Block storage: correct steps order #701
Conversation
rajathere
force-pushed
the
block-storage-correct-steps-order
branch
from
cfe2290
to
d95a5a3
Compare
|
Build failed (check pipeline). Post https://softwarefactory-project.io/zuul/t/rdoproject.org/buildset/0fd7d44d3da2451095ad70e511dd46f8 ❌ adoption-standalone-to-crc-ceph FAILURE in 2h 12m 18s |
rajathere
force-pushed
the
block-storage-correct-steps-order
branch
2 times, most recently
from
cb9171b
to
43d9b3b
Compare
|
Build failed (check pipeline). Post https://softwarefactory-project.io/zuul/t/rdoproject.org/buildset/38fef15ee8ed4f7497ece2ec21154f5b ❌ adoption-standalone-to-crc-ceph FAILURE in 52m 35s |
|
recheck unrelated error while deploying standalone |
|
Build failed (check pipeline). Post https://softwarefactory-project.io/zuul/t/rdoproject.org/buildset/6256563d3d834913b7395d64660c4e80 ❌ adoption-standalone-to-crc-ceph FAILURE in 2h 03m 21s |
rajathere
force-pushed
the
block-storage-correct-steps-order
branch
from
43d9b3b
to
b3b85d0
Compare
This is to support a followup patch where we want to clean up the old scheduler and backup services before deploying the cinder services.
To deploy block storage service i.e. cinder, we should first remove the service entries from old deployments to avoid any issues in the new RHOSO deployment. For that we need to rewrite the steps to: 1. Deploy cinder-api pod 2. List all scheduler and backup services from old deployment 3. Remove the old scheduler and backup services 4. Deploy scheduler, volume and backup services 5. Perform DB migration This will ensure that our RHOSO deployment is fresh and does not contain traces of errors from services of old deployment.
rajathere
force-pushed
the
block-storage-correct-steps-order
branch
from
b3b85d0
to
0201e42
Compare
|
Build failed (check pipeline). Post https://softwarefactory-project.io/zuul/t/rdoproject.org/buildset/f888d266e238491892ada48ac263a7e5 ❌ adoption-standalone-to-crc-ceph FAILURE in 1h 37m 50s |
| delay: 2 | ||
|
|
||
| - name: Get cinder-backup down services | ||
| ansible.builtin.shell: | |
There was a problem hiding this comment.
@rajathere {{ shell_header }} is missing here and at L34
i see the ci [1] failed with this "ERROR! failed at splitting arguments, either an unbalanced jinja2 block or quotes: {{ oc_header }}"
[1] https://softwarefactory-project.io/zuul/t/rdoproject.org/build/ef8a4e5832b34336b15d2c047d1f84a9
There was a problem hiding this comment.
That is intentionally removed.
If you see, shell_header provides the following command, set -euxo pipefail
The -o pipefail results in the whole command failing if any of the piped command fails, which we don't want here.
The comment on L#20 describes the scenario and we also do, set -eux avoiding the -o pipefail part
Strangely this playbook runs for me locally but failing here, will try to debug more
To deploy block storage service i.e. cinder, we should first remove
the service entries from old deployments to avoid any issues in
the new RHOSO deployment.
For that we need to rewrite the steps to:
This will ensure that our RHOSO deployment is fresh and does not
contain traces of errors from services of old deployment.