Comment 10 for bug 1453599

Andy are you building from the Kilo branch/tag?

If so then the Christian's fix wont be there as it was deployed to master. It will be in the next swift release (which will hopefully be not too far away as swift releases more regularly then the rest of OpenStack).

Having said that, the bug should only exist if the CLI version doesn't match the recon middleware in the storage servers (unless your running the CLI that Christian has patched).

The part of swift that pulls the quarantined data is the recon middleware which should be in the pipelines on the storage servers (account, container, object). Becuase it's a middleware it runs with the servers. This code, in kilo, will default storage policies to {}, meaning so long as your CLI and servers are all running the same code, they should work. If you have rolled out the new code to the entire environment, then the CLI and recon middleware should match.. if you have rolled out and still erroring, then the server in question hasn't been restarted (as it is using the old recon middleware in memory) or missed the upgrade.

That's my 2 cents, how ever if you have found a bug, awesome! I will go fix it.