Comment 4 for bug 1264029

@Zhiyan: I was thinking something like the cloud admin/operator was one guy/company but the storage operator was another. But I admit, that was more a theoretical scenario - I have no proof or figures of how (un)common it is or would be in real deployments.

I think the middle road is good - we log those errors for the sake of debugging and diagnostics, but don't necessarily expose them to the end user.

And I agree - even though we log them, we must be careful not to expose sensitive information (even in the logs).

I'll upload the patchset on Monday when I get back to work. I hope you'll help out in the reviews :).