Hi,
A problem has been identified with WADI 2.0-M8 used by Geronimo 2.1
causing the failure of clustered applications when a specific node
(the node hosting the singleton partition rebalancing service) is
killed (a normal shutdown is fine). WADI 2.0-M9 addresses this
problem. This release is backward compatible with 2.0-M8 and hence it
simply needs to be installed into the geronimo repository to upgrade
from 2.0-M8 to 2.0-M9.
There are convenience download links for the 2.0-M9 artifacts on
WADI's home page: http://wadi.codehaus.org/
On a related notes, I will cut a 2.0 release for the next version of
Geronimo which adds the following features:
* State, i.e. HTTP sessions and SFSB instances, are paged on disc in
var/temp/SessionStore after a configurable period of time in memory; and
* Monitoring of global or Service Space Envelopes received and sent
per peer: http://docs.codehaus.org/display/WADI/6.+Monitor+Global+or
+Service+Space+Envelopes.
When I checked in some basic support for SFSB clustering, I added the
ability to register arbitrary distributed services. I believe this
can be quite handy to implement the optional distributed job
execution features of the JEE concurrency API under development
within the sandbox. I wrote a WIKI page describing what I mean by
distributed services: http://docs.codehaus.org/display/WADI/2.
+Distributed+Services
If people are interested by working on a Geronimo caching
implementation on top of the WADI's infra, then please ping me as
this is the next big feature I will be working on.
Thanks,
Gianny
- WADI 2.0-M9 - Replace buggy 2.0-M8 Gianny Damour
-
