Bug 1626140
Summary: | Custom roles deployment with controller pcmk + controller systemd roles deployment fails caused by Keystone project/users not being created | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | Red Hat OpenStack | Reporter: | Marius Cornea <mcornea> | ||||||
Component: | openstack-tripleo-heat-templates | Assignee: | Steven Hardy <shardy> | ||||||
Status: | CLOSED ERRATA | QA Contact: | Sasha Smolyak <ssmolyak> | ||||||
Severity: | urgent | Docs Contact: | |||||||
Priority: | medium | ||||||||
Version: | 13.0 (Queens) | CC: | agurenko, aschultz, asimonel, dbecker, dpeacock, emacchi, mburns, mcornea, michele, morazi, sbaker, shardy, tonyb, yprokule | ||||||
Target Milestone: | z3 | Keywords: | Triaged, ZStream | ||||||
Target Release: | 14.0 (Rocky) | ||||||||
Hardware: | Unspecified | ||||||||
OS: | Unspecified | ||||||||
Whiteboard: | |||||||||
Fixed In Version: | openstack-tripleo-heat-templates-9.2.1-0.20190119154856.fe11ade | Doc Type: | If docs needed, set a value | ||||||
Doc Text: | Story Points: | --- | |||||||
Clone Of: | Environment: | ||||||||
Last Closed: | 2019-07-02 20:08:26 UTC | Type: | Bug | ||||||
Regression: | --- | Mount Type: | --- | ||||||
Documentation: | --- | CRM: | |||||||
Verified Versions: | Category: | --- | |||||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||
Cloudforms Team: | --- | Target Upstream Version: | |||||||
Embargoed: | |||||||||
Attachments: |
|
Description
Marius Cornea
2018-09-06 16:11:13 UTC
Created attachment 1481356 [details]
roles and environment files
Created attachment 1481397 [details]
roles_data.yaml
Ok so update on this, there are two related bootstrapping problems we need to solve: 1. Bootstrapping performed via puppet uses a boostrap ID that's only unique per-role, e.g if you deploy something on more than one role it will run the boostrapping tasks more than once. This is resolved by the series I posted here: https://review.openstack.org/#/q/status:merged+topic:bug/1792613 2. Bootstrapping that happens via docker_puppet_tasks only ever happens on the primary controller, due to the ansible conditionals that check against bootstrap_nodeid To solve this we instead need to calculate the bootstrap node per *service* either the the deploy_steps_tasks ansible or inside docker-puppet.py The second part is still TODO, but if backported should resolve the specific case in the bug (I'm not sure if we will also need (1) to resolve the configuration Marius tested, but we'll definitely need (2)) Ok I pushed https://review.openstack.org/#/c/625078/ which aims to resolve the (2) problem from comment #9 - I think that is the root-cause of this bug and (1) is a related issue which may not be a blocker for this particular configuration. Currently need further testing/review of https://review.openstack.org/#/c/625078/ but if that works as expected we may be able to backport that and the tripleo-common patch https://review.openstack.org/#/c/605046/ idependent of the (fairly long) series of other patches which address (1) ref https://review.openstack.org/#/q/topic:bug/1792613+(status:open+OR+status:merged) Marius - I'll needinfo you, perhaps you can help me with this testing when you have some time, and we can figure out if the partial-backport approach is viable? We could backport the entire series but I'm a little wary as it involves a lot of changes and the problem where a service spans multiple roles is probably a less common case than the one reported in this bug? (In reply to Steven Hardy from comment #10) > Ok I pushed https://review.openstack.org/#/c/625078/ which aims to resolve > the (2) problem from comment #9 - I think that is the root-cause of this bug > and (1) is a related issue which may not be a blocker for this particular > configuration. > > Currently need further testing/review of > https://review.openstack.org/#/c/625078/ but if that works as expected we > may be able to backport that and the tripleo-common patch > https://review.openstack.org/#/c/605046/ idependent of the (fairly long) > series of other patches which address (1) ref > https://review.openstack.org/#/q/topic:bug/1792613+(status:open+OR+status: > merged) > > Marius - I'll needinfo you, perhaps you can help me with this testing when > you have some time, and we can figure out if the partial-backport approach > is viable? > Hey Steve! I tested the rocky backport(https://review.openstack.org/#/c/632714/) and it allowed the deployment to complete successfully. *** Bug 1668641 has been marked as a duplicate of this bug. *** Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2019:1672 |