Red Hat Bugzilla – Bug 1573922
[UPGRADES] "overcloud upgrade converge" should update Heat stack to persist changes
Last modified: 2018-06-27 09:55:51 EDT
Description of problem: The OSP 13 upgrade related clis that is (openstack overcloud...) "upgrade" "update" and "ffwd-upgrade" are all currently *not* running a heat stack update. Instead they update the stored deployment plan in swift, especially to unset all OS::Heat::None resources set during the upgrade/update/ffwd-upgrade. However, the deployment code [1] isn't consulting the stored plan meaning subsequent stack updates (e.g. for scaling to add nodes) will still be broken as the 'nooped' resources will continue to be set to OS::Heat::None. [1] https://github.com/openstack/python-tripleoclient/blob/c09f50f1ad7f79dae5c67e5f047b3217c357ee74/tripleoclient/v1/overcloud_deploy.py#L369-L377 Version-Release number of selected component (if applicable): How reproducible: every time Steps to Reproduce: 1. run upgrade or ffwd-upgrade or update converge as the last step in the workflow: openstack overcloud ffwd-upgrade converge 2. Try to perform a heat stack update on the overcloud, e.g. to scale 3. OS::TripleO::DeploymentSteps and other resources will not be applied (they are still set to https://github.com/openstack/tripleo-heat-templates/blob/e24316c4ecb659d120a2d6288e8e1a992ce8e0ff/environments/lifecycle/upgrade-prepare.yaml#L4-L8) We can use one BZ to track the fix here more can be filed if necessary
adding the alternative proposal from jistr in trackers. The goal is to reach consensus today on which one and merge for end of week.
Just a heads up -- either of the patches is likely to break or timeout the gating update job. I posted 2 related patches, linked via depends-on on the patch. (No need to link them here as they are just for CI.) Given this and the fact that we'll need to backport to stable/queens too, i think landing it this week is not certain.
reassigning to jistr and removing the tripleo-common alternative
can you please clone this for update/upgrade/ffwd-upgrade as the qe team needs to test/track independently
Ok let's keep this one for upgrade, will clone for update + ffwd
Successfully scaled up extra compute after upgrade from RHOS-12 to RHOS-13 with python-tripleoclient-9.2.1-12.el7ost.noarch
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/RHEA-2018:2086