[Neo] User Digest, Vol 25, Issue 9

Ashwin Jayaprakash ashwin.jayaprakash at gmail.com
Tue Apr 14 20:22:59 CEST 2009


Hi Emil, out of curiosity - if the locks are acquired per Node in a Txn,
then how would you avoid deadlocks in a graph?

Ashwin.


On Tue, Apr 14, 2009 at 3:00 AM, <user-request at lists.neo4j.org> wrote:

>
> Hi Ashwin,
>
> There are no artificial limitations on the amount of concurrent
> read/write requests the Neo4j kernel can handle. In general, a read
> operation doesn't block other reads and write operation will acquire a
> write lock for every resource it mutates. That write lock will block
> other writing requests but allow concurrent read requests to read the
> old value.
>
> So the actual upper limit on number of concurrent operations will be
> set by the JVM / OS and how much contention your access patterns will
> lead to on your data set.
>
> Cheers,
>
> --
> Emil Eifr?m, CEO [emil at neotechnology.com]
> Neo Technology, www.neotechnology.com
> Cell: +46 733 462 271 | US: 206 403 8808
>
>
> ------------------------------
>
> _______________________________________________
> User mailing list
> User at lists.neo4j.org
> https://lists.neo4j.org/mailman/listinfo/user
>
>
> End of User Digest, Vol 25, Issue 9
> ***********************************
>


More information about the User mailing list