SST Inconsistency

[Daniel Doubleday <da...@gmx.net>]

Hi all

we are running c* 1.0.8 and found some strange row level tombstone problems.

Some rows (~50 in around 2B keys) have markedForDeleteAt timestamps in 
the future (so they 'drop' all writes) and 0 values as localDeletionTime.

A non-thorough check didn't bring up any code paths that could lead to a 
0 localDeletionTime and our client code is not capable to produce the 
markedForDeleteAt values we are seeing.
First we thought that it's file corruption. But we have multiple SSTs on 
different nodes that show the same pattern (though I did not check yet 
whether all belong to the same parition). Also all files are perfectly 
readable and the columns in the deleted rows are correct too.

I searched Jira but couldn't find anything related. So I was wondering 
if anyone is aware of a bug that could have led to this?

I'm going to patch scrub to fix the SSTs so it's not a big deal. Still 
strange ...

Cheers,
Daniel