Friday, 2021-10-15

*** dviroel|rover|afk is now known as dviroel|out00:00
opendevreviewArun S A G proposed openstack/ironic master: Fix format of kickstart options  https://review.opendev.org/c/openstack/ironic/+/81408700:30
opendevreviewMerged openstack/ironic-lib stable/wallaby: json_rpc: list_opts() should return tuple  https://review.opendev.org/c/openstack/ironic-lib/+/81357004:22
opendevreviewMerged openstack/ironic-lib stable/wallaby: json_rpc: setup.cfg is missing entrypoint for json_rpc's list_opts  https://review.opendev.org/c/openstack/ironic-lib/+/81357104:22
*** mike is now known as MikeCTZA06:13
*** MikeCTZA is now known as Guest296806:13
iurygregorygood morning Ironic, happy friday06:16
MikeCTZAcarrying on with my ironic troubleshooting today, I found yesterday that there was a dhcpsnooping issue so packets weren't getting from the BM node to the DHCP agent06:16
MikeCTZAnow when I DHCP the file is now put in the dhcp folder with the mac-address but has ,ignore so trying to see why06:18
MikeCTZAdigging into the inspector to see if thats where my problem lies with the filter as reading that feels likely to be the case06:24
*** Guest2968 is now known as MikeCTZA06:46
MikeCTZAthe network we are using for our ironic_dnsmasq_interface and ironic_dnsmasq_dhcp_range do we need to add that top openstack as a provider network? I'm not too clear if I need to or not06:55
opendevreviewManuel Schönlaub proposed openstack/sushy master: Add support for additional network resources.  https://review.opendev.org/c/openstack/sushy/+/81385007:08
rpittaugood morning ironic! o/07:24
MikeCTZAat which stage should the ignore be removed from inside the MAC of the box so it can DHCP, and then PXE? I can't quite figure that out, or if something I don't have correctly configured is causing it to be marked as ignore08:01
dtantsurmorning ironic, finally Friday09:22
dtantsurMikeCTZA: ironic's dnsmasq stuff is only used during inspection09:23
dtantsurif you're deploying, "ignore" is what you expect to see09:23
MikeCTZAmorning, has been a looooong week09:23
MikeCTZAI think I have an idea what is wrong09:23
dtantsurbecause the neutron's DHCP agent manages its dnsmasq business separately09:23
MikeCTZAwe have the VLAN as native on the compute / BM side but on the controllers it is not native it is a tagged vlan09:24
iurygregorymorning rpittau dtantsur 09:24
MikeCTZAso I suspect I need to move that to be native on the interface on the controller so that PXE / DHCP can function 09:24
* iurygregory just got back from a long morning trying to solve my vaccination certificate -.-'09:24
dtantsuryou mean, you're trying to PXE boot on a tagged vlan? yep, won't work09:24
dtantsuriurygregory: ouch, what happened?09:24
MikeCTZAIve logged the request with out networking people to change it so as soon as they action that I can see if I have another hurdle09:25
iurygregorydtantsur, the system says I didn't get any vaccine :D09:25
dtantsurOo09:25
dtantsurhave they solved it now?09:25
iurygregorynot 100%09:26
MikeCTZAour system / certificates here the same, some people arent captured correctly and people cant get their certtificates09:26
dtantsur:(09:26
iurygregorynow I have a printed certificate, but the app doesn't save the information and I have to scan every time if I want to use :D09:26
dtantsurwhere are you MikeCTZA?09:26
dtantsur>_<09:26
MikeCTZACape Town, South Africa09:26
dtantsuroh cool09:27
timeuMikeCTZA: there are also some difference between using flat networking and neutron networking for BM in the overcloud. For flat, we usually attach the provisioning NIC of the BM to the ctrlplane network of the controller because that's usually untagged. With neutron integration the plugin that configures the switches will create the corresponding untagged VLANs. You could09:33
timeualso consider switching to redfish and virtualmedia and forgoe the entire PXE issue. 09:33
iurygregorydtantsur, facepalm moment the doctor told me the wrong name of the app :D (my friend just told me I should use other app to have the certificate on my phone) \o/09:37
dtantsuromg09:37
dtantsur(I think we have 2 or 3 apps which can be used)09:37
iurygregorynow I'm using Tecka 09:38
iurygregorythe doctor told me ctecka :D09:38
Inwerejosic1Good morning ironic09:39
dtantsursounds like an implementation in C?09:39
dtantsurmorning Inwerejosic1 09:39
iurygregorydtantsur, LOL :D seems legit I would say09:39
Inwerejosic1dtansur, how are you today?09:39
iurygregorymorning Inwerejosic1 09:40
dtantsurwell, better than yesterday, I would say :) which is not a high bar to take09:40
dtantsurand you?09:40
Inwerejosic1I am in Lagos Nigerian and you?09:40
Inwerejosic1I am good09:40
Inwerejosic1iurygregory hi, how are you?09:41
dtantsurnice! I'm in Düsseldorf, Germany.09:41
iurygregoryInwerejosic1, doing good =) 09:41
Inwerejosic1oh, Gutten Morgen09:41
dtantsur:)09:42
Inwerejosic1I am very glad to be here...09:42
Inwerejosic1so, to understand the frames works I am looking to work with, I practiced with flask and pecan, by creating small sites which I served from my local machine. But I discovered that pecan document structure was more straight forward than flash for me. Did anyone have similar discovery?09:45
dtantsurhonestly, I never touched pecan outside of the ironic context09:45
dtantsurflask is definitely considered a more mainstream framework in the python world nowadays09:46
Inwerejosic1I have been learning and practicing Django and flask before now, yesterday was my first day running pecan09:46
Inwerejosic1Yeah, it is; but I think that is as a result of the publicity it is having around it as against the lack of publicity for pecan09:47
dtantsurI also don't quite like object-dispatch (https://pecan.readthedocs.io/en/latest/routing.html) even though I imagine it looks appealing from the docs09:47
dtantsurthings quickly get ugly when you have a complex URL structure09:48
dtantsurwhile all documentations I've seen concentrate on very straightforward concepts like books :)09:48
Inwerejosic1I will look up the documentation and compare it with a flask routing too and will communicate my findings09:49
dtantsurInwerejosic1: when looking at flask, don't forget to check out their blueprints09:49
dtantsurthis feature is often overlooked09:49
Inwerejosic1I haven't tried it on a complex projects yet09:49
dtantsur(mostly because the flask documentation also concentrates on hello-world level examples)09:49
Inwerejosic1YES, and that hello world context is very limiting09:50
Inwerejosic1I will practice the with Ironic this weekend to help me begin my contributions to the project at hand.09:52
dtantsur++09:54
opendevreviewAija Jauntēva proposed openstack/sushy master: Remove deprecated Task monitors and Volume methods  https://review.opendev.org/c/openstack/sushy/+/81412409:54
dtantsurTIL people who get J&J vaccine should get a booster now10:21
MikeCTZAOK so I've have out VLAN made native on both controller and BM node now, so I'm now seeing if I can PXE once I've checked a few things, an initial test didn't work10:31
MikeCTZAso I can see the DHCP request come into the physical host but nothing gets sent back to the client10:57
arne_wiebalckGood morning, Ironic!11:23
*** dviroel|out is now known as dviroel|rover11:41
dtantsurhey arne_wiebalck 11:47
arne_wiebalckhi dtantsur o/11:48
dtantsurjanders: do we need to backport https://review.opendev.org/c/openstack/ironic/+/812795 ?11:55
iurygregorywe do 11:56
iurygregorymorning arne_wiebalck o/11:56
* iurygregory back from lunch11:56
dtantsurit's a question-slash-request :)11:56
iurygregorydtantsur, ohhh I see =)11:56
arne_wiebalckhey iurygregory o/11:56
jandersdtantsur: thinking if there are any disadvantages (Dave Cain got new firmware that no longer does that)11:57
jandersI think risk of breakage is low (hopefully not "famous last words") so I think we should11:57
jandersI will add it to my TODO for Monday11:58
dtantsurjanders: there may be people out there with old firmware still11:58
dtantsurthanks!11:58
jandersdtantsur: no worries :)11:58
MikeCTZAI'm finding config issue after config issue with this setup of ours, but I'm getting there slowly but surely, my goal is to try get ironic up today ... but with 1hr left in my work day I'm doubtful11:59
dtantsurthere will be another week and another work day :)12:00
MikeCTZAso glad we have a test setup of old hardware we can play like this to get to know it all better, I've learned a decent bit today already12:00
janderssee you on Monday Ironic12:06
jandershave a great weekend everyone o/12:06
dtantsursee you12:07
rpittauhaving long lunch with relatives is great, until the 3rd bottle of wine12:13
dtantsur:D12:13
dtantsurso, Friday is over for you?12:13
iurygregorybye janders o/12:13
rpittauwell, I'll try not to break too much stuff12:14
dtantsuriurygregory: who of us is going to be at the TC-PTL sync on PTG Monday? Initially Julia asked me to do it, and it may make sense for you to be with the team, but I'm ready to swap.12:14
iurygregoryrpittau, don't break the bottles!12:14
dtantsurref https://etherpad.opendev.org/p/tc-ptl-interaction12:14
iurygregorydtantsur, WOOT?! we have something on monday?!12:15
iurygregoryholy moly12:15
dtantsuriurygregory: it's the same slot as ours12:15
iurygregoryin theory im on PTO on monday .-.12:15
iurygregorybecause I will be in Prague with a friend and his family12:15
dtantsuroh right..12:16
dtantsurhonestly.. I don't want to upset TheJulia, but looking at the Monday's schedule, I have a feeling we should cancel that slot12:16
dtantsurthe topics are more midcycle-style IMO12:16
iurygregoryhummm12:17
iurygregorylet me get more coffee to think about this 12:18
rpittaucofee sounds great!12:19
* rpittau goes for coffee too12:19
dtantsur++12:21
dtantsurtime for a 2nd cup12:21
ajyaHappy Friday, can these bugfix/18.1 backports get some attention https://review.opendev.org/c/openstack/ironic/+/812209 https://review.opendev.org/c/openstack/ironic/+/809139 thanks12:22
opendevreviewMerged openstack/ironic master: SRBAC - Prepare for additional services  https://review.opendev.org/c/openstack/ironic/+/81241112:24
opendevreviewVerification of a change to openstack/ironic master failed: Remove debian packages file for devstack  https://review.opendev.org/c/openstack/ironic/+/81031412:24
rpittauhey ajya! happy friday!12:35
iurygregorydtantsur, oh D: all topics on monday are from TheJulia and I don't think Julia will be around today .-.12:38
dtantsurmmm, and not necessarily on Monday as well..12:39
iurygregoryyeah =)12:39
iurygregoryI'm wondering how to handle if we cancel... maybe we can do a virtual midcycle in January?12:40
iurygregoryor we book more slots during the PTG?12:40
iurygregoryor we do something in the week after the PTG LOL12:41
dtantsurI don't think we can fit more into the PTG12:41
iurygregoryyeah12:41
rpittauwe should probably at least exchange monday schedule with another day if TheJulia won't be present12:45
iurygregoryrpittau, the problem is that most of the topics are from TheJulia =)12:47
rpittaummmm yeah12:47
dtantsurand we don't know TheJulia's schedule12:48
MikeCTZAlooks like our networking guys havent done the changes for my dhcp/pxe correctly so probably Monday I'll carry on with this12:51
*** dviroel|rover is now known as dviroel|afk12:54
arne_wiebalckTheJulia: do you have more details on the "IPMI does not set the correct boot device" issue mentioned in the schedule?13:01
opendevreviewVerification of a change to openstack/ironic master failed: Remove debian packages file for devstack  https://review.opendev.org/c/openstack/ironic/+/81031413:04
opendevreviewDmitry Tantsur proposed openstack/ironic master: Remove redundant node_periodic tests  https://review.opendev.org/c/openstack/ironic/+/81417013:54
opendevreviewMerged openstack/ironic-python-agent bugfix/8.1: Software RAID: Call udev_settle before creation  https://review.opendev.org/c/openstack/ironic-python-agent/+/81375314:04
opendevreviewManuel Schönlaub proposed openstack/sushy master: Add support for additional network resources.  https://review.opendev.org/c/openstack/sushy/+/81385014:23
TheJuliaI'll be around on monday14:25
TheJuliawhats the question?14:25
dtantsurTheJulia: morning! The problem: I'll be on the TC sync, iurygregory will be on PTO.14:26
TheJuliaokay14:26
dtantsurso if you'll be there for the PTG, you can drive the discussion14:26
dtantsurotherwise its fate would be unclear14:26
iurygregoryyeah14:26
TheJuliayes, with copious notes :)14:26
dtantsurspeaking of which, is there anything I should bring to the TC meetup?14:26
TheJuliaarne_wiebalck: so, I *think* that is in regard openbmc, which doesn't really properly grok these sorts of things very well :\14:27
TheJuliaarne_wiebalck: and added conundrum, forks of the upstream code for vendor bmcs, so odds we will ever see fixes are minimal14:28
iurygregoryTheJulia, also fyi the info about Stopping use of wiki.openstack.org is now in https://etherpad.opendev.org/p/ironic-wiki-bye-bye :D14:28
TheJuliadtantsur: beyond the release cycle model is hurting operators as a whole and the TC wants to continue business as usual... nah14:28
iurygregorythere is a link to it in the yoga-ptg etherpad also14:28
arne_wiebalckTheJulia: ok, thanks! I was just interested :)14:28
arne_wiebalckwhat release cycle would operators prefer ... was there a discussion somewhere?14:30
TheJuliadtantsur: I have strong opinions regarding the TC, in part revolving around it's name no longer really matches what it's community role is at this point14:31
dtantsurmm, I suspect they won't appreciate me showing up with "hey folks, what you doing here really?" :D14:31
TheJuliaarne_wiebalck: so this has been a semi-long standing thing for me, most operators of larger deployments I have a chance to talk to often say it takes them just 6-9 months to certify new versions before even performing upgrades or deployment to production14:32
dtantsurTheJulia: given your opinions, will it make sense to swap us, so that you go there and I drive the discussion in our room?14:32
dtantsuralthough I may miss some context on your items..14:32
TheJuliaso us trying to churn out new major releases as a wider community is just putting operators more and more behind14:32
TheJuliadtantsur: they have made it abundantly clear they don't want to listen to my feedback in that arena, so its kind of moot14:33
arne_wiebalckhmm, yeah14:33
dtantsurI see14:33
arne_wiebalckeven in our deployment (where we try to stay close), we have fallen behind lately14:33
TheJuliaand, the other conundrum is there *is* value regarding being able to generate a press release every so often14:33
arne_wiebalckif you only focus on upgrades, it is ok, but don't do anything in addition or you miss the train :)14:34
TheJuliathis all needs to be balanced, and focusing on low level technical details doesn't help them stay on top of the bigger picture14:34
TheJuliaarne_wiebalck: Train, indeed :)14:34
arne_wiebalck:)14:35
* TheJulia ponders breakfast before starting laundry and suitcase packing14:35
TheJuliadtantsur: express any feelings you have to them, I'm kind of out of spoons dealing with them and the spoons never regrew14:37
TheJuliaPartially why I've pushed the whole "maybe we should be a top level project" aspect... which would make the foundation super happy to be honest14:37
TheJuliain part, because press releases ;)14:37
TheJuliabut only part14:38
dtantsurheh14:41
rpittaubye everyone, have a great weekend! o/15:01
dtantsurc u rpittau 15:01
arne_wiebalckbye rpittau o/15:05
ade_leeTheJulia, stevebaker[m] hey - just reminder about https://review.opendev.org/c/openstack/ironic/+/79773915:10
dtantsurComputerSystem 1.13.0 introduced a link to VirtualMedia \o/15:20
dtantsurI think it's something we should start using15:21
opendevreviewTzu-Mainn Chen proposed openstack/ironic master: Fix restricted allocation creation  https://review.opendev.org/c/openstack/ironic/+/81200715:43
Kathy_hi15:46
opendevreviewDmitry Tantsur proposed openstack/sushy master: Add some VirtualMedia fields from 1.3.0 and 1.4.0  https://review.opendev.org/c/openstack/sushy/+/81419115:49
dtantsurHi Kathy_ 15:51
dtantsurTheJulia: do you by chance have a minute to look at https://storyboard.openstack.org/#!/story/2009294 ?15:51
*** pmannidi is now known as pmannidi|AFK15:52
*** dviroel|afk is now known as dviroel|rover15:52
NobodyCamGood Morning Ironic'ers 16:22
NobodyCamand OFC...16:22
NobodyCamTGIF!16:22
dtantsurmorning NobodyCam 16:28
dtantsurI've just learned of an ironic bug that can win the nomination "facepalm of year"16:28
dtantsurquiz: what is wrong with this line? https://opendev.org/openstack/ironic/src/branch/master/ironic/drivers/modules/image_cache.py#L16416:30
JayFthat's going to write to `/` everytime, isn't it?16:30
JayFer, resolve to `/`16:31
dtantsurJayF: nope, more interesting16:32
dtantsurthat would be always broken, I think16:32
dtantsurand this bug was not noticed until... (spoiler hidden)16:32
JayFah; I see it16:33
TheJuliadtantsur: no objection to the rfe, no real questions. I guess at  ahigh level, it soudns good to me16:41
dtantsurcool, thx!16:41
NobodyCamhey hey dtantsur :) 16:42
opendevreviewDmitry Tantsur proposed openstack/ironic master: Do not use any parts of image URL in temporary file names  https://review.opendev.org/c/openstack/ironic/+/81420116:59
dtantsurJayF: wdyt ^^^?16:59
JayFIt always makes me sad when I see a redhat bug with no upstream bug, but for something like this :shrug:17:03
JayFthat's actually going to be better overall too, because it'll be on the same disk17:03
JayFhm, I guess it always was before, too17:04
dtantsuryeah, we rely on hard links17:04
dtantsurbut now it's always a UUID rather than a random part of the URL17:04
JayFgerrit looks mildly more fancy now17:04
JayF+2'd17:04
dtantsuryep, a recent update17:04
dtantsurthanks!17:04
dtantsuron this positive note I'm wishing everyone a great weekend17:13
JayFo/ 17:16
JayFhave a good weekend, you deserve it17:16
-opendevstatus- NOTICE: The Gerrit service on review.opendev.org will be offline starting in 5 minutes, at 18:00 UTC, for scheduled project rename maintenance, which should last no more than an hour (but will likely be much shorter): http://lists.opendev.org/pipermail/service-announce/2021-October/000024.html17:59
*** dviroel|rover is now known as dviroel|rover|afk18:01
*** dviroel|rover|afk is now known as dviroel|rover18:09
*** dviroel|rover is now known as dviroel|rover|afk21:08
opendevreviewJulia Kreger proposed openstack/ironic master: There is no aim, we do deploy/manage baremetal.  https://review.opendev.org/c/openstack/ironic/+/81423922:53
*** pmannidi|AFK is now known as pmannidi23:38

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!