Container replicator reports warning about sharded ranges as error

[Mark Kirkwood]

I am in the process of upgrading from Swift 2.7 to 2.23. I am seeing a vast array of errors on any upgraded storage node (i.e 2.23) like:

May 25 19:56:17 obj4 container-replicator: /srv/node/con01/containers/769/585/c072f72bc62d5bcf8f6d3d914c815585/c072f72bc62d5bcf8f6d3d914c815585.db is unable to replicate shard ranges to peer 192.168.123.64:6001/con01; peer may need upgrading
...
May 25 19:56:17 obj4 container-replicator: 0 successes, 1 failures

One of these for every container! Some experimentation shows that in fact - containers *are* replicated (e.g deleting them on the 2.7 nodes and they reappear), but Swift is reporting this as an error (so recon does too) and that makes for a storm of monitoring 'errors' about container replication - which is not great, when in fact they have replicated fine.

Some nuance here would be nice.