WIP: Basic tests to verify setup and Gluster volume#73
Conversation
Two tests are available: - Setup test: Tests expected number of pods and Running status - CSI and Gluster volume test: Creates 1GiB gluster volume using Persistent Volume claim and runs nginx pod to verify mount and persisted data. TODO: - Dependencies need to be installed(`pytest` and `kubernetes`) - Sleeps in tests needs to be removed - Internal modules needs to be refactored - Integrate Pytest with tox and nose similar to project https://github.com/gluster/libgfapi-python - Deployment of these test scripts to master node - Tests running infra creation How I tested: - Copied `/root/.kube/config` from one of the node(kube1) to my local machine(`/home/aravinda/.kube/config`) - Run tests from local machine ``` cd gcs pytest e2e ``` Signed-off-by: Aravinda VK <avishwan@redhat.com>
aravindavk
requested review from
JohnStrunk,
Madhu-1,
amarts,
atinmu,
kshlm,
mykaul,
sabose and
sac
ghost
assigned
| cnt = 0 | ||
| expected = 3 | ||
| for pod in pods: | ||
| if pod.metadata.namespace == "gcs" and pod.metadata.name.startswith("gluster-kube"): |
| op_cnt_expected = 1 | ||
| for pod in pods: | ||
| if pod.metadata.namespace == "gcs": | ||
| if pod.metadata.name.startswith("etcd-operator"): |
| op_cnt += 1 | ||
| continue | ||
|
|
||
| if pod.metadata.name.startswith("etcd-"): |
| nodeplugins_cnt += 1 | ||
|
|
||
| if pod.metadata.name.startswith("csi-provisioner-glusterfsplugin"): | ||
| provisioner_cnt += 1 |
| return client.CoreV1Api() | ||
|
|
||
|
|
||
| def gluster_pods_verify(pods): |
There was a problem hiding this comment.
expected pod count needs to be the input
There was a problem hiding this comment.
we can have a generic function which takes namespace pod name and expected count as input.
| @@ -0,0 +1,19 @@ | |||
| --- | |||
There was a problem hiding this comment.
cant we make use of existing examples
There was a problem hiding this comment.
from where? Sorry I didn't understand this comment
There was a problem hiding this comment.
currently we have an example templates in https://github.com/gluster/gcs/blob/master/deploy/examples can we make use of it?
There was a problem hiding this comment.
got it. Let me check if I can use it. In that volume claim and pod creation is added in single yaml file.
- Also, application pods will change when we introduce more tests, we can reuse volume claims examples.
- Once we start running application tests inside pod then those examples will look complicated for users.
There was a problem hiding this comment.
We need to carefully consider the trade-off:
- Using the examples makes sure they keep working
- It creates interdependencies that may be non-obvious & potentially complicate the examples.
- I'm not sure the test pods, whatever they end up being, will be particularly useful as "examples".
Perhaps a compromise would be to have a hybrid approach where we run the examples, ensuring they continue to work, but put the bulk of the custom test pods w/ the e2e?
| @@ -0,0 +1,13 @@ | |||
| --- | |||
There was a problem hiding this comment.
same here, if it's needed to have a separate pod and PVC template we can separate it out in the example folder.
| metadata: | ||
| name: gv1 | ||
| annotations: | ||
| volume.beta.kubernetes.io/storage-class: glusterfs-csi |
There was a problem hiding this comment.
Should be provided as spec.storageClassName: glusterfs-csi
| # FIXME: Remove sleep | ||
| time.sleep(100) | ||
|
|
||
| # Read the data again to see if it is persisted |
| @@ -0,0 +1,19 @@ | |||
| --- | |||
There was a problem hiding this comment.
We need to carefully consider the trade-off:
- Using the examples makes sure they keep working
- It creates interdependencies that may be non-obvious & potentially complicate the examples.
- I'm not sure the test pods, whatever they end up being, will be particularly useful as "examples".
Perhaps a compromise would be to have a hybrid approach where we run the examples, ensuring they continue to work, but put the bulk of the custom test pods w/ the e2e?
4 participants
Two tests are available:
Persistent Volume claim and runs nginx pod to verify mount and
persisted data.
TODO:
pytestandkubernetes)https://github.com/gluster/libgfapi-python
How I tested:
/root/.kube/configfrom one of the node(kube1) to my localmachine(
/home/aravinda/.kube/config)Signed-off-by: Aravinda VK avishwan@redhat.com