A different issue, but the way clean is implemented with shrinking bucket sizes and rolling, can this not cause a situation where the operation times out, but the cleaning has actually been performed? I.e OperationTimeout is raised, but the clean() was actually successful? What would timeout mean in this situation?
The reason I am asking is because I think that this might be the situation in my test. Since clean() throws the OperationTimeout but there are no events left when I manually check eventcount.
Secondly, a new sidenote: Isn't the timeout on 60 seconds that now has been put in place also a breaking change? It broke a little in helmut since cleaning sometimes takes longer than 60 seconds, and before Index.clean() ran indefinitely. What I am also curious about, is 60 seconds a good default value? I mean, our test logfile only has 10000 events and it is timing out sometimes, when cleaning out the index holding the data.
A different issue, but the way clean is implemented with shrinking bucket sizes and rolling, can this not cause a situation where the operation times out, but the cleaning has actually been performed? I.e OperationTimeout is raised, but the clean() was actually successful? What would timeout mean in this situation?
The reason I am asking is because I think that this might be the situation in my test. Since clean() throws the OperationTimeout but there are no events left when I manually check eventcount.
Secondly, a new sidenote: Isn't the timeout on 60 seconds that now has been put in place also a breaking change? It broke a little in helmut since cleaning sometimes takes longer than 60 seconds, and before Index.clean() ran indefinitely. What I am also curious about, is 60 seconds a good default value? I mean, our test logfile only has 10000 events and it is timing out sometimes, when cleaning out the index holding the data.