstack-ironic13:14
*** Bhujay has joined #openstack-ironic13:15
*** bfournie has joined #openstack-ironic13:25
*** dnuka has quit IRC13:26
*** Bhujay has quit IRC13:33
*** rh-jelabarre has joined #openstack-ironic13:33
*** e0ne has quit IRC13:42
*** e0ne has joined #openstack-ironic13:45
digaHi Guys13:47
iurygregoryhello13:47
*** pcaruana has quit IRC13:50
jiapeiGood morning everyone13:51
iurygregorymorning jiapei o/13:57
openstackgerritRiccardo Pittau proposed openstack/networking-baremetal master: [ DNM/TEST ] testing zuulv3 jobs  https://review.openstack.org/62521113:58
rpittauhi jiapei :013:58
rpittau:)13:58
digaI am trying to backport this patch to https://review.openstack.org/#/c/617951/ to stable/ocata and stable/pike release but getting error14:04
patchbotpatch 617951 - ironic - Fix OOB introspection to use pxe_enabled flag in i... (MERGED) - 17 patch sets14:04
digaError - Chery pick failed - merge conflict14:05
digaCan someone help me on this ? I am doing this from UI14:05
*** pcaruana has joined #openstack-ironic14:05
*** baha has joined #openstack-ironic14:09
*** dnuka has joined #openstack-ironic14:10
digaiurygregory: Hi14:10
iurygregorynot sure if the backport is possible, probably had huge code changes during ocata -> stein14:10
iurygregoryif any core is around they will prob know =)14:11
iurygregorybut most of them are on PTO14:11
digaokay14:11
digahmm14:11
digaI am able to successful create patches for rocky and queens relealse14:12
digagetting error for pike and ocata only14:12
digaLooks like a big task :)14:12
rpittaudiga, I'm pretty sure if that backport is even possible, you will have to resolve a lot of merge conflicts, doing that from the UI is not recommended14:12
digarpittau: I am not aware about doing it from backend14:13
digarpittau: do you have any link how to do it ?14:13
*** rloo has joined #openstack-ironic14:14
rpittaudiga, https://docs.openstack.org/project-team-guide/stable-branches.html14:16
digarpittau: will take a look14:16
iurygregorydiga, also try to ping some cores to see if the patch is backportable =)14:17
digaiurygregory: sure14:18
digaLooks like nobody from core is around, I will ping them later14:18
iurygregoryyeah14:18
iurygregoryrpittau, pxe_ipmitool should be ipmi only no? https://review.openstack.org/#/c/625211/5..6/zuul.d/networking-baremetal-jobs.yaml14:19
patchbotpatch 625211 - networking-baremetal - [ DNM/TEST ] testing zuulv3 jobs - 6 patch sets14:19
*** moshele has joined #openstack-ironic14:20
*** mmethot has joined #openstack-ironic14:22
rpittauyes, that was actually a mistake, it's not even starting the daemon14:22
iurygregoryrpittau, it happens XD14:23
jrollmorning everyone14:26
rpittauhi jroll :)14:27
jrolldiga: iurygregory: that looks backportable to me14:28
iurygregorymorning jroll o/14:28
jroll\o14:28
iurygregorydiga, you problably will need to do manualy since you are getting error from UI like rpittau said =)14:29
*** dnuka has quit IRC14:30
*** toabctl has quit IRC14:57
*** toabctl has joined #openstack-ironic14:57
TheJuliaGood morning everyone14:58
rpittauhi TheJulia :)14:59
TheJuliadiga: best advice I can give you is to try and resolve the merge conflict as you would any other merge conflict on stable/pike. Update the commit message to include that there was a merge conflict, and once we've landed that then backport that to stable/ocata15:00
iurygregorymorning TheJulia o/15:00
*** beekneemech is now known as bnemec15:00
*** e0ne has quit IRC15:02
*** moshele has quit IRC15:03
jiapeio/15:04
* iurygregory thought that TheJulia would be on PTO =)15:06
*** e0ne has joined #openstack-ironic15:06
TheJuliaso, if I were to start PTO to use all my time, I would have had to have started.... last tuesday15:07
iurygregoryyou can transfer everything to next year? O.o15:07
TheJuliaI might take a day or two off this week15:07
iurygregorygot it =)15:07
TheJuliaiurygregory: Laws where I live let me hold on to it for another three months15:07
iurygregoryTheJulia, awesome \o/15:08
TheJuliaI have some high priority discussions and I'm working on a bit a massive spec this week.15:09
iurygregorygood luck o/15:09
*** coreywright has quit IRC15:13
TheJuliajroll: replied with my context to your question on https://review.openstack.org/#/c/62441315:19
patchbotpatch 624413 - ironic-specs - Support for Software RAID - 1 patch set15:19
TheJuliaarne_wiebalck: You should likely chime in as well and provide your perspective ^^^15:20
* TheJulia cleans up the whiteboard15:21
jrollTheJulia: good context, but doesn't answer my question :) is it some sort of default, or will layouts that are not a RAID 1 + RAID N be blocked? is "just one RAID 1" allowed?15:22
TheJuliaThat was the part of the question that I suspect you were getting at which is why I wanted arne_wiebalck to chime in15:23
TheJuliaor better question is "what do we want it to be"?15:23
TheJuliaAny raid config causes an automatic raid1 would make sense to me15:23
TheJuliafor this driver15:23
*** derekh has quit IRC15:24
arne_wiebalckTheJulia: Will do, planned to do so since the morning, need to finish off some other stuff first, though :-/15:24
digaTheJulia: Thank you15:25
digawill do that same15:25
TheJuliaat least CI mellowed somewhat last week, at least from my perception15:25
jrollTheJulia: okay, thanks, I'll post on the review15:25
*** coreywright has joined #openstack-ironic15:25
*** derekh has joined #openstack-ironic15:27
*** pcaruana has quit IRC15:29
TheJuliaI'm going to +A https://review.openstack.org/#/c/504952/19 in half an hour.. last chance!15:30
patchbotpatch 504952 - ironic-specs - Deploy Templates - 19 patch sets15:30
arne_wiebalckjroll: The current proposal is to have RAID-1 + RAID-N always.15:30
arne_wiebalckjroll: Only RAID-1 would work, but only RAID-0 would be more complicated15:30
jrollarne_wiebalck: what if I just want RAID 1? and is this a default, or configurable (e.g. which disks are which), etc?15:31
arne_wiebalckHence, we thought it’d be simpler to always have a RAID-1 for the system.15:31
TheJuliaI think raid-0 as root is not really viable unless the spacing was huge and... yeah15:31
jrollyep, all makes sense. just logistics questions I'm trying to answer.15:31
* TheJulia tries to think of the right word for spacing15:31
TheJuliastripe15:31
arne_wiebalckjroll: if you only want one huge RAID-1, the current proposal will give you two RAID-1s :)15:31
arne_wiebalckjroll: The idea is to avoid boot issues and have sth that looks the same, independent from the RAID level you want.15:32
arne_wiebalckThe RAID-1 should be small, up to 100GB maybe.15:33
arne_wiebalckIt should not eat much of your disks.15:33
arne_wiebalckjroll: ATM, the system RAID-1 is not configurable.15:34
jrollarne_wiebalck: yes, understood. I'm asking for more info on what this looks like from the user perspective.15:34
arne_wiebalckjroll: Ah, ok.15:34
jrollarne_wiebalck: the size of the system RAID 1 isn't configurable? why not?15:34
arne_wiebalckjroll: No reason, just b/c it was simpler.15:34
arne_wiebalckjroll: We had to change it once already, so I wsh it was configurable :-D15:35
TheJuliajroll: also upstreaming, so we could potentially make it configurable as time goes on if we don't want to outright15:35
TheJuliaoh, well, that is a good reason to make it configurable then15:35
arne_wiebalckTheJulia: Should be simple15:35
TheJuliayup15:35
arne_wiebalckTheJulia: jroll: I planned to make this configurable.15:35
arne_wiebalckAll of this is not user configurable anyway, as these are node properties.15:36
arne_wiebalckAnd this is done during cleaning.15:36
jrollerm15:36
arne_wiebalckNo?15:36
jrollwell,15:36
arne_wiebalckIt’s someone with admin privs who would need to set this.15:36
arne_wiebalckMAybe I’m missing sth .15:36
TheJulianode properties from a standpoint of matching capability content?15:36
jrollwe have a mechanism for providing RAID already. I think we should stay with that.15:37
TheJuliafrom custom filters?15:37
arne_wiebalckjroll: target_raid_config?15:37
jrollI don't think we should have a significantly different method for doing software RAID.15:37
jrollarne_wiebalck: yes, that15:37
arne_wiebalckCan the user set this?15:37
jrolloh, we're getting the word "user" mixed up :)15:37
arne_wiebalckI thnk so :-D15:37
arne_wiebalckUser == person who creates an instance.15:38
arne_wiebalck(for me)15:38
jrollsorry, when I said "user perspective" before, I meant the API/CLI user, i.e. the admin.15:38
jrollright right15:38
jrollmy fault :)15:38
arne_wiebalckjroll: ah, ok!15:38
arne_wiebalckYes, the admin can set this, of course.15:38
TheJuliaI guess my next question is if target_raid_config == {} and the raid interface is enabled, will a raid1 just be created by default?15:38
arne_wiebalckAnd it’s using the target_raid_config.15:38
arne_wiebalckjroll: No.15:38
jrollok, excellent15:39
arne_wiebalckIf the target_raid_config is emtpy, nothing will be done.15:39
TheJuliaokay15:39
jroll++15:39
TheJuliaThat seems super reasonable to me15:39
arne_wiebalckAnd the standard deploy will happen.15:39
* TheJulia is onboard15:39
arne_wiebalckFor the “system RAID-1“: I think grub can understand more complex RAID layouts.15:40
jrollarne_wiebalck: when you have some time, do you mind taking "predefined layout" on line 62 of the spec, and changing that to talk about what a valid layout is instead? for example, a RAID 1 must be defined (and should be small, but not required), and a RAID-N for data may be defined, anything else will error?15:40
arne_wiebalckjroll: In the current proposal, a RAID-N is mandatory.15:42
arne_wiebalcktarget_raid_config can take multiple RAIDs, no?15:42
jrollarne_wiebalck: why? what if I only have two small SSDs?15:42
jrollyes, it can15:42
arne_wiebalckeven on two small SSDs you’ll need to store the system somewhere15:43
arne_wiebalckand it does not reduce space15:43
jrollright, the RAID 1 is for the system, right?15:43
arne_wiebalckbut15:43
arne_wiebalckyes15:43
arne_wiebalckbut15:43
arne_wiebalckwe could change it15:43
jrollwell... let's define system. / or /boot?15:44
arne_wiebalckand make the target raid config have the RAID-1 and the RAID-N15:44
arne_wiebalckand make the RAAID-N optional.15:44
TheJuliaarne_wiebalck: Yeah, there are raid0/5/6 modules, but I think they assume default stripes for linux and md deices, not lvm raid15:44
TheJuliaso re /boot, most cloud images do not mount it and overwrite it...15:44
jrollto be clear, I'm not opposed to mandatory RAID-N, but I want to understand why15:44
arne_wiebalckTheJulia: Yeah, i was suspecting that otehr images may do different things from our image.15:45
TheJuliajroll: your raising great questions! :)15:45
arne_wiebalckjroll: ATM, the admin would only define a RAID-N.15:45
arne_wiebalckin the target raid config15:45
arne_wiebalckand the RAID-1 would be there automatically15:46
arne_wiebalckto avoid boot issue if the RAID-N is a RAID-0, for instance15:46
arne_wiebalcks/issue/issues/15:46
jrollI see. hm.15:46
arne_wiebalckWhat we could do is allow: (RAID-1 + RAID-N) or (RAID-1)15:47
arne_wiebalckWhere N != 1 … if that makes sense :)15:47
jrollyeah, I think being explicit is best here15:48
arne_wiebalckOur basic uses cases are RAID-1 (for services/pets) and RAID-0 (for batch/cattle)15:48
arne_wiebalckjroll: I agree15:48
jrollright, makes sense15:48
jrollI assume there's some RAID-10 use cases for databases and hypervisors out there15:48
arne_wiebalckYes, we have RAID-10 now as well.15:49
arne_wiebalckAlso for services/pets.15:49
jrollyep15:49
TheJuliajroll: so one thing that I do like about this is that it is moderately opinionated, , but I do agree with you at the same time15:49
arne_wiebalckPutting the “system RAID-1” into the target raid config makes it explicit and configurable.15:49
arne_wiebalckI think that’s a good enhancement to our opiniated first shot :)15:50
jrollTheJulia: opinionated is good! I just prefer to be explicit and error when the user's opinion is wrong, rather than to surprise the user15:50
TheJuliaI worry that by doing that people might not get it, but at the same time... we can't prevent every issue or headache if someone needs to do something outside of what the community ends up stating is a happy path15:50
jroll(s/user/admin)15:50
TheJulia++15:51
arne_wiebalckTheJulia: That’s a good point as well, it becomes somewhat more complicated.15:51
*** pcaruana has joined #openstack-ironic15:51
arne_wiebalckOTOH, admins should be used to define the RAID layout of servers.15:52
TheJuliaThen again, maybe someone needs raid5 for their root and they are willing to make it work?15:52
TheJuliaarne_wiebalck: ++15:52
arne_wiebalckRAID-5 for root would not be supported in this approach.15:52
jrollI have to run as my lunch is about ready, but I'll drop a comment on the spec about this discussion and will +2 a matching re-spin :)15:52
arne_wiebalckjroll: thx, I’ll change that part!15:53
arne_wiebalckjroll: TheJulia: Thanks a lot for your support on this!15:53
TheJuliaarne_wiebalck: Thank you for pushing forward15:55
*** gyee has joined #openstack-ironic16:10
*** diga has quit IRC16:12
*** Bhujay has joined #openstack-ironic16:16
*** Emine has quit IRC16:24
*** pcaruana has quit IRC16:28
TheJuliaDeploy templates approved16:33
openstackgerritRiccardo Pittau proposed openstack/networking-baremetal master: Correcting a typo in plugin.sh  https://review.openstack.org/62565216:34
openstackgerritMerged openstack/ironic-specs master: Deploy Templates  https://review.openstack.org/50495216:43
openstackgerritDebayan Ray proposed openstack/ironic master: Fixing Redfish inspection  https://review.openstack.org/62565916:45
TheJuliaoh fun, that redfish one is super valid16:47
*** Emine has joined #openstack-ironic16:48
*** Bhujay has quit IRC16:56
*** derekh has quit IRC17:21
*** jtomasek has quit IRC17:21
*** trown is now known as trown|lunch17:25
*** dsneddon has quit IRC17:26
*** e0ne has quit IRC17:46
*** Emine has quit IRC17:54
*** Emine has joined #openstack-ironic17:54
*** Emine has quit IRC17:58
*** MattMan has quit IRC18:03
*** MattMan has joined #openstack-ironic18:04
rpittaubye all, good evening o/18:16
*** rpittau has quit IRC18:16
*** trown|lunch is now known as trown18:41
*** e0ne has joined #openstack-ironic18:42
*** dougsz has quit IRC18:43
*** e0ne has quit IRC18:46
*** e0ne has joined #openstack-ironic19:01
openstackgerritMerged openstack/networking-baremetal master: Correcting a typo in plugin.sh  https://review.openstack.org/62565219:09
*** Emine has joined #openstack-ironic19:10
*** sthussey has joined #openstack-ironic20:32
*** dsneddon has joined #openstack-ironic20:55
*** dmellado has quit IRC21:05
*** yolanda has quit IRC21:18
*** Emine has quit IRC21:23
*** rcernin has joined #openstack-ironic21:28
*** trown is now known as trown|outtypewww21:58
*** bfournie has quit IRC22:18
*** rh-jelabarre has quit IRC22:21
*** e0ne has quit IRC22:25
openstackgerritJulia Kreger proposed openstack/ironic-specs master: Ironic in container infrastuctures  https://review.openstack.org/62573022:36
TheJuliamy brain falls over now22:39
TheJuliabrain not work now22:42
*** baha has quit IRC22:47
*** Emine has joined #openstack-ironic23:04
*** Emine has quit IRC23:08
*** absubram has joined #openstack-ironic23:53
*** Emine has joined #openstack-ironic23:59
« Sunday, 2018-12-16 Index Tuesday, 2018-12-18 »

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!