This must possibly be a common occurrence but we noticed through hbck some regions (5 out of 20000) that were left unassigned and the master doing nothing about it. The thing to be noted is that all these regions seem to be daughter regions after a split (all of them had the regioninfo name in META table called splitB).
We had been very conservative about splits all this while but we had to turn them on due to region skews. (Still, the splits are far and few though: at most a handful a day) We use HBase 0.90.0 for now. The last few times this happened, we had restarted HBAse but it isn't a sustainable solution. Restarting the master perhaps, so as to trigger an assignment? V
