You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
File Explorer sends request to user bucket even when accessing group bucket #1007
When navigating to the File Explorer in Onyxia and clicking on a non-user bucket (such as a group bucket), Onyxia still sends a request to the user bucket in the background.
Steps to Reproduce:
Go to the File Explorer in the Onyxia interface.
Click on a group bucket or any bucket that is not the user's personal bucket.
Observe that a request is still made to the user's personal bucket, even though it was not selected.
Expected Behavior:
No request should be made to the user bucket if the user is interacting with a different bucket (e.g., a group bucket).
In fact when we initialize an S3 configuration , we are trying to create or get access to the current working directory.
I think this behaviour is a bit strange with the new behaviour of the explorer.. as we don't care anymore about the working directory
We should not trying to create any storage when we initialize S3 clients silently like that.
The working directory is a concept that need to be re-evaluated, it serves multiple purpose :
automatic creation of this bucket when we initialize a S3 connection
injection where helm chart ask for it ( automatic fill S3 path for differents purposes , mostly used for databases that want a path on S3 and where we try to provide a good path )
as a first step @ddecrulle can still do the automatic creation ( in order to have the same gobal UX regarding bucket creation ) but be resiliant to any error during this "optionnal / convenient " creation
Activity
fcomte commentedon Jul 11, 2025
In fact when we initialize an S3 configuration , we are trying to create or get access to the current working directory.
I think this behaviour is a bit strange with the new behaviour of the explorer.. as we don't care anymore about the working directory
We should not trying to create any storage when we initialize S3 clients silently like that.
fcomte commentedon Jul 16, 2025
The working directory is a concept that need to be re-evaluated, it serves multiple purpose :
as a first step @ddecrulle can still do the automatic creation ( in order to have the same gobal UX regarding bucket creation ) but be resiliant to any error during this "optionnal / convenient " creation
do not throw on unknown errors when creating S3 bucket