Hope this issue can be resolved in manner that isn't toxic. I actually ran into this issue like this on a fairly large collection 2 months ago (almost 300GB, ha

Hope this issue can be resolved in manner that isn't toxic. I actually ran into this issue like this on a fairly large collection 2 months ago (almost 300GB, happened on a single replica set). It looked like an index got irreversibly corrupted. Luckily for us this collection got rebuilt everyday, so we could just delete the whole thing, but the pain was we weren't aware of the issue until a user reported he couldn't access data. When ever the corrupted data was accessed, it would lead to an assertion failure. Seemed like one of those dreadful moments when mongo decides to corrupt your data.