data size reported by 2i is very high even though only one snapshot is open according to 2i stats.

Open snapshots is 1(i.e snapshot_create is 8952 and snapshot_close is 8951) and items_count is 20M. From the raw files the data sizes are 12 GB and 9GB. The file with 9GB data is the file being created by compaction run. At this point, 2i statistic report that data is 115GB in size and raw disk size is 200+GB. From this, it appears that 2i is holding open handles to unnecessary snapshots.

"bucket-1:myindex:data_size”:123511019463 (115GB)
"bucket-1:myindex:disk_size”:223842385920
"bucket-1:myindex:timings/storage_snapshot_close":"8951 195711085476 3844338222836990240”,
"bucket-1:myindex:timings/storage_snapshot_create":"8952 1246994716383 3949738262081300211”
"bucket-1:myindex:items_count”:20000001

"bucket-1:myindex:num_docs_indexed":80297320

logs:
https://s3.amazonaws.com/cb-customers/performance/15785/collectinfo-2015-07-27T035451-ns_1%40172.23.100.16.zip