Definitely open a bug report, that should never happen. However, can you provide the full panic output? That will tell us whether it was in blazer or restic. You seem convinced it is in blazer.
Also, how do you know it was hanging? (the first backup with 0.9.0 will take much longer because of the new archiver code… so it might not have been hanging, just taking a while)
I have now compiled the restic binary myself directly on my Debian ARM server (i.e. not cross-compiled from my Ubuntu amd64 desktop). I used the most recent master HEAD from github (i.e. not the v0.9.0 tag).
This seems to solve the problem. Therefore, there might be a problem with the Linux/ARM binary that is provided on the 0.9.0 release page.