This will happen if you end up with duplicate bucket ids in your index while splunk is running.
Some really weird problem with network filesystem might be responsible (do not use network filesystems for index and search unless you can guarantee their reliability). Bugs in a custom warmToCold script could be responsible. Pointing two splunk executables at one directory could be responsible.
It's pretty hard to make this happen.
But if it does happen, splunk may rename the warm bucket to this duplicate name.
Recommended steps:
- Determine if the 'duplicate' bucket is an exact duplicate of one in cold, in which case it can simply be deleted
- Determine if the 'duplicate' bucket is so old you don't care about the contents, in which case it can be deleted
- If neither of the above is true, it might make sense to manually rename the bucket to a new id that does not conflict with any of the ones in your index. Likely it will be old and will be moved to cold (and possibly archived) in short order.
