Re: Backup failing with "failed to clone extents" error

[Andrei Borzenkov <arvidjaar@gmail.com>]

On 01.09.2021 20:52, Darrell Enns wrote:
>> Most likely. Did you simply make received subvolume read-write?
> 
> I believe so.

Well, here you are. There was proposed patch to clear received_uuid on
clearing read-only subvolume property, but it has never been applied.

> Is there a different way of doing it that would have
> avoided this situation? 

Yes. You should have created clone (read-write snapshot) of received
sibvolume and use it as your root. Received_uuid is automatically
cleared in writable snapshot.


> On a new subvol, would the received uuid
> normally be blank? 

Yes.

> Any suggestions on how to "fix" mine?
> 

I am not sure if it is even possible to clear received_uuid from user
space - the only available IOCTL is for setting it (there are extra
mandatory parameters and I do not know how to set them in this case).
You may try ./examples/set_received_uuid.py from
git://github.com/knorrie/python-btrfs while setting stransid and stime
to zero.

What will surely work is cloning you root subvolume and switching to
clone. If course you will need to restart all replication streams
beginning with full send of new root. But as it looks like you will need
to restart it anyway.