[seam-dev] Re: Blocked threads

[Sanne Grinovero]

Hi all,
I'm sorry I didn't look at the sources, but if the problem is a
synchronized HashTable.get
did you consider this two very clean options:

A)
java.util.concurrent.ConcurrentHashMap
scales much better than a standard synchronized Hashmap

B)
using a standard HashTable but protect it
with java.util.concurrent.locks.ReadWriteLock instead of synchronized,
if you mostly use "get" instead of "put" you can use a "readlock" for each get,
readlocks permit many to be taken at same tame, so that no blocking occurs,
but acquiring a write lock blocks all other write and read ops (and has to wait
for all previous to be released).

I hope to be helpful, but must say I don't know the internals of Seam well.

Sanne

2008/6/5 Christian Bauer <christian.bauer at gmail.com>:
> The option to turn the ManagedEntityIdentityInterceptor off has been closed
> with "we need it for passivation":
>
> http://jira.jboss.com/jira/browse/JBSEAM-468
>
> I'm not sure we need it for a system with a) only javabean components and b)
> in-memory/non-persistent http sessions and c) no clustering/replication of
> http sessions. Which well, would be really a common scenario. :)
>
> We definitely can't have each thread run into a synchronized (readers block
> other readers) HashTable.get() thousands of times during a regular request.
> _______________________________________________
> seam-dev mailing list
> seam-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/seam-dev
>