[06:50:19] I'm looking for someone who could help me package a new version of python-pynetbox for bullseye - https://gerrit.wikimedia.org/g/operations/debs/pynetbox [07:26:37] XioNoX: how does that relate to the versions in Debian? Are you wanting a later Debian-packaged-version backported, or something more complicated? [07:26:57] (I see bookworm and later have 7.0.0-1) [07:27:36] Emperor: I need https://github.com/netbox-community/pynetbox/releases/tag/v7.4.0 on the cumin hosts, which are running bullseye [07:27:59] I don't think it's complicated but I don't really know the steps [07:37:24] I think, if I were doing it, I'd start a fresh package following https://wikitech.wikimedia.org/wiki/Debian_packaging/Tutorial but maybe pinch the existing dh/{control,changelog,rules,compat} from the debian branch in operations/debs/pynetbox . That might not be what v.olans would do, though. [07:37:52] (if you'd like me to have a go at that and let you look at the result, I can probably do so in the next couple of days) [07:38:53] [which would be a slightly different approach to that was taken back in 2021 when the 5.3.0-2 package was done, but we didn't have the gitlab ci stuff available then] [07:46:46] Emperor: yeah 2021 is a long time ago ! I'll read the doc and see if it seems doable for me and will ping you either way [07:51:34] 👍 [09:43:46] hi all i finnaly recived my leaving preset https://photos.app.goo.gl/wvnd66ubeH7QidPM8 i love it <3 thanks all its a great gift. The artist selcetion was specifically impressive, it an artist that brings a lot of shared memories for me and my wife so well done to whoever eeked that bit of infomraiton out of me and rembered. Thanks all :) [09:46:11] <3 [09:55:02] jbond: <3!!!! [10:01:19] Emperor: https://gitlab.wikimedia.org/repos/sre/pynetbox/-/jobs/348632/artifacts/browse/WMF_BUILD_DIR/ :) I followed https://wikitech.wikimedia.org/wiki/Debian_packaging_with_dgit_and_CI [10:01:37] Emperor: not sure about the Bookworm vs. bullseye though, will the generated package works on bullseye? [10:04:30] jbond: <3 [10:06:52] XioNoX: the .changes file says Distribution: bullseye-wikimedia so that should be good. [10:07:26] Emperor: nice, that's so much easier than what I remembered :) [10:07:44] XioNoX: \o/ [10:08:20] I'll try the package manually on cumin2002 after lunch, then send it to reprepo if all good [10:12:37] jbond: <3 <3 [12:33:38] jbond: <3 <3 [12:46:48] jbond: <3 [12:52:21] arnoldokoth: you have a pending puppet diff. Is that safe to `puppet-merge`? [12:52:38] Yes please. [12:52:41] alright, on it! [12:53:22] Thank you. [12:55:20] jbond: awesome pic thanks for sharing :) [12:55:26] hope you are keeping well [12:56:37] thanks for all the love everyone right back at yu <3 and yes im doing very well thanks, finnaly progressing with sme of the house \o/ :) [13:13:36] jbond: <3 [13:14:42] jbond: :wave: [13:27:10] Folks just a heads up myself and sukhe are going to do a planned failover of lvs2013 traffic to lvs2014 shortly [13:27:35] this is to validate lvs2014 is working as expected following some topology changes yesterday [13:27:43] 👍 [13:27:44] shouldn't be any affect on anything but just to be aware [13:27:54] I need some assistance with docker-pkg update. I can't seem to make it see both _only_ the image that I am updating (kserve/build) and its dependency (golang1.22) [13:28:23] `docker-pkg -c config.yaml update build --version 0.13.0 --reason '* Version bump to v0.13' ./images/kserve/` results in "RuntimeError: Image golang1.22 (dependency of docker-registry.wikimedia.org/kserve-build:0.11.2-2-20240818) not found" [13:28:49] whereas dropping kserve from the second parameter and adding it to the build one results in just nothing being done [13:31:28] also note the wrong version in the first error (0.11.2-2-20240818) [13:51:36] ah, this did the trick: [13:51:44] docker-pkg -c config.yaml update --version 0.13.0 --reason '* Version bump to v0.13' kserve\*build ./images [13:52:20] (kserve-build also works) [14:25:16] cdanis: auto-telemetry worked beautifully! however... it's about 80% `fs.statSync` by volume which feels both useless and spammy. i'm gonna keep dumpster-diving the docs, but, do you happen to know offhand if it's possible to just filter that out? or is it fine that it hits our jaeger instance? [14:26:13] zip: in production we can filter spans before ingestion time. buuuut I think you should be able to set an env var [14:26:40] maybe try OTEL_NODE_DISABLED_INSTRUMENTATIONS="fs" [14:27:40] https://opentelemetry.io/docs/zero-code/js/configuration/ [14:53:24] aha, thanks [17:36:46] brett: just looking at your email, Safari 11-14 and the Android stock browser seem to still have Grade A support [17:37:32] I wonder particularly how many Safari users would be affected [17:43:42] My napkin maths for Miraheze (which users insntant commons so would impact) is about 1.1% of our iOS users [19:31:17] Nice, thanks! [20:30:45] brett: not really, the fact it's Grade A means MediaWiki still fully supports so you should probably look at usage for safari versions [20:31:06] Might be worth seeing what plans are to increase Grade A minimum version [20:32:26] I meant thanks for the info [20:33:29] Ah [20:54:27] RhinosF1: It's worth noting that the TLS 1.2 deprecation is a long, long ways away. [21:02:46] brett: that gives a fair amount of time to factor in when grade A support from mediawiki for those browsers is dropped too