Created attachment 1587018 [details] ansible log file where it is failing for brick capacity of BHV. Description of problem: As of today in upgrade playbook we are just checking the capacity of bricks. We are not actually verifying whether bricks are related to block hosting volume or not. Because it is not necessary that if BHV brick is full block volume will be also full. So from preventing upgrade failure, We should verify above as well. Version-Release number of selected component (if applicable): openshift-ansible-3.11.123-1.git.0.db681ba.el7.noarch How reproducible: Steps to Reproduce: 1. Upgrade glusterfs cluster with upgrade playbook with 98% used block hosting volume. 2. 3. Actual results: 2019-07-02 20:13:41,017 p=125669 u=root | fatal: [dhcp46-218.lab.eng.blr.redhat.com -> dhcp46-218.lab.eng.blr.redhat.com]: FAILED! => {"attempts": 120, "changed": false, "msg": "bricks near capacity found: {u'dhcp46-19.lab.eng.blr.redhat.com': [u'/var/lib/heketi/mounts/vg_966da3e87480fa95ddacd2ad91f49518/brick_7d2df4baf0b1a547ed6f38f57aff7969'], u'dhcp47-107.lab.eng.blr. redhat.com': [u'/var/lib/heketi/mounts/vg_ab9409e4645b20dd76bf550cab8ef1eb/brick_0cf9feb79abf33fc48eedcab4b3d4c54'], u'dhcp46-35.lab.eng.blr.redhat.com': [u'/var/lib/heketi/mounts/ vg_4c95c52d5ee3d5d5d685c57aeaf5ac04/brick_2571c4a94a15468a37952641e5176da2']}", "state": "unknown"} Expected results: It should exclude bricks which are part of BHV. Additional info:
I see this as an RFE and not a technical requirement for the upgrade playbook. Marking it as such.
Hello Arun, Can you please set the right test coverage flag here ? Thanks kasturi