What version of Dgraph are you using?
v1.1.0
Have you tried reproducing the issue with the latest release?
yes
What is the hardware spec (RAM, OS)?
n/a
Steps to reproduce the issue (command/config used to run Dgraph).
Set expired AWS credentials env variables (or none at all) on alpha server, then make
backup request overriding them as detailed here.
Expected behaviour and actual result.
The credentials provided in the backup request are expected to be used throughout.
Instead, if the current env credentials are stale dgraph attempts to get the latest
manifest in s3 using the old creds instead of the provided ones.
This can be seen in source here: https://github.com/dgraph-io/dgraph/blob/master/dgraph/cmd/alpha/admin_backup.go#L103
What version of Dgraph are you using?
v1.1.0
Have you tried reproducing the issue with the latest release?
yes
What is the hardware spec (RAM, OS)?
n/a
Steps to reproduce the issue (command/config used to run Dgraph).
Set expired AWS credentials env variables (or none at all) on alpha server, then make
backup request overriding them as detailed here.
Expected behaviour and actual result.
The credentials provided in the backup request are expected to be used throughout.
Instead, if the current env credentials are stale dgraph attempts to get the latest
manifest in s3 using the old creds instead of the provided ones.
This can be seen in source here: https://github.com/dgraph-io/dgraph/blob/master/dgraph/cmd/alpha/admin_backup.go#L103