opendevreview | OpenStack Proposal Bot proposed openstack/ironic-inspector master: Imported Translations from Zanata https://review.opendev.org/c/openstack/ironic-inspector/+/927701 | 03:36 |
---|---|---|
opendevreview | Dmitry Tantsur proposed openstack/ironic master: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/928931 | 08:16 |
opendevreview | Dmitry Tantsur proposed openstack/ironic master: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/928931 | 08:35 |
opendevreview | Dmitry Tantsur proposed openstack/ironic master: Fix inspection if bmc_address or bmc_v6address is None https://review.opendev.org/c/openstack/ironic/+/928885 | 08:36 |
dtantsur | both patches need an urgent review ^^^ | 08:36 |
opendevreview | Szymon proposed openstack/metalsmith master: Add ability to specify columns used for json output https://review.opendev.org/c/openstack/metalsmith/+/928993 | 09:03 |
masghar | So in ironic, when a node is deleted, is it also cleaned as a result? Say, using automated cleaning? | 11:42 |
JayF | No, you mark node retired and on undeploy, after cleaning, it goes to manageable instead of available | 12:14 |
JayF | So you have a clean node ready to be deleted | 12:14 |
masghar | I see! Thanks Jay | 12:38 |
dtantsur | masghar: note that metal3 sorta automates it for you (without retirement) | 12:39 |
masghar | okay | 12:39 |
dtantsur | attention cores: everything is quite broken until we merge https://review.opendev.org/c/openstack/ironic/+/928931 and https://review.opendev.org/c/openstack/ironic/+/928885 | 12:39 |
dtantsur | JayF, TheJulia, the 3 of us seem to be the only active cores now | 12:40 |
JayF | I single core approved the CI fix and put a plus 2 on the inspection fix | 12:42 |
dtantsur | thx! | 12:42 |
JayF | dtantsur: by now mean for this point in time? | 12:42 |
JayF | I count five. Me, you, Julia, Ricardo, Iury | 12:43 |
masghar | Riccardo and Iury are on PTO these days | 12:43 |
dtantsur | right ^^ | 12:43 |
dtantsur | Riccardo is back next week, Iury even later | 12:43 |
JayF | Okay that's what I thought just making sure that there hadn't been any shuffling around that I wasn't aware of | 12:44 |
JayF | 😅 | 12:44 |
TheJulia | Wait everything broken could be good right? Vacation?!? | 12:54 |
TheJulia | Joking, kind of :) | 12:54 |
TheJulia | going to be a slow start for me this morning. Had to take meds that kept me up until 1 am | 12:55 |
TheJulia | JayF: your up really early | 13:00 |
dtantsur | TheJulia: take it easy, rest well | 13:01 |
dtantsur | nothing can be merged anyway :D | 13:01 |
masghar | yep, definitely take it easy | 13:01 |
JayF | TheJulia: I'm often up this early and lurking IRC on my phone. Just rare that I'm going to break the silence | 13:06 |
TheJulia | Ahh | 13:07 |
TheJulia | Time to move to the coffee maker now that the cats have left me. | 13:07 |
opendevreview | Dmitry Tantsur proposed openstack/sushy master: Make default request timeout configurable on Connector https://review.opendev.org/c/openstack/sushy/+/929055 | 13:22 |
TheJulia | so at one point https://review.opendev.org/c/openstack/ironic/+/928931 was similar to other configuration we had. It occurs to me that broke because one of the prior clouds broke because all public intertubes were over a tunnel. | 13:29 |
TheJulia | but, should that ever happen again, we can cross that bridge at that point | 13:29 |
JayF | Well essentially the discovery is that you simply cannot use IPv6 if the MTU is lower than that | 13:47 |
JayF | So if that change ends up breaking, we'll have to re-architect something about how that networking works for us | 13:47 |
opendevreview | Merged openstack/ironic master: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/928931 | 13:50 |
dtantsur | yay | 13:50 |
dtantsur | now only metal3/bifrost is not feeling well | 13:50 |
dtantsur | no clean cherry-picks because of the job disabling, sigh | 13:51 |
frickler | JayF: I was asking on the patch, but since this is maybe semi-urgent: do you want to release team to override the ptl-approval to release the cve fixes? https://review.opendev.org/c/openstack/releases/+/928862 | 13:54 |
cardoe | So total aside but I see dtantsur's requests timeout. Yeah that's a super important thing with Python requests... you gotta set a timeout. Its terrible it doesn't do that by default. BUT as a follow on, with the goal to go away from eventlet. I know the recommendation was to use aiohttp in the future. I don't believe it's in the global requirements. But we should nudge that to happen (or httpx if people wanna switch to | 14:01 |
cardoe | that). | 14:01 |
cardoe | Then start using aiohttp or httpx in place of requests. | 14:01 |
cardoe | Because it's a trivial sed to switch a code base with either of those from the sync backend that requests uses to the asyncio backend. | 14:01 |
dtantsur | As another aside, we need to start sushy-async :) | 14:03 |
JayF | dtantsur: are you still a release manager? If so can you go approve that releases patch | 14:05 |
dtantsur | I'm not | 14:05 |
JayF | frickler: I am +1 to that | 14:05 |
JayF | If any cores object please say so now | 14:05 |
frickler | JayF: dtantsur: TheJulia: maybe we can do this the other way round and have at least one other core +1 the release patch? then I'd feel more comfortable with overriding | 14:38 |
opendevreview | Merged openstack/ironic master: Fix inspection if bmc_address or bmc_v6address is None https://review.opendev.org/c/openstack/ironic/+/928885 | 14:41 |
TheJulia | JayF: Interesting, metal3 use of ipv6? | 14:49 |
JayF | No, basically any ironic tempest job on rackspace flex | 14:50 |
JayF | it starts at a lower MTU dropping us below the limit | 14:50 |
TheJulia | eww | 14:51 |
JayF | exactly | 14:51 |
TheJulia | EWWWWW | 14:51 |
TheJulia | :( | 14:51 |
JayF | you can read backlog on #openstack-qa / #openstack-neutron if you want | 14:51 |
TheJulia | is rackspace flex their new deployment? | 14:51 |
JayF | but basically that's the result | 14:51 |
JayF | yep | 14:51 |
clarkb | as mentioned in that thread it isn't ideal but it also isn't that uncommon (its not the first cloud to do this and people with pppoe + dsl may experience similar if running things locally). I did provide this as feedback to them that they may want to look at jumbo frames in the underlying infrastructure so that neutron networks can have 1500 MTUs | 14:53 |
TheJulia | clarkb: thanks | 15:00 |
opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/26.0: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/929128 | 15:05 |
opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/25.0: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/929129 | 15:05 |
opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/24.0: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/929130 | 15:05 |
opendevreview | Dmitry Tantsur proposed openstack/ironic stable/2024.1: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/929131 | 15:06 |
opendevreview | Dmitry Tantsur proposed openstack/ironic stable/2023.2: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/929132 | 15:07 |
opendevreview | Dmitry Tantsur proposed openstack/ironic stable/2023.1: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/929133 | 15:07 |
opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/26.0: Fix inspection if bmc_address or bmc_v6address is None https://review.opendev.org/c/openstack/ironic/+/929134 | 15:08 |
opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/25.0: Fix inspection if bmc_address or bmc_v6address is None https://review.opendev.org/c/openstack/ironic/+/929135 | 15:08 |
opendevreview | Dmitry Tantsur proposed openstack/ironic bugfix/24.0: Fix inspection if bmc_address or bmc_v6address is None https://review.opendev.org/c/openstack/ironic/+/929136 | 15:08 |
opendevreview | Dmitry Tantsur proposed openstack/ironic stable/2024.1: Fix inspection if bmc_address or bmc_v6address is None https://review.opendev.org/c/openstack/ironic/+/929137 | 15:09 |
opendevreview | Dmitry Tantsur proposed openstack/ironic stable/2023.2: Fix inspection if bmc_address or bmc_v6address is None https://review.opendev.org/c/openstack/ironic/+/929138 | 15:09 |
dtantsur | sorry for spam :D | 15:09 |
TheJulia | ain't spam when it's legitimate :) | 15:18 |
JayF | I worked at an email marketing co for my first job; I was doing "deliverabililty" (aka begging email hosts to not mark us as spam). At the first MAAWG (anti-spam) conference, I gave out lunchboxes with cans of spam promising it was the only spam they'd get from us | 15:22 |
JayF | I always think about that when talking about spam now, the fact I worked at an email marketing company and /literally gave the postmasters for AOL/Yahoo/etc a can of spam/ | 15:22 |
TheJulia | how... meta | 15:31 |
JayF | Well, we were doing a re-launch as a startup, so they were "launch boxes" | 15:35 |
JayF | I'm realizing telling this story in writing that it might be the most jay thing I've ever done | 15:36 |
clarkb | your next campaign could've been something like "We serve only the best spam: in fried rice and musubis" | 15:42 |
JayF | my next campaign involved quitting when a no-grey-area-spammer -- our largest company -- that I convinced the company to fire -- was reinstated when they agreed to pay a multiple of our typical rate | 15:43 |
JayF | s/largest company/largest customer/ | 15:43 |
JayF | I've never quite seen such an explicit demonstration of money directly paying to make someone look the other way for an ethical and reputation issue | 15:43 |
JayF | (... and that led to me being at Rackspace, which led to me being here) | 15:44 |
JayF | TheJulia, others: https://review.opendev.org/c/openstack/ironic/+/928974 needs to land for yoga CI, and then we can merge those followups | 15:46 |
JayF | it'd be extremely helpful to me if we can try to get these unmaintained fixup patches in back to xena today, or by eow | 15:49 |
TheJulia | ack, I just approved it | 15:52 |
JayF | you want me to remove the req bump from your patch or will you? | 15:53 |
TheJulia | lets keep rolling then, I'm *almost* done with my train version of the nightmare | 15:53 |
JayF | either way one of us who doesn't has to +2a | 15:53 |
TheJulia | JayF: if you could, that would be kickass, I'd approve it anyway as it is an unmaintained branch at this point but we both care and that is good enough for me | 15:54 |
JayF | ack | 15:54 |
JayF | am I crazy? I thought https://review.opendev.org/c/openstack/ironic/+/928709 had a requirements change with it | 15:55 |
JayF | guess not, easier this way anyway | 15:58 |
TheJulia | nah, I was trying to repro locally and complaining in irc | 15:58 |
JayF | I think we could justify a setuptools addition to u-c in yoga | 15:59 |
JayF | but those u-c updates in yoga have been agressively nacked in the past | 15:59 |
TheJulia | yeah, keeping it working in current state on an uncapped dependency is more important | 16:00 |
JayF | I agree with "don't change requirements" | 16:01 |
JayF | but for things like setuptools, I feel like that's mostly a "shipped with OS" thing | 16:01 |
TheJulia | we really are not, from what it was developed with | 16:01 |
TheJulia | yeah | 16:01 |
JayF | really only developers and weirdoes running gentoo hit it ;) | 16:01 |
clarkb | setuptools is also special because people think that putting it in requirements or constraints allows you to control its version (it does not) | 16:03 |
clarkb | you haev to use pyproject.toml to do that reliable | 16:03 |
* TheJulia finds issue on train backport and gets the wax and buffer back out | 16:04 | |
JayF | clarkb: having it there matters in tox cases, afaict | 16:05 |
JayF | clarkb: and that's mainly the only case | 16:05 |
JayF | which is why in that yoga fix patch, I put setuptools<71 in the tox.ini instead of requirements | 16:05 |
clarkb | that only works if people use tox right. But if you just pip install kaboom | 16:06 |
clarkb | if you want pip install to reliably fetch setuptools you need to use pyproject.toml to specify the dep | 16:06 |
JayF | which is not a change we're going to backport | 16:06 |
JayF | which is why we end up in these weird situations :/ | 16:07 |
clarkb | ya I'm not saying you should I'm just calling out why you have to be careful with pretending to manage setuptools | 16:07 |
clarkb | it gives people the feeling of a solution when it doesn't work in the most common case | 16:07 |
JayF | Yeah. It's not easy trying to maintain software for a half-decade when it's built on a quicksand foundation ;/ | 16:08 |
JayF | python packaging tools have yolo-broken more packages that I work on than any other set of libraries combined | 16:08 |
opendevreview | OpenStack Release Bot proposed openstack/ironic bugfix/22.0: Update .gitreview for bugfix/22.0 https://review.opendev.org/c/openstack/ironic/+/929149 | 16:14 |
JayF | I approved dmitry's changes above to unblock CI across our branches. Once they land I'll evaluate if we need to send 'em to unmaintained, too | 16:18 |
opendevreview | Merged openstack/ironic master: Set node "alive" when inspection finished https://review.opendev.org/c/openstack/ironic/+/927828 | 16:46 |
opendevreview | Merged openstack/ironic bugfix/26.0: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/929128 | 16:46 |
opendevreview | Merged openstack/ironic bugfix/25.0: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/929129 | 16:46 |
opendevreview | Merged openstack/ironic bugfix/24.0: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/929130 | 16:46 |
TheJulia | JayF: what do you need from me right now for the branch patches? | 17:03 |
JayF | nothing, I think | 17:03 |
JayF | I went ahead and +2A all of those UM patches so we could make them get landed; the yoga one needs a recheck once my fix-it patch lands | 17:04 |
JayF | but I expect we'll have em all landed by eod | 17:04 |
JayF | I'll ping ya if I need something specific :D | 17:04 |
JayF | TheJulia: one question: is there a reason the cve follow-up isn't on victoria? Just making sure before I cherry-pick it | 17:08 |
TheJulia | I likely forgot | 17:09 |
TheJulia | I've been getting pulled in a few different directions the past few days | 17:09 |
JayF | no problem, I'm on it | 17:09 |
JayF | just making sure no unknowns | 17:09 |
TheJulia | I think it was more trying to figure out what was going on with yoga | 17:09 |
TheJulia | Thank you for your attention to detail and follow-up | 17:09 |
JayF | https://review.opendev.org/c/openstack/ironic/+/929165 probably best you +2A this since I picked it | 17:10 |
TheJulia | done | 17:16 |
JayF | ty | 17:20 |
opendevreview | Merged openstack/ironic stable/2024.1: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/929131 | 17:53 |
opendevreview | Merged openstack/ironic stable/2023.2: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/929132 | 17:56 |
JayF | https://review.opendev.org/q/topic:%22mtu-fix%22 I cherry-picked those MTU fixes onto unmaintained before we lose context; they are there if you wanna review but CI is still running. | 18:09 |
TheJulia | ++ | 18:24 |
TheJulia | Thanks | 18:24 |
TheJulia | I'm adding some stuff to the PTG etherpad before I forget | 18:24 |
JayF | FYI 11-14 days until final Ironic releases *must* be cut | 18:25 |
JayF | so if we need to land something, we should do it soon | 18:25 |
TheJulia | ++ | 18:25 |
TheJulia | I'm sure that will be "release right meow!" next monday. | 18:25 |
opendevreview | Merged openstack/ironic stable/2023.1: Try limiting MTU to at least 1280 https://review.opendev.org/c/openstack/ironic/+/929133 | 18:28 |
TheJulia | Folks: reminder, for PTG etherpad stuffs, put an estimated amount of time on the item | 18:35 |
TheJulia | JayF: did the move of ironic to formally be part of VMT ever go forward or was it pending cores to +1.. or what?! | 18:36 |
JayF | TheJulia: sometimes hard to predict the time until we see the async discussion; I know when I was PTL I often filled in expected times across many topics | 18:37 |
TheJulia | True, yeah | 18:37 |
TheJulia | but starting to scope it early is really critical | 18:38 |
TheJulia | because then your saying "this could be big" | 18:38 |
TheJulia | or "this should be quick" | 18:38 |
TheJulia | and when some of that async starts pushing back, it is easy to know | 18:38 |
JayF | https://review.opendev.org/c/openstack/ironic-python-agent/+/927985 and friends need review, trivial, just adding the "we are insecure" warning to older ipas | 18:38 |
JayF | https://review.opendev.org/c/openstack/ossa/+/928005 I have that W-1 until after rpittau is around to approve it | 18:39 |
JayF | I know he will be unlikely to have opposition to it, but I want to respect that he gets a vote :D | 18:39 |
JayF | I need to also check the vmt checklist, ensure we've done all the things there | 18:40 |
* JayF adds it to his list for tomorrow, along with doing an ironic security vid | 18:40 | |
JayF | I put general time based comments on mine | 18:41 |
TheJulia | cool cool, thanks | 18:45 |
TheJulia | I think my cat supervisor thinks I need to call it a day soon | 18:49 |
cardoe | TIL loading interfaces can swallow exceptions and exit silently | 19:36 |
cardoe | https://github.com/rackerlabs/understack/pull/275 | 19:45 |
TheJulia | yup, the exception hits the surface | 19:50 |
TheJulia | Okay, two trips, PTG, and Jury duty now on my calandar for october. | 22:04 |
TheJulia | Please, nothing else. | 22:04 |
clarkb | maybe you'll get lucky and they'll say they don't need jurors on the day you are supposed to go in | 22:05 |
TheJulia | It is the whole week, but depending on how they feel I might have to report each and every day or not | 22:07 |
TheJulia | best case, they just punt it each day | 22:07 |
clarkb | oh wow in my county it has only ever been a single day unless you get picked | 22:08 |
clarkb | then if picked it is for the duration of the trial | 22:08 |
TheJulia | It really depends with what is going on, but similar most of the time. | 22:12 |
TheJulia | Go in, fill out form. Either get disqualified in the first or second round of paperwork. | 22:12 |
JayF | The experience here in Washington is similar to what Julia experienced... I was basically told to be prepared to come in to be in a jury pool for four out of the five business days of the week, and if we got selected for a trial we would serve and no matter how long it took we were done | 22:17 |
JayF | I luckily was struck from having to be on a case that would have been extremely unpleasant to be a juror on | 22:17 |
cardoe | PTG gonna be available virtually too? | 22:49 |
TheJulia | It will be entirely virtual | 23:24 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!