Transaction Management


We've had another post about memory leaks just the other day, here:
Do you think what you encountered is a duplicate of that problem or is it something different?


Hello. I have been having memory leaks while using Janusgraph and found that there were a few places that I was not explicitly closing transactions and thought that might be the culprit. I am now closing all transactions explicitly but still find that I get out of memory errors. Someone suggested to me to go look at my cassandra database txlog table to see if there are any stray transactions there. I see 14 rows in that table. Before I go and mess with that table I would like to understand what it is for and what the rows inside it indicate. Are they always indicative of stale transactions?

I have done some local tests and did a whole bunch of queries where I closed the transactions and never saw anything pop up in this table. then I did a whole bunch of queries where I didn't close the transaction and one row appeared in this table. 

Can someone please explain what is going on here?