What message would be on the log, or not on the log, what would you be grep-ing for?
Also, why were you interested in startup procedure? Is there some "best practices: for that which could impact this? On Tue, Mar 25, 2014 at 2:55 PM, Robert Coli <rc...@eventbrite.com> wrote: > On Tue, Mar 25, 2014 at 1:55 PM, Batranut Bogdan <batra...@yahoo.com>wrote: > >> I believe that no matter how I stop cassandra, I should not be missing >> data. Even if compaction is in progress. >> > > Well, excepting that what you describe is impossible if you "crash" > cassandra (kill -9, etc.), I agree. It's possible that you are not actually > missing data, that the fragmentary SSTable you find "missing" is a > partially completed file from compaction? I suggest grepping in logs from > before the restart to determine if the SSTable you are finding missing > "should" actually exist. > > If you repeatably get this error, adding a test for actually missing data > should be pretty easy? > > =Rob > >