Wednesday, 2021-07-28

opendevreviewJacob Anders proposed openstack/sushy master: [WIP] Remove optional attributes from insert_media  https://review.opendev.org/c/openstack/sushy/+/80245200:17
opendevreviewJacob Anders proposed openstack/sushy master: Change defaults - optional insert_media attributes  https://review.opendev.org/c/openstack/sushy/+/80245200:28
opendevreviewJacob Anders proposed openstack/sushy master: Change defaults - optional insert_media attributes  https://review.opendev.org/c/openstack/sushy/+/80245200:30
opendevreviewJacob Anders proposed openstack/ironic master: Remove hardcoded parameters from insert_media call  https://review.opendev.org/c/openstack/ironic/+/80264301:31
TheJuliajanders: hey, it would be good if you revised commit messages to indicate why in the grand scheme of things and if there is a related rfe/story/bug/etc02:49
jandersTheJulia: thanks, will include BZs02:50
TheJuliakeep in mind with commit messages, it is good to be able to capture the overall why just from the commit log02:50
TheJuliaNobodyCam: Ahh, yes. We took it to the max!02:52
opendevreviewJacob Anders proposed openstack/ironic master: Remove hardcoded parameters from insert_media call  https://review.opendev.org/c/openstack/ironic/+/80264304:19
opendevreviewJacob Anders proposed openstack/sushy master: Change defaults - optional insert_media attributes  https://review.opendev.org/c/openstack/sushy/+/80245204:19
jandersTheJulia: added BZ references/links ^^ - apologies for forgetting these earlier04:20
opendevreviewMerged openstack/ironic stable/train: Provide a path to set explicit ipxe bootloaders  https://review.opendev.org/c/openstack/ironic/+/80166904:34
iurygregorygood morning Ironic!05:56
cennegood mornings iurygregory.06:02
cennegood morning Ironic!06:02
iurygregorymorning cenne o/06:03
cenneo/ :)06:03
stendulkeriurygregory, cenne o/06:49
iurygregoryhey stendulker o/06:50
cennehey o/ stendulker06:54
*** rpittau|afk is now known as rpittau07:02
rpittaugood morning ironic! o/07:02
iurygregorymorning rpittau o/07:04
rpittauhey iurygregory :)07:05
iurygregoryHey everyone, so far we have 8 people that are planning to attend the MidCycle, if you are planning in participate please provide some input in the doodle https://doodle.com/poll/mhh959u4s4rturxi ( I will close the doodle tomorrow morning)07:13
dtantsurmorning ironic07:14
iurygregorymorning dtantsur 07:14
opendevreviewDmitry Tantsur proposed openstack/ironic stable/wallaby: Fix regression in ramdisk deploy kernel parameters  https://review.opendev.org/c/openstack/ironic/+/80266407:29
iurygregorydtantsur, iscsi removal was like this https://classicprogrammerpaintings.com/post/624257359339995136/engineers-remove-dead-code-after-dropping-a ? hehehe07:29
dtantsurexactly :D07:30
jandersgood morning Ironic o/08:00
jandersstendulker rpittau thank you for prompt reviews, will fix up the things you pointed out later this evening08:00
iurygregoryhey janders o/08:04
opendevreviewMerged openstack/ironic stable/victoria: Use openstack-tox for ironic-tox-unit-with-driver-libs  https://review.opendev.org/c/openstack/ironic/+/80246208:41
opendevreviewMerged openstack/ironic bugfix/18.1: Fix regression in ramdisk deploy kernel parameters  https://review.opendev.org/c/openstack/ironic/+/80246408:41
jandersstendulker rpittau w/r/t your comments in https://review.opendev.org/c/openstack/sushy/+/802452/6/sushy/resources/manager/virtual_media.py#96 - would you like me to change the "fixes" section of the release note to "upgrades", or keep it as-is and add a separate "upgrades" containing only the changed-defaults warning?08:41
iurygregorydo we plan to backport the change?08:43
rpittauiurygregory: I don't think we can backport that08:43
iurygregoryyeah, since we need to change default values etc =(08:43
rpittauyep :/08:44
rpittaujanders: I think we need to change from fixes to upgrades08:44
jandersthank you iurygregory rpittau08:44
jandersrpittau is it okay to reference downstream BZ in an upstream change like I did, or do I need to create a matching bug upstream and reference that?08:45
dtantsurwe definitely need to backport that08:46
dtantsurthe default values are wrong, why cannot we change them?08:46
dtantsuralso -1 to not adding "fixes". it's a bug fix.08:47
dtantsur(even if it does have upgrade impact, a bug fix has to be mentioned together with the upgrade note)08:47
dtantsurjanders, iurygregory, rpittau ^^08:48
iurygregoryif the defaults are wrong it should be fine to backport, on the ironic side we will probably need to add the parameters so we can set if the request has the information08:48
iurygregoryyeah ++ to upgrade + fixes 08:49
dtantsurI don't think we ever set these. I'm not even sure why they were added08:49
dtantsur(probably for completeness)08:49
iurygregorydtantsur, https://review.opendev.org/c/openstack/ironic/+/802643/2/ironic/drivers/modules/redfish/boot.py08:49
iurygregorywe send inserted=True/write_protected=True08:50
dtantsur#facepalm08:50
iurygregoryhehehe08:50
dtantsuroookay, just keel them with fire08:50
dtantsurehmm, kill08:50
iurygregory(╯°□°)╯︵ ┻━┻08:50
rpittauif we backport that, don't we potentially break everyuthing that relies on the old values ?08:51
dtantsurnamely?08:51
rpittauI don't have a specific example, I'm just saying in general08:52
dtantsurin general it's possible. in practice we only know examples of the opposite.08:52
rpittauok then08:53
dtantsurokay, if we go really paranoid, this is what we can do:08:53
dtantsur1) make a sushy change that handles None values, but don't change the defaults08:53
dtantsurehh, I started typing and realized it won't work :(08:54
dtantsurokay, another attempt:08:55
iurygregorymaybe if ironic can send None we can handle in sushy to not include in the payload?08:55
dtantsurwe cannot backport ^^^08:55
dtantsurbecause it relies on a specific version of sushy08:55
iurygregoryoh =(08:55
dtantsurold sushy will send "null" to the BMC08:55
dtantsurokay, so:08:55
dtantsur1) change sushy to not send any values if they match the official defaults in the standard (true and true)08:55
dtantsur2) change ironic to provide the values matching the standard (true and true again)08:56
dtantsur3) change sushy to default to None only on master08:56
iurygregoryinteresting... (2 we are already doing this...)08:56
dtantsur4) change ironic to not provide anything only on master08:56
dtantsureven better!08:56
iurygregory++ makes sense to me08:56
dtantsurrpittau, janders, wdyt?08:57
dtantsurI know it's a bit more work, but I think it resolves the backporting concerns08:57
jandersdtantsur re-reading ^^ the third time to make sure I got it right. I have no concerns about extra work if it gets the job done and doesn't risk breaking stuff.08:57
rpittauthat would work with my level of paranoia :)08:58
janderscertain degree of paranoia is healthy in engineering08:58
iurygregorystep2 you can ignore since we are already sending true true by default :D08:58
jandersotherwise we end up with quad-engine aircraft that can fly on two engines, but only if they are on two different wings08:59
iurygregoryOMG08:59
janders(not quite in line with how gas turbines usually fail when they fail)08:59
jandersso - im all for a healthy dose of paranoia :)09:00
rpittauwell if they alternate between a set and another, possibly they can fly :P09:00
janderspoor Concorde didn't fly far in 200009:00
rpittauehm right09:00
jandersso to sum up steps 1)-4) - we have 3) and 4) done, right?09:03
janders2) is the current state without my patch (and this will remain outside master)09:03
jandersso I will just need to implement 1) for the non-master patch right?09:04
iurygregory1 needs to go to master I would say, and you can do the backports from it09:05
dtantsuryep09:05
iurygregoryafter 1 merges you can apply 3 409:05
dtantsurwe always start on master, unless impossible09:05
jandersok! makes sense09:06
janderslastly - w/r/t https://review.opendev.org/c/openstack/ironic/+/802643/2/releasenotes/notes/remove-hardcoded-insert-media-attrs-6621a811f46cf8d3.yaml - should I just drop the release note entirely? 09:06
iurygregoryif you have 1 and 3 together you would need to remove things in the backport, so it's better to have a clean backport =)09:06
dtantsurjanders: I think both 1 and 3 will have an impact worth mentioning in a release note09:07
jandersdtantsur: ACK09:07
jandersdinner break, will address remaining review bits and sort out change 1) in a bit09:21
iurygregoryjanders, take your time o/ enjoy the dinner09:22
opendevreviewRiccardo Pittau proposed openstack/ironic master: Increase version of hacking and pycodestyle  https://review.opendev.org/c/openstack/ironic/+/80267509:33
*** sshnaidm|afk is now known as sshnaidm09:45
opendevreviewMerged openstack/ironic stable/ussuri: Use openstack-tox for ironic-tox-unit-with-driver-libs  https://review.opendev.org/c/openstack/ironic/+/80246310:02
jandersdtantsur: Should I add [DNM] tag on patches in change https://review.opendev.org/q/Ieab097c63f218c3c262f013c29495f5393dd3cfc (steps 3) and 4) from our plan) or is workflow-1 sufficient?10:06
dtantsurjanders: W-1 is fully sufficient10:08
jandersdtantsur: ty!10:08
dtantsurI only use [WIP] on patches that are not finished and [DNM] on patches that should never be merged10:08
jandersdtantsur: ACK10:10
opendevreviewJacob Anders proposed openstack/sushy master: Change defaults - optional insert_media attributes  https://review.opendev.org/c/openstack/sushy/+/80245210:16
opendevreviewJacob Anders proposed openstack/ironic master: Remove hardcoded parameters from insert_media call  https://review.opendev.org/c/openstack/ironic/+/80264310:29
jandersallright - I hope ^^ should sort out 3) and 4) from the discussion above, 2) should be all good without any changes, now onto 1)10:31
dtantsurjanders: you'll need to rebase 3) on top of 1)10:37
jandersdtantsur: makes sense10:38
iurygregoryyeah, just to avoid merge conflict :D10:38
dtantsurand to figure out the release notes so that they don't contradict or duplicate each other10:38
iurygregoryyup, make sense10:39
dtantsurlike 1) fixes: support for X11, X12; upgrade: no longer send the fields when they match the standard defaults; 2) upgrade: changes the method default to None10:41
jandersdtantsur: thanks!10:42
jandersdtantsur: quick one - now that I am more mindful about the current defaults, I realised that sushy defaults to 'Inserted': True, 'WriteProtected': False while Ironic sets 'WriteProtected': True10:48
jandershttps://opendev.org/openstack/sushy/src/branch/master/sushy/resources/manager/virtual_media.py#L9610:49
jandershttps://opendev.org/openstack/ironic/src/branch/master/ironic/drivers/modules/redfish/boot.py#L19910:49
janderswhat was the reason for that, do you remember?10:50
janders(checking history too)10:50
dtantsurI suspect it was accidental..10:53
jandersdtantsur: ACK, thank you10:53
opendevreviewJacob Anders proposed openstack/sushy master: Removing optional fields from insert_media payload  https://review.opendev.org/c/openstack/sushy/+/80269012:05
opendevreviewJacob Anders proposed openstack/sushy master: Removing optional fields from insert_media payload  https://review.opendev.org/c/openstack/sushy/+/80269012:07
janders^^ change 1) as per the plan above; first pass12:08
jandersnow - updating commit message/release note for 3)12:08
opendevreviewJacob Anders proposed openstack/sushy master: Change defaults - optional insert_media attributes  https://review.opendev.org/c/openstack/sushy/+/80245212:18
jandersdtantsur rpittau iurygregory: w/r/t sushy vMedia fixes I think I have 1) 3) and 4) ready (and 2) shouldn't require any code changes)12:25
janders1) https://review.opendev.org/c/openstack/sushy/+/80269012:26
janders3) https://review.opendev.org/c/openstack/sushy/+/80245212:27
janders4) https://review.opendev.org/c/openstack/ironic/+/80264312:27
janderslet me know if these make sense to you and I will do more testing (and fixing if required) tomorrow12:27
janderssee you tomorrow Ironic o/12:27
iurygregoryjanders, tks! good night!12:28
rpittaubye janders :)12:29
jandersthank you for your insights iurygregory rpittau dtantsur12:30
opendevreviewJacob Anders proposed openstack/sushy master: Removing optional fields from insert_media payload  https://review.opendev.org/c/openstack/sushy/+/80269012:38
opendevreviewBob Fournier proposed openstack/sushy-tools master: Add additional BIOS Settings  https://review.opendev.org/c/openstack/sushy-tools/+/80269512:44
TheJuliagood morning13:05
dtantsurmorning TheJulia 13:08
iurygregorygood morning TheJulia 13:16
dtantsura friendly reminder: the bugfix CI is still not configured: https://review.opendev.org/c/openstack/ironic/+/801876 https://review.opendev.org/c/openstack/ironic-inspector/+/801873 https://review.opendev.org/c/openstack/ironic-python-agent/+/80189815:32
opendevreviewDmitry Tantsur proposed openstack/bifrost master: [WIP] Move Nginx code to a new role bifrost-nginx-install  https://review.opendev.org/c/openstack/bifrost/+/80253215:52
opendevreviewDmitry Tantsur proposed openstack/bifrost master: DNM test the upgrade job  https://review.opendev.org/c/openstack/bifrost/+/80067316:00
opendevreviewDmitry Tantsur proposed openstack/bifrost bugfix/11.0: Install libvirt-python from source instead of a wheel  https://review.opendev.org/c/openstack/bifrost/+/80274916:05
rpittaugood night! o/16:21
*** rpittau is now known as rpittau|afk16:21
opendevreviewDmitry Tantsur proposed openstack/bifrost master: Keep sushy-emulator state directory in /var/lib  https://review.opendev.org/c/openstack/bifrost/+/80275416:37
opendevreviewDmitry Tantsur proposed openstack/bifrost master: Keep sushy-emulator state directory in /var/lib  https://review.opendev.org/c/openstack/bifrost/+/80275416:44
dtantsuro/16:51
TheJuliagoodnight dtantsur 16:51
opendevreviewMerged openstack/sushy-tools master: Add additional BIOS Settings  https://review.opendev.org/c/openstack/sushy-tools/+/80269517:02
opendevreviewMerged openstack/ironic bugfix/18.1: Configure CI for bugfix/18.1  https://review.opendev.org/c/openstack/ironic/+/80187617:14
opendevreviewMerged openstack/ironic-inspector bugfix/10.7: Configure CI for bugfix/10.7  https://review.opendev.org/c/openstack/ironic-inspector/+/80187317:14
opendevreviewMerged openstack/ironic-python-agent bugfix/8.1: Configure CI for bugfix/8.1  https://review.opendev.org/c/openstack/ironic-python-agent/+/80189817:14
opendevreviewMerged openstack/python-ironic-inspector-client master: Update min version of tox to use allowlist  https://review.opendev.org/c/openstack/python-ironic-inspector-client/+/79639817:22
opendevreviewMerged openstack/virtualbmc master: Update min version of tox to use allowlist  https://review.opendev.org/c/openstack/virtualbmc/+/79640117:24
iurygregoryHey everyone, so far we have 9 people that are planning to attend the MidCycle, if you are planning in participate please provide some input in the doodle https://doodle.com/poll/mhh959u4s4rturxi ( I will close the doodle tomorrow morning my time 10am UTC+2)17:42
stevebakergood morning20:42
jandersgood morning Ironic o/21:11
TheJuliagood morning stevebaker janders 21:42
jandershey TheJulia21:45
jandersTheJulia I dropped some messages w/r/t vMedia attribute tweaks on the internal channel - if you have time I'd like to chat further (getting coffee/bfast :)21:46
TheJuliaLets chat here, and honestly coffee does sound really good :)21:50
* TheJulia goes and makes an espresso based beverage21:50
* TheJulia returns with a chocolate raspberry soy latte21:59
TheJuliawell, I guess it is more mocha raspberry soy....21:59
jandersOK, back22:24
TheJuliawb22:25
jandersw/r/t removing vMedia InsertMedia optional attributes - the original change was https://review.opendev.org/c/openstack/sushy/+/80245222:27
TheJuliaokay22:27
jandersI CCed Aija and Shivanand on this one, got some good feedback22:27
jandersbut in discussion on IRC Riccardo/Iury/Dmitry had concerns about backporting this and that is how we came up with two changes doing the same thing in different ways22:28
jandersshort summary of this discussion is in that JIRA comment I shared in the internal channel in case you wanted to have a look22:29
jandersthe idea behind all this is we specify optional attributes to InsertMedia Redfish calls that aren't needed and their presence breaks SuperMicro vMedia22:30
TheJuliaI saw it, that entire jira story is basically the context your missing upstream22:30
jandersso why not remove these optional attributes22:30
TheJuliawell, the root seems to be 1) make it compliant to the standard and 2) also get supermicro gear working22:31
TheJuliaat least, 2021.1 schema it seems22:31
jandersyeah bit of a moving target isn't it :)22:31
jandersbut 2021.1 it is for the time being22:31
TheJuliaExtreme Programming Moving Target, The Television Show!22:31
jandersbring it on! :)22:31
* TheJulia wonders how gets producer credentials22:31
TheJulias/how/who/22:32
opendevreviewMerged openstack/ironic-specs master: Clean up released features/items  https://review.opendev.org/c/openstack/ironic-specs/+/80165022:32
jandersw/r/t testing - I haven't tested these particular changes to sushy as I am waiting for the implementation details to settle (there's still a lot of movement as reviews come in)22:32
jandersbut22:32
jandersI did test it on the RF call level22:32
jandersand haven't found a machine that needs these attributes22:32
janders(tested on Dell, HP and a few SuperMicros)22:33
TheJuliathat is good, would be ideal to document on at least the commit message22:33
jandersOK! I can do that22:33
TheJuliaif they *are* entirely redundant, then gives us more evidence and comfort that they can just be removed on a backport espescialy if they are redundant to the spec22:33
jandersnow - with the context - if it were an upstream RFE, I'd create a story, tasks, etc22:34
TheJuliaIt is all about setting context in that few paragraphs22:34
TheJuliaIt still is, and should be tracked as such22:34
jandersthis is a very-downstream (not even OSP) bug at the moment22:34
jandersso what is the right way to go about this?22:34
TheJuliaDoesn't hurt to take 5 minutes dump some information in and go from there. It does sound like a bug in the behavior where the spec also seems to be simplier *now*22:35
jandersdo we still need a story for a bug reported against OCP22:35
TheJuliaYes, because RH entries can be made private and have private posts22:35
jandersok, I can do that22:36
TheJuliaso either the commit messages need to provide every bit of context required, or we need to have a little information upstream to explain what/why22:36
TheJuliaThere is always some overlap there, but it goes quick :)22:36
jandersjust was concerned about how many different entities we'll end up with with for a trivial fix (BZ, GH for ironic-images re-spin, story, multiple changes)22:37
jandersbut I guess it is what it is22:37
janderss/GH/GH PR22:37
TheJuliaone story, two tasks, the bz for downstream tracking22:37
TheJuliadownstream pr for the re-spin once the patches make it through the build pipeline22:38
jandersOK, I can do this22:39
jandersI will focus on https://review.opendev.org/c/openstack/sushy/+/802690 first as it should be pretty close and that's the one that will be needed for 4.8 fixes22:40
TheJuliaIt gets worse for downstream backports of course if you have to clone stuff :\22:40
jandersyeah22:40
TheJuliajanders: cool, well anything else?22:40
janders80-90% of the work on this one will be creating tracking bits, but whatever gets the job done :)22:40
TheJuliathat is the way with all backport heavy things22:41
TheJuliathat and waiting for pipelines22:41
jandersjust quickly checking if you had any specific comments w/r/t master-only fix22:41
TheJuliaI'd prefer whatever gets backported lands to master first22:41
TheJuliaPersonally22:42
jandersyeah that's the plan22:42
TheJuliaif that can't be the case, then the need is typically just explaining why22:42
jandersI think we covered all the comments regarding https://review.opendev.org/c/openstack/sushy/+/802452 in the discussion about the backportable one too22:42
jandersthanks TheJulia22:42
TheJuliaYeah, that is what I seemed to garner22:43
jandersso - for the story22:43
janders1) does it need to be against Ironic or sushy specifically?22:44
TheJuliajanders: I'd file it against ironic and just assign one of the tasks to sushy22:44
jandersok! you answered my second question before I asked it22:44
janders:D22:44
TheJuliaExcellent!22:44
TheJuliaWell with that being done, I think it is time to go exercise22:44
jandersand then both backportable and non-backportable patches go against the sushy task?22:44
TheJuliayeah22:45
jandersI think that is a fine plan! :)22:45
jandershave a good evening TheJulia, enjoy!22:45
janders(I will get my share of hill climbing later today)22:45
*** pmannidi|AFK is now known as pmannidi23:08

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