We are trying to do a multithread graph traversal by using a connection pool. 
Currently, each thread has a new connection and we close each connection when 
it is done. If we do not specifically close the connection, the number of 
connections will reach to the maximum capacity in a very short time.
We are not quite understanding how graph objects and connections are being 
reused in the pool. Our guess is that we are not using the connection pool in a 
correct way. Killing the connection right away once the thread is done seems to 
solve the too-many-active-connection problem but it seems very inefficient and 
makes the pool seem meaningless. 
What we want to achieve is that we establish certain number of connections at 
the beginning in the pool. When a request comes, a connection is then assigned 
to the thread to do the graph traversal. If too many requests come at the same 
time, the requests that are not being able to be serviced can be placed in a 
queue to be further serviced.
I wonder if anybody can provide some help on how to do the multithread graph 
traversal correctly. 
Thanks much!
Cong Sun

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to