Comment 2 for bug 1733588

The warning log was introduced in [1] (back in 2013) and reading the commentary on that review it seems to have been intended to warn about transient version mismatch when proxys had not been upgraded to same version as object servers i.e. in [1] object servers started to expect X-Delete-At-Container to be sent by proxy rather than calculate the container name in the object server, but if an older proxy server does not send the header then fall back to calculating the container name locally and log the warning.

It's now unlikely that the proxy vs object server version mismatch exists, but the warning manifests as reported above when number of backend object requests is > number container replicas, which is likely to occur with many EC policies, so it is just an annoying warning that has no significance.

Possible solutions:
1. drop the warning altogether
2. make the warning conditional on the other delete-at headers such as X-Delete-At-Partition being present (which would be the case if the proxy was older). AFAICT the object server only expects X-Delete-At-Container when X-Delete-At-Partition and X-Delete-At-Host are also present.

I prefer solution #2

[1] https://review.openstack.org/#/c/31584/