DO370 ch04s02: Registry storage still uses filesystem instead of S3

[Joseph5]

Step 4.4 has me patch the image registry config to use noobaa.

Step 4.5 has me output the image registry config and validate that it's set to S3.

Step 4.6 has me monitor the redeployment of the image-registry pod. However, a redeployment did not occur. The footnote states: if the redeployment did not occur, run oc edit and verify that spec.storage is correct.

I run `oc edit configs.imageregistry/cluster` and verify that spec.storage has the correct bucket name, region, and region endpoint (with https). spec.storage exactly matches the output of 4.5

I redeploy the pods using `oc rollout restart deploy/image-registry` . Then I run the check script in 4.7. However it outputs `REGISTRY_STORAGE fileystem` rather than the expected `REGISTRY_STORAGE s3`.

What am I missing?

[SANJURAJ]

Greetings Joseph. I have tried out this Guided exercise and found that it works perfectly. 

I am attaching the screenshot of the same.Step 4.1Step 4.2Step 4.3Step 4.4Step 4.5Step 4.6Step 4.7

You may recheck the entries in the file imageregistry-patch.yaml. It could be an indentation issue or incorrect bucket name.

Thanks & Regards,

Sanju Raj

[SANJURAJ]

Also the script shows REGISTRY_STORAGE as filesystem, it indicates that the image registry is still using the PVC storage. So you may rechceck the entries in the imageregistry-patch.yaml or the spec.storage section in the configs.imageregistry/cluster resource 

[Chetan_Tiwary_]

@Joseph5 Have you tried redoing the same in a freshly created lab ?

[Chetan_Tiwary_]

@Joseph5 working fine for me as well :

[Chetan_Tiwary_]

@Joseph5 please share the screenshots of the steps in case you face this issue still. 

or better raise a support case so that we can check in your lab.