Release Notes NCache 3.6.2 (Service Pack 2)
Thursday, April 21, 2010
NCache 3.6.2 Service Pack 2 (SP2) contains important fixes and enhancements that have
been reported by customers. It is a recommended upgrade for all users of NCache. Those
using NCache 3.6.x version can upgrade without re-building/re-compiling the application.
Enhancements and New Additions
Following are some enhancements made in this release:
- Logging of Important events into event viewer:
An event will be logged in windows event viewer and cache-error log when:
- you stop or start a cache
- cache fails to start.
- cache node joins/leaves a cluster.
- state transfer start or stop.
- client connect or disconnect a cache.
- cache memory crosses a certain threshold specified in service config file.
- you are on last 10 days of NCache evaluation and a log warning event will
be logged per day until you extend the evaluation or activate NCache with a
These events are categorized as information, warning and information.
<add key="NCacheServer.EventLogLevel" value="error | warning| all"/>
User can specify only one of the following levels:
- error: Only the error events will be logged.
- warning: Both error and warning events will be logged.
- all: This level allows events of all categories to be logged. This is the default level.
Client Connection Log entry for event viewer/log-file can be enabled from NCache
Service configuration file. By default this option is disabled but you can enable it by
modifying the following attribute in NCache service config file.
<add key="NCacheServer.LogClientEvents" value="false" />
NOTE: The above change requires NCache service to restart.
Oracle 11g is supported:
NCache now supports Oracle 11g for database dependency and synchronization
List of Bugs Fixed:
- 736225 FIX: Some of NCache clients stop responding when the cache cluster is under
stress and you add a new caching server or the cache servers are using shared
network card on VM.
- 736224 FIX: NCache service (Alachisoft.NCache.Service.exe) and NCache Manager
stop responding when the cache is under stress and you add a new caching server or
the cache servers are using shared network card on VM.
- 736223 FIX: NCache service stops responding due to Hashmap corruption. This is
similar to the above mentioned service hanging scenario, however in this scenario,
service hangs for indefinite time.
- 736222 FIX: Cache stops expiring items from the cache if a WMI call does not respond
for a long period of time because of WMI unresponsive state.
- 736221 FIX: Client unbalancing in Partitioned and Partitioned Replica on node joining
is fixed now. This only happens in very rare situations.
- 736220 FIX: Clients were not able to get real performance benefits of clientoptimization feature of Partitioned Replica if the clients are 32-bit and cache is 64bit or
- 736219 FIX: Cache state transfer (starts when a cache node joins the cluster) does not
transfer those items which have multi-level key dependent items.
- 736218 FIX: Java client environmental issues have been fixed. Now you can pass
client.ncconf file path from API. Java client now supports both Windows and Linux
- 736217 FIX: Java sample application bugs are fixed now.
- 736216 FIX: NCache service does not start if the socket call fails to bind itself network
card. NCache now tries to establish the connection without binding.
- 736215 FIX: When you unplug the network cable Client hanging issue when cable is
unplugged during stress is fixed now. This normally happened in replicated topology
and the client operations were hanged for sometime.
- 736214 FIX: NCache tag query throws “Null Reference” in partitioned and partitioned
replica topologies when used with LIKE operator and the tag does not exist.
- 736213 FIX: NCache throws “Data group mismatch” exception when you update and
item by calling INSERT method with locking and group parameters.
- NCache Enterprise Edition 3.6.x
- NCache Professional Edition 3.6.x