This message was deleted.
# epinioa
adamant-kite-43734
10/10/2024, 5:48 AMThis message was deleted.
b
broad-dream-81849
10/10/2024, 6:28 AMHi @rapid-optician-64537, I guess the choice was made to cache the files, and reuse them during a rebuild
r
rapid-optician-64537
10/10/2024, 6:34 AMThank you for answer,
Are cache files actually being reused?
During my testing, it seemed like a new binary format bucket name was generated every time.
rapid-optician-64537
10/10/2024, 6:36 AMThere is no step in the service flow that removes the files, and it also doesn't seem like they are being reused in a cache format. I'm not sure if this is how it's supposed to work.
b
broad-dream-81849
10/10/2024, 6:38 AMIf you do a rebuild (not a push) they should be reused. The epinio server will get the same bucket pushed for that build and reuse it
broad-dream-81849
10/10/2024, 6:39 AMI don't remember if there is something to set a number of builds to keep, or if there is some configuration on MinIO that can be used to avoid this issue
r
rapid-optician-64537
10/10/2024, 6:44 AMThank you for the response.
I believe the purpose of using Epinio is to create an OCI (Open Container Image) after the build.
If the OCI is stored in a Docker Registry, wouldn't it be better for integrity to reuse the OCI from the registry?
Wouldn't it be more effective to implement a feature that clears the information in the bucket after the build?
b
broad-dream-81849
10/10/2024, 7:13 AMI was checking the docs and I found this
When you successfully push a new version of your application, Epinio removes the resources of the earlier staging process from the Kubernetes cluster. It also deletes the earlier version of the sources from S3. This way, Epinio doesn't store more than it needs on the S3 storage and no manual cleanup is necessary.https://docs.epinio.io/howtos/customization/setup_external_s3
broad-dream-81849
10/10/2024, 7:13 AMso maybe the behavior you are experiencing is not expected
✅ 1
r
rapid-optician-64537
10/10/2024, 7:25 AMThank you!!
rapid-optician-64537
10/10/2024, 7:55 AMI just checked the source code, and when the stage API is called, the unstage API is not triggered, which means the files are not deleted from S3. Only when the deploy API is called, the unstage process is designed to remove all the existing resources. I believe that in the case of the stage API as well, all previous resources should be removed once the image is pushed. It seems like additional handling is required.
Thank you very much for your response!!
b
broad-dream-81849
10/10/2024, 8:15 AMyou are welcome, and feel free to write an issue to discuss about this if you want 🙂
👍 1
3 Views