| clarkb | like gitea ^ that hit the 404 against the buildest registry error that we think is actually the docker hub rate limit error | 00:34 |
|---|---|---|
| clarkb | unlike the system-config run jobs we don't force these jobs to talk to ipv4 docker hub | 00:34 |
| clarkb | but maybe we should? | 00:34 |
| clarkb | anyway I'll recheck that change now and cross my fingers but its dinner time so won't be able to followup until tomorrow likely | 00:36 |
| opendevreview | Merged opendev/system-config master: Add setuptools to python-builder assemble https://review.opendev.org/c/opendev/system-config/+/946722 | 04:16 |
| mnaser | clarkb: I think guilhermesp can help update them for you | 05:29 |
| *** mrunge_ is now known as mrunge | 07:23 | |
| opendevreview | Jaromír Wysoglad proposed openstack/project-config master: Add Aetos under the telemetry project https://review.opendev.org/c/openstack/project-config/+/946739 | 07:52 |
| opendevreview | Jaromír Wysoglad proposed openstack/project-config master: Add jobs for new Aetos repository https://review.opendev.org/c/openstack/project-config/+/946742 | 08:17 |
| opendevreview | Jaromír Wysoglad proposed openstack/project-config master: Add Aetos under the telemetry project https://review.opendev.org/c/openstack/project-config/+/946739 | 08:27 |
| opendevreview | Jaromír Wysoglad proposed openstack/project-config master: Add Aetos under the telemetry project https://review.opendev.org/c/openstack/project-config/+/946739 | 08:49 |
| *** Guest12857 is now known as dhill | 11:35 | |
| *** dhill is now known as Guest13235 | 11:36 | |
| jamesdenton | clarkb i'm sorry, i squirrel'd and lost track of that issue. Didn't really have a chance to make progress on it. If you are OK with deleting it, that's fine | 13:30 |
| Clark[m] | jamesdenton: I'm happy either way. If you think it is important to debug further we can keep it around | 13:31 |
| jamesdenton | yeah, lemme take another stab at it. Can you get me the UUID, if you have it? | 13:32 |
| fungi | i can get it | 13:34 |
| fungi | jamesdenton: 7e608502-6e33-4a59-aa1a-b4e2257d0a11 | 13:37 |
| jamesdenton | thank you | 13:37 |
| guilhermesp | o/ ive heard you guys needs some reserve dns for some ips? | 13:40 |
| fungi | guilhermesp: reverse, yes | 13:41 |
| fungi | guilhermesp: these two ip addresses (one v4, one v6) should reverse to review03.opendev.org: 162.253.55.233 2604:e100:1:0:f816:3eff:fe31:926a | 13:43 |
| guilhermesp | cool i will arrange that and let you know once its done | 13:43 |
| fungi | we're already serving the corresponding a/aaaa rrs: https://opendev.org/opendev/zone-opendev.org/src/branch/master/zones/opendev.org/zone.db#L690-L691 | 13:43 |
| fungi | thanks! | 13:43 |
| opendevreview | Jaromír Wysoglad proposed openstack/project-config master: Add Aetos under the telemetry project https://review.opendev.org/c/openstack/project-config/+/946739 | 14:02 |
| opendevreview | Jaromír Wysoglad proposed openstack/project-config master: Remove telemetry groups https://review.opendev.org/c/openstack/project-config/+/946767 | 14:02 |
| opendevreview | Merged openstack/project-config master: Add Aetos under the telemetry project https://review.opendev.org/c/openstack/project-config/+/946739 | 15:42 |
| JayF | Is there any way to persist or download the log of a PTG session chat? | 15:44 |
| JayF | (from meetpad) | 15:44 |
| fungi | i do not know | 15:44 |
| fungi | though if you re-enter the same room i believe it will persist | 15:45 |
| fungi | looks like there's a firefox extension for doing it: https://addons.mozilla.org/en-US/firefox/addon/jitsi-chat-exporter/ (came up in a random web search) | 15:46 |
| fungi | https://community.jitsi.org/t/how-to-how-to-save-chats-and-speaker-stats/80904 talks about a server-side plugin | 15:47 |
| JayF | to others: doon't use this | 15:47 |
| JayF | it blanked existing chat in the window and only copied timestamps to my clipboard | 15:48 |
| JayF | reload did get me the messages back though | 15:48 |
| fungi | yeah, i can't make recommendations, right now i'm just serving as a meat-based search engine proxu | 15:48 |
| fungi | proxy | 15:48 |
| JayF | yeah, nbd just closing the loop in case other folks see this | 15:49 |
| JayF | looks like there's no great answer which is fine, we just need to be better at 301-redirecting live chat commentary to the etherpad | 15:49 |
| guilhermesp | https://www.irccloud.com/pastebin/zMSchARw/ | 16:40 |
| guilhermesp | btw cc fungi :) | 16:41 |
| fungi | thanks guilhermesp! | 16:42 |
| fungi | i can confirm they're resolving for me now too | 16:43 |
| fungi | clarkb: ^ | 16:43 |
| *** dansmith is now known as dansmith_pto | 16:52 | |
| corvus | fungi: are we ready for https://review.opendev.org/946274 ? (i'm slightly out of the loop) | 17:42 |
| fungi | corvus: oh! yes, past due, sorry | 17:46 |
| opendevreview | Merged opendev/zuul-providers master: Revert "Disable rax-flex-dfw3" https://review.opendev.org/c/opendev/zuul-providers/+/946274 | 17:46 |
| fungi | we merged the corresponding revert for project-config a while ago | 17:46 |
| clarkb | fungi: guilhermesp thank you! | 17:49 |
| clarkb | fungi: I'm back now and I think we can reapprove/reenqueue the gitea upgrade? | 17:49 |
| fungi | i'm going to be afk betwen ~18:00-19:00 but maybe it won't deploy before then anyway | 17:51 |
| clarkb | ya the jobs take about an hour minimum | 17:51 |
| clarkb | shoudl I enqueue to the gate or recheck? | 17:54 |
| clarkb | if I recheck it will definitely be done well after you get back. Maybe thats the best option | 17:54 |
| clarkb | rechecked | 17:55 |
| fungi | yeah, sorry, trying to wrap up a ptg session i'm leading | 17:58 |
| fungi | rcheck is fine | 17:58 |
| clarkb | ah sorry, I've admittedly been a bit disconnected from PTG happenings this time around | 17:58 |
| opendevreview | James E. Blair proposed opendev/zuul-providers master: Build vhd images https://review.opendev.org/c/opendev/zuul-providers/+/946799 | 17:58 |
| corvus | that's still needed for rax classic, right? | 17:59 |
| clarkb | yes | 17:59 |
| clarkb | +2 from me | 17:59 |
| corvus | zuul-launcher seems to think so -- it's in a rebuild loop because we keep running that job and not producing a vhd image :) | 17:59 |
| clarkb | corvus: the conversion needs to use the specail built tool too | 17:59 |
| clarkb | corvus: not sure if that is installed by the jobs but if not it will need to be | 17:59 |
| corvus | if we just tell dib to do that will it work? | 18:00 |
| clarkb | corvus: dib knows to use vhd-util, but that has to be installed. We install it to the nodepool container images but I'm not sure about your jobs | 18:00 |
| corvus | our jobs :) | 18:00 |
| clarkb | ++ | 18:01 |
| fungi | okay, disappearing, should return in an hour-ish | 18:01 |
| clarkb | corvus: the nodepool dockerfile does echo "deb http://ppa.launchpad.net/openstack-ci-core/vhd-util/ubuntu focal main" >> /etc/apt/sources.list then an apt get update then installs vhd-util | 18:01 |
| clarkb | I think if you install that toolchain then dib will automatically use it | 18:01 |
| corvus | so we probably just need an ansible package task for that in the job | 18:01 |
| corvus | i don't suppose we have any ansible that installs that? so i can just copypasta it... | 18:02 |
| clarkb | I'm not aware of any | 18:03 |
| clarkb | I think we've baked it into the container images so we don't have to think about it. Another option is to pull the container image and exec the command out of it. I think that is what tonyb did when converting images recently | 18:03 |
| corvus | dib container images? | 18:04 |
| clarkb | corvus: nodepool-builder container images | 18:04 |
| clarkb | but now that youv'e said dib container images maybe we should be building those | 18:04 |
| corvus | well that is not a long-term solution :) | 18:04 |
| clarkb | ya I think building dib images is somethign we should consider but they don't exist yet | 18:05 |
| corvus | (there will not be nodepool container images in future) | 18:05 |
| opendevreview | James E. Blair proposed opendev/zuul-providers master: Build vhd images https://review.opendev.org/c/opendev/zuul-providers/+/946799 | 18:13 |
| corvus | clarkb: ^ does that look right? | 18:13 |
| clarkb | corvus: yes, the only thing I wonder about is apt_repository might default to using the codename for the node it runs against and we may need toforce it to focal? | 18:16 |
| clarkb | http://ppa.launchpad.net/openstack-ci-core/vhd-util/ubuntu/dists/ ya focal is the latest | 18:17 |
| corvus | oh we're running "focal" packages on something else? | 18:17 |
| clarkb | ya | 18:17 |
| clarkb | corvus: its a tool from xenserver that mordred patched to do the two phase vhd conversion that rax needs built into a package that we last built for focal | 18:17 |
| clarkb | its all kinds of pain but it works and has for some time | 18:17 |
| corvus | yeah i remember that, just not which things it's built for | 18:18 |
| mordred | might not be terrible to make newer versions of that? | 18:18 |
| clarkb | mordred: ya its possible we can reupload the existing source package stuff for newer releases and it would work. I'm not sure | 18:18 |
| mordred | also - IIRC - the functionality that's in that MIGHT be in one of the normal utilites now | 18:19 |
| clarkb | mordred: the last time we checked we had to revert | 18:19 |
| opendevreview | James E. Blair proposed opendev/zuul-providers master: Build vhd images https://review.opendev.org/c/opendev/zuul-providers/+/946799 | 18:19 |
| mordred | there were outstanding patches 5 years ago and I feel like I remember at least one of them being close | 18:19 |
| clarkb | and I think ianw recentlyish double checked if that was fixed and decided it wasn't? | 18:19 |
| corvus | there's a "codename" argument so i set it to focal | 18:19 |
| mordred | sigh | 18:19 |
| clarkb | corvus: +2 | 18:20 |
| corvus | clarkb: cool, i'll +3 -- it's very self-testing | 18:20 |
| clarkb | infra-root as far as timing for landing the change to add review03 to the inventory goes I think tomorrow is good for me. I expect I will be home all day (unlike today) | 18:25 |
| clarkb | so if you've got time to sanity check that doing so between now and then would be great | 18:25 |
| clarkb | looking at grafana it looks like rax-flex sjc3 might be capping out at ~8 or 9 nodes | 18:42 |
| clarkb | I wonder if we leaked resources there | 18:42 |
| clarkb | server list there returns a 500 error | 18:44 |
| clarkb | dan_with: jamesdenton ^ not to pile on but I just noticed that when trying to debug why we aren't booting as many nodes as expected. limits show --absolute does seem to indicate we are using most of our memory so I do suspect we have leaked servers but I can't list them currently | 18:44 |
| corvus | (if the api under server list returns an error, then nodepool or zuul-launcher can't launch them either -- or, rather, mark them as ready after booting them) | 18:53 |
| clarkb | ah ya so maybe nodepool is in a bit of a limbo due to that specific error and fixing that would allow nodepool to do its thing and clean up | 18:55 |
| corvus | that too. if zuul-launcher leaked them... cleanup isn't implemented yet. that's #2 on my list. :) | 18:55 |
| jamesdenton | getting a 500 error on list? | 18:58 |
| clarkb | jamesdenton: yes `openstack server list` | 18:58 |
| clarkb | <class 'KeyError'> | 18:59 |
| jamesdenton | right, ok. i've seen this before but need to track it down | 18:59 |
| clarkb | /v2.1/tenantuuid/servers/detail?deleted=False seems to be the request it is making | 19:00 |
| clarkb | using our other tenant listings work | 19:00 |
| jamesdenton | there is likely something about a particular instance causing that. Not sure if we tracked it down to a client bug or not | 19:01 |
| opendevreview | Tim Burke proposed opendev/yaml2ical master: Add ability to have per-meeting chairs https://review.opendev.org/c/opendev/yaml2ical/+/946824 | 19:23 |
| clarkb | corvus: jamesdenton has pointed out three servers in an ERROR state that are expected to be the problem with listing things. All three appear to be booted by zuul-launcher not nodepool (based on non sequential npXYZ digits and the flavors being small and large) | 19:24 |
| clarkb | corvus: jamesdenton the reason they are in ERROR is due to hitting quota limits which I think is expect with zuul-launcher and nodepool both competing for the quota | 19:25 |
| fungi | back, errand took longer than expected | 19:26 |
| clarkb | still well before gitea is done gating | 19:27 |
| fungi | yeah, looks that way | 19:28 |
| fungi | system-config-run-gitea just now starting in gate | 19:28 |
| timburke | cross-post from #openstack-dev: i sent an email yesterday about swift's ops-feedback session happening tomorrow, but i haven't seen it show up in the discuss archives -- is it still held for moderation or something? | 19:29 |
| clarkb | fungi: ^ is typically a moderator for the openstack discuss list | 19:31 |
| clarkb | corvus: jamesdenton: so I think the reason nodepool graphs show less usage is zuul-launcher is also consuming the quota. Thats all fine and doing a server list I'm able to see that now (they fixed something on the backend and ti works) | 19:32 |
| clarkb | so other than the error'd nodes and failed listing I think this si working as expected | 19:32 |
| corvus | ack -- so we should try manually deleting those since zl won't do it for us yet | 19:37 |
| clarkb | corvus: rax already did it for us. That was the fix for server listings | 19:38 |
| clarkb | which puts us in an awkward spot where we can't actually see the broken nodes to delete them. But I don't think zuul-launcher would either... | 19:39 |
| clarkb | My hunch is this is a nova bug when you have multiple simultaneous requsets and some subset go over quota | 19:39 |
| corvus | oh this doesn't sound good? | 19:39 |
| clarkb | but that is very hand wavy | 19:39 |
| corvus | like, it sounds really likely that we will keep producing ERROR servers with no recourse to delete them? | 19:39 |
| clarkb | yes that seems likely. We can probably mitigate by using max-servers values that are smaller in nodepool and allow zuul-launcher to fill that space? | 19:40 |
| corvus | yeah, but zuul-launcher will, by design, occassionally exceed quota | 19:40 |
| corvus | so if we're in a situation where we must absolutely never hit quota otherwise the cloud breaks... then we need to limit it far below the real quota, or something similarly drastic | 19:41 |
| jamesdenton | clarkb fungi corvus looks like those issues might be a result of https://bugs.launchpad.net/nova/+bug/2095364 | 19:41 |
| jamesdenton | waiting on another +2 and backport | 19:42 |
| clarkb | oh that bug may point to a workaround. If we use an older microversion to list servers then we'd see them maybe? | 19:42 |
| jamesdenton | MAYBE it can be worked around by specifiying microversion w/ the client? | 19:42 |
| jamesdenton | beat me to it | 19:42 |
| clarkb | ya I'm going to test that now. Though I can't test if it fails but if I get back data with an older microversion then that would probably work | 19:42 |
| clarkb | corvus: ^ and if that does work then zuul-launcher and us have an out | 19:42 |
| jamesdenton | worth a shot | 19:42 |
| corvus | ++ | 19:43 |
| clarkb | I just have to figure out how to incant the microversion version (the docs say a bunch of things about what versiosn you need to support which flags but no obvious this is how you set the version info...) | 19:44 |
| corvus | keep in mind that nodepool and zl both have their own exceedingly optimized server list method (they don't use sdk; only the keystone session object). not sure if that makes that easier or harder. | 19:44 |
| corvus | clarkb: ^ i mention that to say: don't waste too much time figuring out how to do it with sdk; look directly at the nodepool/zl driver instead | 19:45 |
| fungi | clarkb: yeah, streams crossed, i started replying in #openstack-dev because i was pinged there first | 19:45 |
| corvus | clarkb: https://opendev.org/zuul/nodepool/src/branch/master/nodepool/driver/openstack/adapter.py#L721 | 19:45 |
| clarkb | corvus: ack | 19:45 |
| corvus | that's identical in niz | 19:45 |
| clarkb | the docs I found seem to say use --os-compute-api-version but if I set that to 2.95 it says that isn't in 2 or 2.1 | 19:47 |
| clarkb | so I think that is trying to select the major api version not the microversion | 19:47 |
| clarkb | looking in the code openstackclient hardocdes a lot of requests to microversion 2.1 | 19:54 |
| clarkb | but it still isn't clear to me how you as the user say "use this microversion for this request" | 19:55 |
| clarkb | since apparently --os-compute-api-version only allows for 2 or 2.1 | 19:55 |
| jamesdenton | i did find that my openstackclient will append a header for the microversion: -H "OpenStack-API-Version: compute 2.54", for example | 19:56 |
| jamesdenton | openstack --os-compute-api-version 2.54 --os-cloud default server list | 19:57 |
| clarkb | -H "X-OpenStack-Nova-API-Version: 2.95" | 19:57 |
| clarkb | ok I see that too | 19:57 |
| clarkb | so the bug here is in UX saying that is an invalid version then it sends the header anyway | 19:57 |
| jamesdenton | ok yeah, actually two headers...-H "X-OpenStack-Nova-API-Version: 2.54" and -H "OpenStack-API-Version: compute 2.54" | 19:58 |
| opendevreview | Tim Burke proposed opendev/irc-meetings master: swift: Fix meeting channel and chair https://review.opendev.org/c/opendev/irc-meetings/+/946833 | 19:58 |
| opendevreview | Tim Burke proposed opendev/irc-meetings master: Swift: Reinstate 0700 meeting slot https://review.opendev.org/c/opendev/irc-meetings/+/946834 | 19:58 |
| corvus | if that's all we have to do, then it should be very easy to drop that into simpleServerList | 19:59 |
| clarkb | corvus: reading https://docs.openstack.org/api-guide/compute/microversions.html#client-interaction I think zuul-launcher and nodepool-launcher may be ok as is beacuse that doc says nova will use the oldest possible micrversion if not specified | 19:59 |
| corvus | ohhh nice | 19:59 |
| clarkb | corvus: and we're just doing a get against a url there I don't think that will get us those headers? So this is a set of human interface bugs but the automated tools are ok I suspect | 20:00 |
| corvus | cool, we can probably double check that with logs from the error time period | 20:00 |
| corvus | would have to be from nodepool since i don't think we launched anything with zl lately | 20:00 |
| corvus | but basically: if we don't have any nodepool-launcher tracebacks in server list during that time, we should be good | 20:01 |
| clarkb | I think there are 11 zl launched nodes in sjc3 right now. I don't know how old they are and maybe they too leaked? | 20:01 |
| clarkb | let me see if nodepool logs from just before 1900 show successful listings | 20:02 |
| opendevreview | Merged opendev/system-config master: Update gitea to 1.23.7 https://review.opendev.org/c/opendev/system-config/+/946640 | 20:03 |
| clarkb | corvus: 2025-04-09 18:50:01,589 DEBUG nodepool.OpenStackAdapter.raxflex-sjc3: API call detailed server list in 0.6331690889783204 | 20:03 |
| clarkb | let me triple check there aren't any exceptions around that message | 20:03 |
| clarkb | ya it seems clean so I suspect that indicates we are ok | 20:04 |
| clarkb | the last question I have is whether or not those 11 zl nodes need manual cleanup or not | 20:04 |
| opendevreview | Merged opendev/irc-meetings master: swift: Fix meeting channel and chair https://review.opendev.org/c/opendev/irc-meetings/+/946833 | 20:05 |
| clarkb | I'll haev to switch gears and double check that gitea upgrade nowish (its behind hourly updates but taht shouldn't take long) | 20:06 |
| fungi | looks like the gitea deploy is behind the hourly jobs, but those are going to be done at any moment | 20:07 |
| fungi | underway now | 20:11 |
| fungi | container images are pulling | 20:11 |
| fungi | and now being stopped on 09 | 20:11 |
| clarkb | only ssh has stopped though the others are in the process fo exiting | 20:12 |
| clarkb | and up again | 20:13 |
| clarkb | https://gitea09.opendev.org:3081/opendev/system-config/ lgtm I'm going to test a clone now | 20:13 |
| fungi | Powered by Gitea Version: v1.23.7 | 20:13 |
| clarkb | clone worked for me | 20:14 |
| fungi | yeah, working here from 09 too | 20:14 |
| clarkb | 10 is done now too. So far this is looking happy | 20:15 |
| fungi | yeah, so far, so good | 20:15 |
| clarkb | all are done now | 20:21 |
| clarkb | and the job just succeeded | 20:21 |
| fungi | and everything seems to be working fine for me | 20:21 |
| clarkb | the last major item on my mental checklist is seeing replication occur | 20:21 |
| clarkb | I don't see any recent changes and don't have anything loacl I can push other than a throwaway | 20:23 |
| clarkb | anyone have a change to update that is more interesting than a noop change from me? | 20:23 |
| fungi | i do not | 20:24 |
| fungi | https://review.opendev.org/c/starlingx/update/+/946835 | 20:26 |
| clarkb | perfect | 20:26 |
| fungi | https://opendev.org/starlingx/update/commit/f1ea563 | 20:27 |
| clarkb | that did replicate (I did my git fetch origin refs/changes/35/946835/2 && git show FETCH_HEAD thing because I can never remember the url) | 20:28 |
| fungi | ignore the commit time on it, that's when it was created locally, not when it was pushed | 20:28 |
| fungi | i find it easier to grab the commit id for the patchset and then load that commit directly in gitea's webui | 20:29 |
| fungi | rather than fussing with the refs/changes named refs | 20:30 |
| clarkb | ya for whatever reason I can remember the magical git refs but not the gitea urls | 20:30 |
| fungi | i'm not sure there even is a gitea url to pull them up that way like cgit and gitweb could | 20:30 |
| opendevreview | Tim Burke proposed opendev/yaml2ical master: Add ability to have per-meeting chairs https://review.opendev.org/c/opendev/yaml2ical/+/946824 | 20:31 |
| fungi | gitea does some odd custom filtering of the refs tree | 20:31 |
| fungi | if we get consensus on 946824 then that will be a good opportunity to go ahead and do another yaml2ical release | 20:32 |
| fungi | since there's an irc-meetings change pending that depends on it | 20:33 |
| opendevreview | Tim Burke proposed opendev/irc-meetings master: Allow per-schedule meeting chairs https://review.opendev.org/c/opendev/irc-meetings/+/946838 | 20:34 |
| clarkb | I have to say yaml2ical being in opendev feels like an odd fit when it comes to changes like this beacuse opendev is very much the opposite in terms of making everythign liek meeting chairs rigid. | 20:34 |
| clarkb | Its probably a good feature it just occurs to me that we're apparently maintainnig software we basically never use | 20:35 |
| fungi | yeah, when we shook up the openstack-infra snowglobe there were clearly things than landed on the opendev and the openstack sides of the landscape, but also some stuff that was near the intersection that might have gotten misclassified | 20:37 |
| fungi | yaml2ical and irc-meetings repos are part of that questionable middle ground | 20:38 |
| clarkb | I approved the yaml2ical change. The irc-meetings change has a small typo in it that I called out | 20:42 |
| fungi | i tend to look at it as features that meta-opendev (sysadmins and systems automation) doesn't use heavily but projects we host still rely on | 20:45 |
| clarkb | ya and I guess for the general functioanlity they do. But fancy daylight savings time in particular came up recently and got me thinking I don't want to be responsible for getting that right | 20:47 |
| clarkb | it feels like the tool is falling away from general purpose usability to opensatck wants it this way | 20:47 |
| clarkb | (which is fine it just feels weird for me to be on the hook for that) | 20:48 |
| fungi | i don't even think a majority of openstack wants meetings that follos some arbitrary timezone, just a subset of contributors do | 20:48 |
| fungi | if i were in au i'd be extra annoyed by meetings following eu or usa dst | 20:49 |
| clarkb | ya a two hour swing instead of one | 20:49 |
| fungi | it seems like a case of convenience over inclusivity | 20:50 |
| opendevreview | Tim Burke proposed opendev/irc-meetings master: Allow per-schedule meeting chairs https://review.opendev.org/c/opendev/irc-meetings/+/946838 | 20:50 |
| fungi | it's funny-sad to see people talk about wanting more engagement from potential contributors in south america and africa, but then nailing team meetings to northern hemisphere timekeeping | 20:52 |
| JayF | fungi: at some point folks give up on having alternate times with minimal participation; I know that is why Ironic pinned to a single time (which is kinda EU/US friendly but not as APAC friendly) | 20:54 |
| opendevreview | Merged opendev/yaml2ical master: Add ability to have per-meeting chairs https://review.opendev.org/c/opendev/yaml2ical/+/946824 | 20:55 |
| fungi | yeah, but for example chile and california are in compatible timezones if only they weren't on different sides of the equator | 20:55 |
| JayF | I am confused by that comment | 20:56 |
| fungi | i was talking about differencs between northern vs southern hemisphere annual timezone changes | 20:56 |
| fungi | latitude not longitude | 20:57 |
| fungi | there was a suggestion that we e.g. make it easier for a team to have meetings auto-adjust to follow usa dst or eu dst annual changes rather than a year-round tz like utc | 20:58 |
| fungi | and i was saying that i'd be annoyed if i were trying to contribute from the southern hemisphere and everyone just assumd my timezone changes the same ways theirs does on the other side of the equator | 20:59 |
| timburke | the only thing worse than having your meeting time shift by an hour twice a year is having it shift by one to two hours four times a year | 20:59 |
| fungi | also people talking about a "summer schedule" when that's actually winter for half the globe | 21:00 |
| fungi | or fall vs spring | 21:00 |
| JayF | ooh yeah | 21:00 |
| JayF | I would *never* want to move meeting UTC times with DST | 21:01 |
| fungi | there's an unfortunate majority of "developed" countries on one side of the equator who are either prenending that the other side of the planet is irrelevant or merely ignorant to how the whole concept works | 21:01 |
| JayF | but some of that is also working for a UK-based company and experiencing the awkward-weeks of only half of us being DST | 21:01 |
| fungi | s/prenending/pretending/ | 21:03 |
| fungi | but yeah, i think we can be opinionated and draw a line between features like time-specific meeting chairs vs meeting schedule options to track dst changes for some specific point on the globe | 21:04 |
| fungi | i see the periodic changes for teams that reschedule their utc meeting time twice a year to avoid unrelated meeting conflicts among team members as a feature not a bug. they're being explicit that while the meeting time is being kept constant in local time for some set of people it's actually changing utc time | 21:06 |
| opendevreview | Merged opendev/irc-meetings master: Allow per-schedule meeting chairs https://review.opendev.org/c/opendev/irc-meetings/+/946838 | 21:08 |
| fungi | i recently watched the worse version of that go down for a recurring meeting i participate in, because the meeting host scheduled it in a usa timezone but wanted it to follow eu dst changes, so they had to "cancel" the "normal" (usa) meeting time for several weeks and schedule a temporary alternate meeting at a "different" time until eu dst started | 21:09 |
| fungi | tracking international meetings in utc is far less complicated by comparison | 21:12 |
| opendevreview | Merged opendev/zuul-providers master: Build vhd images https://review.opendev.org/c/opendev/zuul-providers/+/946799 | 21:14 |
| clarkb | etherpad has a new release (perfectly timed for the ptg) and I thought it would be an easy update the the dockerfile is much different | 21:20 |
| fungi | here's the changes for yaml2ical since the most recent (0.13.0) tag: https://paste.opendev.org/show/bua5N3eueF58SaH9NKYd/ | 21:24 |
| fungi | i think a 0.14.0 would make sense | 21:24 |
| fungi | the most recent change is the only one of substance | 21:26 |
| fungi | infra-root: objections to me tagging yaml2ical commit 630657cafd6600d6ac40452a7c917ea75ac18ff5 as 0.14.0 ? | 21:35 |
| opendevreview | Clark Boylan proposed opendev/system-config master: Update Etherpad to 2.3.0 https://review.opendev.org/c/opendev/system-config/+/946842 | 21:36 |
| fungi | this will also be a good secondary test of the newer-style python packaging we used in recent bindep which i've replicated in yaml2ical recently | 21:36 |
| clarkb | no objections from me (but I didn't check the hash or the version yet) | 21:36 |
| fungi | this will be the first release in about 3.5 years | 21:37 |
| clarkb | I marked the etherpad change WIP and made a couple notes for thigns I noticed | 21:38 |
| clarkb | fungi: the version and sha both look correct to me | 21:38 |
| clarkb | I'm going to pop out shortly for a bike ride. Its warm and the sun is coming out so I will take advantage | 21:39 |
| fungi | have fun! i'm around but winding down for the day | 21:39 |
| fungi | commit 630657cafd6600d6ac40452a7c917ea75ac18ff5 (HEAD, tag: 0.14.0, origin/master, origin/HEAD, gerrit/master, gerrit/HEAD) | 21:40 |
| fungi | i've pushed it | 21:41 |
| fungi | now to see if the recent job and tenant changes still work | 21:41 |
| fungi | it queued opendev-release-python so that's a good start | 21:41 |
| fungi | finally running | 21:47 |
| fungi | uploaded successfully, looks like | 21:48 |
| fungi | https://pypi.org/project/yaml2ical/ shows the new release | 21:49 |
| fungi | the new package metadata seems to have been parsed correctly there | 21:49 |
| fungi | exercising with a recheck of 946834 | 21:50 |
| fungi | #status log Released yaml2ical 0.14.0 | 21:51 |
| opendevstatus | fungi: finished logging | 21:51 |
| fungi | strangely it still used yaml2ical-0.13.0 | 21:57 |
| timburke | delay in updating mirrors? | 21:57 |
| fungi | maybe too fast for the proxy cache to realize | 21:58 |
| fungi | yeah, i don't see any smoking gun in the log to explain it skipping a present 0.14.0 so i'll just give it a few minutes and try again | 22:00 |
| fungi | i think our apache configs are set to cache indices for up to 10 minutes? | 22:01 |
| fungi | anyway, it's not like it's running on a too-old cpython or constraining/pinning the dependency | 22:01 |
| fungi | so it should "just work" | 22:01 |
| fungi | timburke: okay, that's better, now the only failure is that something's still not quite right in tools/check_chair.py | 22:12 |
| timburke | i'll work on it some | 22:13 |
| timburke | thanks for getting the release out! | 22:13 |
| fungi | my pleasure. i'm glad it worked given all the packaging and zuul job changes i made to it | 22:13 |
| opendevreview | Tim Burke proposed opendev/irc-meetings master: Swift: Reinstate 0700 meeting slot https://review.opendev.org/c/opendev/irc-meetings/+/946834 | 22:14 |
| opendevreview | Tim Burke proposed opendev/irc-meetings master: Update check_chair.py for per-schedule chairs https://review.opendev.org/c/opendev/irc-meetings/+/946851 | 22:21 |
| opendevreview | Merged opendev/irc-meetings master: Swift: Reinstate 0700 meeting slot https://review.opendev.org/c/opendev/irc-meetings/+/946834 | 22:32 |
| opendevreview | Merged opendev/irc-meetings master: Update check_chair.py for per-schedule chairs https://review.opendev.org/c/opendev/irc-meetings/+/946851 | 22:46 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!