Tuesday, 2025-10-07

opendevreviewOpenStack Proposal Bot proposed openstack/project-config master: Normalize projects.yaml  https://review.opendev.org/c/openstack/project-config/+/96255702:25
*** mrunge_ is now known as mrunge06:13
ildikovHi, I'm reaching out with a quick question. The StarlingX vulnerability team received a report regarding https://docs.starlingx.io/backup/ and points out that the directory listing shouldn't be available on that url.14:46
ildikovAs the StarlingX docs site is on the opendev infrastructure, I wanted to check if y'all have run into this before? And might ideas how to resolve it?14:48
clarkbildikov: that happens because there is no index.html in that directory. That seems to be intentional based on the links on this page: https://docs.starlingx.io/#backup-and-restore14:49
clarkbthis isn't an opendev issue, its a content issue. We're serving what is present in the content14:49
ildikovthe vulnerability report Staes that in this case the API response should be 40314:49
clarkbif index-backup-1b466179efc3.html is renamed to index.html or an index.html is added you'll get the content of that index.html file14:50
clarkbildikov: the server is configured to generate an index file autoamtically if one isn't present. This is intentional as it supports use cases like the tarballs hosting. We can theoretically turn that off for starlingx docs in particular since that should be its own vhost, but I think it would be good to undersatnd why fixing it in the content is not possible first14:51
clarkbits also not really a vulnerability as is14:52
clarkbit could potentially expose unwanted data if you push it there, but considered all of the docs are generated from public git repos already the risk is low to nil14:52
ildikovI understand14:53
ildikovthank you for clarifying14:54
fungiildikov: yes, if there is no secret data in that directory, then there is absoutely no vulnerability in listing the contents of it. reports like that are noise and i just send them directly to the trash without even wasting my time responding14:56
fungithere is too little time in the day for dealing with actual vulnerabilities to spend it on that sort of nonsense14:57
clarkblooking at the source for the starlingx docs I think what happened is they stopped updating those particular docs so those are a point in time snapshot which is where the hash suffix comes from14:58
clarkbanother option there is to have a minimal stub for those docs using proper indexes14:58
clarkbindex.html that says "these docs are deprecated and not updated, see historical version here: linkwithsuffic"14:59
*** NeilHanlon_ is now known as NeilHanlon15:06
fungiΠthon is released! https://discuss.python.org/t/python-3-14-0-final-is-here/10421015:32
clarkbTIL upper case π is far less recognizable15:35
fungii contemplated saying πthon instead15:36
clarkbI thought it was a ∩ at first15:44
fungimy terminal fonts make both look too much like latin "n"15:46
*** dhill is now known as Guest2856316:19
*** dhill is now known as Guest2856917:35
clarkbI've deleted old nodepool images in rax flex and ovh. Ovh in particular was holding onto some old images. Still need to do rax classic, vexxhost, openmetal and osuosl17:44
corvusthanks!17:44
tonybI'm going to be a couple of minutes late for the pre-PTG start.17:55
clarkback, its early for you no rush17:57
fungiwe'll keep your seat warm17:57
fungiand coffee hot17:57
clarkbwe've fully entered the time of year where I'm cold until about 2pm and then its too warm. Weather transition problems17:59
fungii'm already in the room, assuming it's the one that matches the planning pad17:59
fungiwhich the pad says it should be17:59
clarkbfungi: I'm there to and don't see you18:00
clarkbhttps://meetpad.opendev.org/p/opendev-preptg-october-202518:00
fungihttps://meetpad.opendev.org/opendev-preptg-october-2025 is what it says i'm in18:01
clarkboh ha18:01
funginow corvus is in mine too18:01
clarkbI copied the etherpad link and left the /p/18:01
clarkbthat will learn me18:01
fungioh, /p strikes again, yep18:01
clarkbthough I guess that works just not the way we want18:01
fungiyeah, low hanging fruit: blanket redirect from https://meetpad.opendev.org/p/(.*) to https://meetpad.opendev.org/$118:04
clarkbok first block done. And now lunch20:05
clarkband then a reminder that I have to pop out in about 1.5 hours for a dentist visit20:05
clarkbI started on the osuosl image cleanups. noble and jammy cleared out ok. older focal and bionic images hit 500 errors21:10
clarkbI'm going to have to pause there, but it seems likely we'll need to ask a cloud admin to cleanup whatever is left when I'm done21:10
fungimakes sense, thanks for working on that!21:11
clarkbok I cleaned up the images in osuosl that I could delete https://paste.opendev.org/show/bC0wmj2PI0SXHcuVwC3P/ are the images I could not delete23:13
clarkbRamereth[m]: absolutely no rush at all on this but https://paste.opendev.org/show/bC0wmj2PI0SXHcuVwC3P/ containts a list of image uploads in the osuosl arm64 openstack cloud that can be deleted but I get http 500 errors when attempting to do so as our normal user account23:14
clarkbI'm working on openmetal next23:14
clarkbopenmetal is done except one jammy image wouldn't delete. Glance claims it is in use, but I'm not sure how as there are no nodepool nodes listed anymore23:18
clarkbI'll pause here. Rax classic and vexxhost are the remaining clouds that need the nodepool image cleanups23:19
clarkbalso its looking like tomorrow after my morning meeting block (pre ptg included) I'll be popping out to help my brother move back into the house I just helped him move out of (the joys of home remodelling and moving twice in 4 months)23:23
clarkbour new ubuntu noble image uploads for builds created after I deleted the image with corrupted git repos have completed. We should check on the daily image builds that run in a couple of hours /me makes a note to check that tomorrow23:26
clarkbMostly want to make sure there isn't some consistent problem that reproduces this issue as our assumption for right now is some fluke in uploads/conversion23:26
tonybcorvus: FWIW, I mostly understand why https://review.opendev.org/c/openstack/openstacksdk/+/931575 fails (at least for the test_create_object case).  I need to learn a little more about swift and openstacksdk to fix it.23:58

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!