Set of SSTables have the same set of ancestors

[Rajath Subramanyam <ra...@gmail.com>]

Hello Cassandra-users,

I logged into my test Cassandra cluster and saw this today. I find it very
unusual that a set of SSTables have the same set of ancestors:

$ sstablemetadata b800ks3-colla1-ka-*-Statistics.db | grep "Ancestors\|SSTable:"
SSTable: ./b800ks3-colla1-ka-1876
Ancestors: [1840, 1745, 1827, 1782, 1862, 1863, 1864, 1865, 1849, 1407]
SSTable: ./b800ks3-colla1-ka-2419
Ancestors: [2417, 2418, 2373, 2374, 2375, 2376, 2377, 2378]
SSTable: ./b800ks3-colla1-ka-2420
Ancestors: [2417, 2418, 2373, 2374, 2375, 2376, 2377, 2378]
SSTable: ./b800ks3-colla1-ka-2421
Ancestors: [2417, 2418, 2373, 2374, 2375, 2376, 2377, 2378]
SSTable: ./b800ks3-colla1-ka-2422
Ancestors: [2417, 2418, 2373, 2374, 2375, 2376, 2377, 2378]
SSTable: ./b800ks3-colla1-ka-2423
Ancestors: [2417, 2418, 2373, 2374, 2375, 2376, 2377, 2378]
SSTable: ./b800ks3-colla1-ka-2424
Ancestors: [2417, 2418, 2373, 2374, 2375, 2376, 2377, 2378]
SSTable: ./b800ks3-colla1-ka-2425
Ancestors: [2417, 2418, 2373, 2374, 2375, 2376, 2377, 2378]
SSTable: ./b800ks3-colla1-ka-2430
Ancestors: [2428, 2429]


What could have potentially caused this ? Nobody has run a sstablesplit on
any of these SSTables and I do not see any "anticompaction" words in the
log file.

Thank you !

Regards,
Rajath
------------------------
Rajath Subramanyam