An automated configuration checker reports on common ENOVIA Live Collaboration
configuration problems that can cause severe performance degradation
or server shutdowns. The checker writes diagnostic information to an
mxAudit.log file. Users can also view diagnostic information via the
MQL Print Config command. See the MQL Guide
for details.
In a typical ENOVIA Live Collaboration environment, an ENOVIA Live Collaboration Server can consist of one or more ENOVIA Live Collaboration kernels, each of which acts like a server engine, encapsulating session level information and
providing database-independent access among other tasks.
The ENOVIA Live Collaboration Server can be set up for different modes of
operation:
- RMI mode--One or more ENOVIA Live Collaboration kernels (RMI servers)
runs in its own Java process, and the application server talks
to the server(s) via the RMI protocol. This is generally a more fault-tolerant
mode; the installation script will prompt for the number of ENOVIA Live Collaboration Server required (gateway setup).
- RIP mode--The ENOVIA Live Collaboration kernel runs in the same
process space as the application server and communication is made using
direct calls. This is faster than RMI mode since all code is within the
same process; however, because of this, RIP mode is not as fault-tolerant
as RMI mode.
- Gateway-- If there is more than one ENOVIA Live Collaboration Server,
an RMI gateway is used to distribute tasks to each server. The server(s)
do not need to be on the same machine as the application server because
the RMI protocol itself runs over TCP/IP, and therefore the ENOVIA Live Collaboration Servers can be running remotely.
For more information about these modes,
refer to Installing ENOVIA Live Collaboration Server.
When each kernel loads, its initialize function calls the configuration
checker, which evaluates the settings listed below. Topics or headings
in the following table have a link to the topic that describes it.
| Setting Checked |
Category |
Error Range |
| JVM Settings for Application
Server |
| -Xms |
Critical |
!=Xmx |
| -Xmx |
Critical |
<256M |
| -Xss |
Critical |
<384K |
| -XX:NewSize |
Warning |
>0.45*Xmx || <0.50*Xmx |
| -XX:MaxNewSize |
Warning |
!=NewSize |
| -XX:SurvivorRatio |
Warning |
!=2 |
| Other Java Settings |
| Obsolete
versions of eMatrixServletXXX.jar |
Critical |
|
| JVM
Mode |
Critical |
!=Server Mode |
| Java
Version |
Critical |
<1.3.1 |
| Key
ENOVIA Live Collaboration Environment Variables |
| MX_CONNECTION_POOL_SIZE |
Critical |
==1 |
| MX_CONNECTION_POOL_SIZE |
Warning |
<10 |
| MX_PROGRAM_POOL_SIZE |
Warning |
<10 |
| MX_BOS_EXPAND_LIMIT |
Warning |
not set |
| MX_BOS_FIND_LIMIT |
Warning |
not set |
| MX_MAX_THREAD |
Warning |
<200 |
| MX_TRANSACTION_TIMEOUT |
Warning |
not set |
| MX_ABORT_DANGLING_TRANSACTION |
Critical |
!=TRUE |
| MX_MEMORY_SYSTEM_LIMIT |
Critical |
not set |
| Trace
Settings |
Warning |
|
* On Windows platforms running non-RIP RMI, application server JVM settings are not audited. Use your application server console to check and set recommended JVM settings. -XX
settings apply to Sun's JVM only. These warnings should be ignored when using WebSphere's JVM. |
Each setting is evaluated by the configuration checker, which reports diagnostic information using the following categories:
- Informational--information-only messages that might be useful when tracking down performance issues and minor problems
- Warning--messages indicating that a setting could cause a problem and should be corrected
- Critical--messages alerting that a setting will cause a problem and must be changed
These messages can be viewed using the mxAudit.log file or the MQL Print Config command.