Wednesday, 2026-07-29

*** zseguin is now known as Guest1435007:21
opendevreviewMauricio Harley proposed openstack/keystone master: Add transparent PBKDF2-SHA512 password rehashing on login  https://review.opendev.org/c/openstack/keystone/+/99289310:10
opendevreviewMauricio Harley proposed openstack/keystone master: Increase PBKDF2-SHA512 default iterations to 600,000  https://review.opendev.org/c/openstack/keystone/+/99105910:24
*** ralonsoh_ is now known as ralonsoh11:01
opendevreviewGrzegorz Grasza proposed openstack/keystone master: Use constant-time comparison for TOTP passcode validation  https://review.opendev.org/c/openstack/keystone/+/99908211:12
opendevreviewGrzegorz Grasza proposed openstack/keystone master: Sign SAML assertions with SHA-256 instead of SHA-1  https://review.opendev.org/c/openstack/keystone/+/99908311:12
opendevreviewGrzegorz Grasza proposed openstack/keystone master: Fix dropped validation constraints in role_assignments schema  https://review.opendev.org/c/openstack/keystone/+/99908411:15
opendevreviewGrzegorz Grasza proposed openstack/keystone master: Fix dropped validation constraints in role_assignments schema  https://review.opendev.org/c/openstack/keystone/+/99908411:23
d34dh0r53#startmeeting keystone15:01
opendevmeetMeeting started Wed Jul 29 15:01:36 2026 UTC and is due to finish in 60 minutes.  The chair is d34dh0r53. Information about MeetBot at http://wiki.debian.org/MeetBot.15:01
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.15:01
opendevmeetThe meeting name has been set to 'keystone'15:01
d34dh0r53Reminder: This meeting takes place under the OpenInfra Foundation Code of Conduct15:01
d34dh0r53#link https://openinfra.dev/legal/code-of-conduct15:01
d34dh0r53#topic roll call15:01
d34dh0r53admiyo, bbobrov, crisloma, d34dh0r53, dpar, dstanek, hrybacki, lbragstad, lwanderley, kmalloc, rodrigods, samueldmq, ruan_he, wxy, sonuk, vishakha, Ajay, rafaelwe, xek, gmann, zaitcev, reqa, dmendiza[m], dmendiza, mharley, jph, gtema, cardoe, deydra15:02
d34dh0r53o/ dmendiza 15:02
gtemao/15:02
moutazchaara[m]/o15:02
cardoeo/ (though I've not looked at keystone since the last time cause I've been neck deep in neutron)15:02
dmendiza[m]🙋‍♂️15:02
dmendiza[m]Appreciate the artisanal bespoke ping15:03
d34dh0r53hand crafted15:03
dmendiza[m]pasture raised, locally sourced15:03
d34dh0r53100% organic15:03
d34dh0r53lol15:03
dmendiza[m]😂15:04
d34dh0r53#topic review past meeting work items15:04
d34dh0r53#link https://meetings.opendev.org/meetings/keystone/2026/keystone.2026-07-22-15.01.html15:04
d34dh0r53I have an action item but I'm not sure if I'm going to get to it in a meaningful time.  I'm on bereavement leave next week, followed by PTO and then some more bereavement leave15:04
d34dh0r53dwilde plan mid-cycle keystone virtual meetup15:05
dmendiza[m]s/mid-cycle/late-cycle/15:05
d34dh0r53yeah, that's what I'm thinking15:06
d34dh0r53going to defer that till later in the cycle15:06
d34dh0r53next up15:06
d34dh0r53#topic liaison updates15:06
d34dh0r53nothing from me15:06
gtemaneither from me15:06
d34dh0r53cool15:07
d34dh0r53#topic specification Secure RBAC (dmendiza)15:07
d34dh0r53#link https://governance.openstack.org/tc/goals/selected/consistent-and-secure-rbac.html#z-release-timeline_15:07
d34dh0r532026.1 Release Timeline15:07
d34dh0r53Update oslo.policy in keystone to enforce_new_defaults=True15:07
d34dh0r53Update oslo.policy in keystone to enforce_scope=True15:08
d34dh0r53Fix config options in keystone-tempest-plugin https://review.opendev.org/c/openstack/keystone-tempest-plugin/+/93082915:08
d34dh0r53keystone-tempest-plugin has merged15:08
dmendiza[m]Nice15:08
d34dh0r53anything else on SRBAC dmendiza ?15:10
dmendiza[m]I don't think so? ... We can probably take it off the agenda now and revisit removing the deprecated policies at the PTG15:11
d34dh0r53cool, I'll archive that spec on the etherpad15:13
d34dh0r53#topic specification Secuirty Compliance Testing (dmendiza)15:13
d34dh0r53#link https://review.opendev.org/c/openstack/devstack/+/95796915:13
dmendiza[m]No updates ... I need to revisit this during an Upstream Friday15:14
d34dh0r53#topic specification User Specified Project/User UUIDs (dmendize, alee)15:14
d34dh0r53https://review.opendev.org/c/openstack/keystone-specs/+/99732015:14
d34dh0r53#link https://review.opendev.org/c/openstack/keystone-specs/+/99732015:14
dmendiza[m]Yeah, Andre (one of our teammates at RH) is going to keep working on these.  Not sure he's in here though?15:15
dmendiza[m]The Project/Domain spec needs to be updated to address gtema 's comments.  Not sure the other one for Users has any comments that need addressing yet15:16
gtemanot from me yet, had no chance to look into user one except seeing it is same huge long15:16
d34dh0r53cool, next up15:18
d34dh0r53#topic keystone-rs15:19
d34dh0r53#link https://github.com/openstack-experimental/keystone15:19
gtemaI am busy with polishing after it was finally deployed in our stage env.15:19
gtemaworks great, performance is awesome15:19
d34dh0r53nice15:19
gtemaneed to apply some fixes in area of missing journalctl support for logging, absence of sql logging, etc15:20
gtemaagain have seen that some internal caching is only having negative performance impact - so no caching15:20
gtemahowever understood, that for the case of side-by-side deployment with python keystone I would need to understand how python keystone caches data to be able to invalidate it remotely15:21
gtematoken validation with 0 caching takes 5ms while python without cached token 200ms, for warm cache 80ms, for hot cache 25ms15:22
gtemafare-well caching15:23
gtemaI am pretty done on features, now really focused on this ops items15:23
gtemathat's it for this week15:23
d34dh0r53that's impressive gtema , thank you15:24
d34dh0r53#topic open discussion15:24
moutazchaara[m]Hi, I have this one here https://review.opendev.org/c/openstack/keystone/+/99772415:25
moutazchaara[m]would be great if you gtema  Grzegorz Grasza have time to take a look at it15:25
gtemaI still remember it. On Rust side I have the same problem and also consider how to implement pagination for composite results (like assignments)15:26
moutazchaara[m]ok, yeah with the ldap the pagination breaks. not sure about the Rust one, i can take a look at it also 15:27
gtemaI mean it is the same issue conceptually that ldap does not support keyset pagination and we need to invent a solution for this15:28
moutazchaara[m]got it15:28
gtemafetching all and calculate the offset is the most "naive" approach, but depending on the number of entries it is so inefficient15:28
moutazchaara[m]yes it is a naive one, but not really sure if ldap support something works out of the box for that thing15:29
gtemaRust can keep the session to the ldap, so there is a chance of at least using the native ldap pagination inside15:29
moutazchaara[m]but in python it won't actually work with session that way ?15:30
gtemanot really - you never know on which node the request for next page lands so you can't reuse the connection cookie15:31
moutazchaara[m]yeah no guarantee 15:31
moutazchaara[m]i would suggest to go at least with the naive approach and then we can optimize it. As it is currently actually a bug for the ldap pagination. 15:31
gtemanot sure it is bug, I guess we are just returning all entries ignoring pagination15:32
moutazchaara[m]np, it is actually broken, you can't get page 2. 15:32
moutazchaara[m]the hash was calcuated incorrectly and passed there which ldap is not able to resolve it and lands always on 1st page15:33
gtemaI was sure we just return all entries unpaged to the client15:34
moutazchaara[m]if client doesn't has limit, then yeah15:34
moutazchaara[m]but if has limit. it won't make it to the client15:34
gtemamy naive fix would be to simply ignore the limit for ldap domain15:34
moutazchaara[m]yes list_limit = 0 and then it will be unlimited, but a pagination is ideal. 15:35
gtemaI know pagination is ideal, but if for the domain with 5000 users and page size 20-100 you fetch all users multiple times just to discard 90% it is such a terrible waste15:38
moutazchaara[m]yeah, in the patch if the user puts the limit to 5k, then he get's 5k and can paginate with it. 15:40
moutazchaara[m]but without the patch if he puts 5k he won't be able to paginate 15:40
gtemaI would definitely have no problem to simply discard the pagination for the LDAP as the first fix improving it further. Discarding so much data on Keystone is something I am not very comfortable with15:42
moutazchaara[m]yeah that's what we did as preliminary fix 15:43
moutazchaara[m]so would you suggest to invest more time into the sessions as there might be a solution for that? 15:44
gtemathere is pretty much definitely no solution to session in the current keystone architecture15:45
gtemakeystone-rs has at least way to address this15:45
gtemasince nodes are form a cluster15:45
gtemas/are//15:46
gtemalet's move on for the sake of the time15:46
moutazchaara[m]yp15:46
d34dh0r53good discussion, thanks both!15:47
d34dh0r53#topic bug review15:47
d34dh0r53#link https://bugs.launchpad.net/keystone/?orderby=-id&start=015:47
d34dh0r53no new keystone bugs15:47
d34dh0r53#link https://bugs.launchpad.net/python-keystoneclient/?orderby=-id&start=015:47
d34dh0r53python-keystoneclient is good15:48
d34dh0r53#link https://bugs.launchpad.net/keystoneauth/+bugs?orderby=-id&start=015:48
d34dh0r53nothing new in keystoneauth15:48
d34dh0r53#link https://bugs.launchpad.net/keystonemiddleware/+bugs?orderby=-id&start=015:48
d34dh0r53keystonemiddleware is good15:48
d34dh0r53#link https://bugs.launchpad.net/pycadf/+bugs?orderby=-id&start=015:48
d34dh0r53nothing new in pycadf15:48
d34dh0r53#link https://bugs.launchpad.net/ldappool/+bugs?orderby=-id&start=015:48
d34dh0r53and ldappool is good15:48
d34dh0r53#topic conclusion15:48
d34dh0r53That's all from me, thanks folks!15:48
d34dh0r53I'm out next week15:49
gtemathanks guys, cy15:49
dmendiza[m]Dave Wilde (d34dh0r53): I can chair next week15:49
d34dh0r53thanks dmendiza !!15:50
d34dh0r53#endmeeting15:50
opendevmeetMeeting ended Wed Jul 29 15:50:22 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)15:50
opendevmeetMinutes:        https://meetings.opendev.org/meetings/keystone/2026/keystone.2026-07-29-15.01.html15:50
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/keystone/2026/keystone.2026-07-29-15.01.txt15:50
opendevmeetLog:            https://meetings.opendev.org/meetings/keystone/2026/keystone.2026-07-29-15.01.log.html15:50
opendevreviewUdayendu Kar proposed openstack/keystone master: Fix updating unified limit resource_limit to zero  https://review.opendev.org/c/openstack/keystone/+/99917518:53

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!