I believe this is very late to solve your issue if you are still experiencing it. However, I experienced this issue recently and got it to work again.
I simply restarted the CM and the fixup tasks resumed. I was removing an indexer from a cluster and it took over 13 hours. Hopefully someone else can benefit from seeing this. The Bucket status is held in the CM memory. When it restarts it will rebuild the list of the buckets and pick up where it left off.