Uninstall playbook for Glusterfs #6918

SaravanaStorageNetwork · 2018-01-29T11:42:14Z

Uninstall playbook for Glusterfs

papr-bot · 2018-01-29T11:42:41Z

Can one of the admins verify this patch?
I understand the following commands:

bot, add author to whitelist
bot, test pull request
bot, test pull request once

SaravanaStorageNetwork · 2018-01-30T15:25:07Z

@jarrpa Please review

jarrpa

The desired solution is to just do a standalone uninstall playbook, not a conditional variable, so get rid of the openshift_storage_glusterfs_state variable.
With this implementation we don't need the wipe variable anymore. Remove that as a separate commit in this PR, and make sure to keep all of the functionality it enables.
There's no need to have | default after every use of glusterfs_name, that's already done.

SaravanaStorageNetwork · 2018-01-31T10:48:02Z

The desired solution is to just do a standalone uninstall playbook, not a conditional variable, so get rid of the openshift_storage_glusterfs_state variable.

So, you mean to call uninstall playbook directly like:
ansible-playbook -i /path/to/glusterfs_uninstall.yml
will address this.

With this implementation we don't need the wipe variable anymore. Remove that as a separate commit in this PR, and make sure to keep all of the functionality it enables.

Sure, duplicate code needs to be removed.
I will address as part of separate PR as it comes as part of install path.

There's no need to have | default after every use of glusterfs_name, that's already done.

Sure. will remove it.

jarrpa

A few things:

Rename the glusterfs_uninstall.yml files to just uninstall.yml.
Don't forget to update registry.yml to use the new facts file.

jarrpa · 2018-02-01T17:33:34Z

roles/openshift_storage_glusterfs/tasks/glusterfs_uninstall.yml

@@ -0,0 +1,96 @@
+---
+
+- include_tasks: glusterfs_config_facts.yml


- block: - include_tasks: glusterfs_config_facts.yml - include_tasks: glusterfs_uninstall.yml when: 'glusterfs' in groups and groups['glusterfs'] | length > 0

mjudeikis · 2018-02-01T19:31:37Z

roles/openshift_storage_glusterfs/tasks/glusterfs_uninstall.yml

+  - kind: "svc"
+    name: "heketi-storage-endpoints"
+  - kind: "secret"
+    name: "heketi-{{ glusterfs_name }}-topology-secret"


It would be good if we could have an optional section to wipe pv, vg, lv. So development would be easier when reusing same nodes... as now we would need to go and delete residual disk configuration.

@mjudeikis Sure - doing it..

jarrpa · 2018-02-02T13:30:41Z

roles/openshift_storage_glusterfs/tasks/uninstall.yml

+- block:
+  - include_tasks: glusterfs_config_facts.yml
+  - include_tasks: glusterfs_uninstall.yml
+    when: groups.glusterfs | length > 0


The group may not always be defined, so check that with 'glusterfs_registry' in groups and groups['glusterfs_registry'] | length > 0.

I have updated it - please check

jarrpa · 2018-02-02T13:31:13Z

roles/openshift_storage_glusterfs/tasks/glusterfs_uninstall.yml

+  - kind: daemonset
+    name: "glusterfs-{{ glusterfs_name }}"
+  - kind: storageclass
+    name: "glusterfs-{{ glusterfs_name }}"


We should also delete the glusterblock-provisioner pod. I believe that's part of a DeploymentConfig.

I dont' see glusterblock provisioner starting up as part of default CNS setup. Also, all the resources in glusterfs namespace gets deleted.

It should be deploying by default, because you have to specify --no-block to not deploy it. Regardless, since we COULD deploy it it should be covered by the uninstall.

Note that we'll need to cover the uninstall of gluster-s3, but I say we can do that later.

sure - will check about the block part

jarrpa · 2018-02-02T13:33:18Z

playbooks/openshift-glusterfs/private/uninstall.yml

+  - name: Run glusterfs uninstall role
+    include_role:
+      name: openshift_storage_glusterfs
+      tasks_from: uninstall.yml


You've changed the file name here but not the actual file's name.

Now, we have two files at roles/openshift_storage_glusterfs/tasks/ 1. uninstall.yml 2. glusterfs_uninstall.yml

uninstall.yml is the top level file there.

...OH RIGHT! I forget my own words sometimes, sorry. :)

Signed-off-by: Saravanakumar Arumugam <sarumuga@redhat.com>

jarrpa · 2018-02-02T15:37:07Z

playbooks/openshift-glusterfs/private/uninstall.yml

+  - name: Run glusterfs uninstall role
+    include_role:
+      name: openshift_storage_glusterfs
+      tasks_from: uninstall.yml


...OH RIGHT! I forget my own words sometimes, sorry. :)

jarrpa · 2018-02-07T14:58:28Z

roles/openshift_storage_glusterfs/tasks/glusterfs_uninstall.yml

+  - kind: daemonset
+    name: "glusterfs-{{ glusterfs_name }}"
+  - kind: storageclass
+    name: "glusterfs-{{ glusterfs_name }}"


It should be deploying by default, because you have to specify --no-block to not deploy it. Regardless, since we COULD deploy it it should be covered by the uninstall.

jarrpa · 2018-02-07T15:00:09Z

roles/openshift_storage_glusterfs/tasks/glusterfs_uninstall.yml

+  register: devices_info
+  delegate_to: "{{ item }}"
+  with_items: "{{ glusterfs_nodes | default([]) }}"
+  failed_when: False


You should use a variable to allow for specifying whether this gets done or not. I saw appropriate the wipe variable for this.

jarrpa · 2018-02-07T15:00:58Z

roles/openshift_storage_glusterfs/tasks/uninstall.yml

+- block:
+  - include_tasks: glusterfs_config_facts.yml
+  - include_tasks: glusterfs_uninstall.yml
+    when: "'glusterfs' in groups and groups['glusterfs'] | length > 0"


Move the indentation two spaces to the left. This applies it to the whole block, not the last task.

jarrpa · 2018-02-07T15:01:03Z

roles/openshift_storage_glusterfs/tasks/uninstall.yml

+- block:
+  - include_tasks: glusterfs_registry_facts.yml
+  - include_tasks: glusterfs_uninstall.yml
+    when: "'glusterfs_registry' in groups and groups['glusterfs_registry'] | length > 0"


Move the indentation two spaces to the left. This applies it to the whole block, not the last task.

jarrpa · 2018-02-07T15:05:50Z

roles/openshift_storage_glusterfs/tasks/glusterfs_uninstall.yml

+  - kind: daemonset
+    name: "glusterfs-{{ glusterfs_name }}"
+  - kind: storageclass
+    name: "glusterfs-{{ glusterfs_name }}"


Note that we'll need to cover the uninstall of gluster-s3, but I say we can do that later.

jarrpa · 2018-02-12T20:53:52Z

/ok-to-test

openshift-merge-robot · 2018-02-13T17:31:54Z

/test all [submit-queue is verifying that this PR is safe to merge]

openshift-merge-robot · 2018-02-13T19:00:56Z

Automatic merge from submit-queue.

openshift-ci-robot · 2018-02-13T19:42:50Z

@SaravanaStorageNetwork: The following test failed, say /retest to rerun them all:

Test name	Commit	Details	Rerun command
ci/openshift-jenkins/extended_conformance_install_crio	`2861f28`	link	`/test crio`

Full PR test history. Your PR dashboard. Please help us cut down on flakes by linking to an open issue when you hit one in your PR.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here.

This reimpliments openshift#6402 which was missed by openshift#6918. Signed-off-by: Jose A. Rivera <jarrpa@redhat.com>

Automatic merge from submit-queue. GlusterFS: Fix uninstall regression This reimpliments #6402 which was missed by #6918. Signed-off-by: Jose A. Rivera <jarrpa@redhat.com>

This reimpliments openshift#6402 which was missed by openshift#6918. Signed-off-by: Jose A. Rivera <jarrpa@redhat.com>

…ll_playbook Automatic merge from submit-queue. Uninstall playbook for Glusterfs Uninstall playbook for Glusterfs