From us to you! We present 20 simple performance tips to optimize your EnterpriseOne installation. The settings that are default for a server may be great for a lab setting but typically are not sufficient for the real world. The business demands placed upon the software can put an extreme load on CPUs, RAM and bandwidth. If the server is not properly configured or properly sized, this will translate into poor performance ultimately disrupting your business.

1. Ensure all Ethernet adapters are at optimal settings.

Ensure all Ethernet adapters are set to the maximum settings, such as 100MB full duplex or 1000MB and ensure these settings are hard coded in the device properties. Do not allow these devices to auto negotiate (autosense) themselves as half duplex will occur.

2. Validating Network Adapters have current drivers to ensure optimal performance.

3. All servers should be connected to the same physical switch.

And all ports for these servers should be hard coded to match each server’s Ethernet adapter settings. For example, if a server is set to 100MB full duplex then the corresponding port would also be 100MB full duplex.

4. It is prudent to have a second server sizing done using the current user load to ensure that the servers have sufficient capacity.

5. Defragment your hard drives on a regular basis.

Defragment your hard drives on a regular basis off hours. Most operating systems have some way of reorganizing the data stored on its hard drives. For example, Windows comes with a manual disk defragmentation utility that can be run occasionally to increase data access on the drives. Depending on type of OS/DB and the amount of fragmentation, this one simple task can return huge dividends in speed.

6. Establish a routine for regenerating indices for all databases.

Priority should be placed on the “live” production database. This will increase the performance of applications as the end user does not have to wait as long for the data requested. This can significantly increase performance in specific areas such as pulling Sales Orders, etc. Also, identifying specific indices that can be added to the database tables can help cumbersome applications run more efficiently.

7. Ensure optimal use of SAN and/or RAID drives.

If using an SAN or Raid array be sure you have the best drive for your situation. For example, if you have a heavily used, large database you would not want to load it on three 100 gig drives that only spin at 10K RPM. You would see better performance if you loaded this database on several 36 gig drives that spin at 15K RPM. This allows for faster data access thus significantly improving performance.

8. Execute a database re-org.

Spreading your larger database files across multiple disks/controllers can have a significant increase in performance.

9. Make sure debug logging is turned off.

This may seem like an obvious item, but you would be surprised how many times a person is troubleshooting an item, turns logging on, resolves the issue and forgets to turn logging off. Active debug logging decreases performance significantly.

10. Ensure that you have your anti-virus properly configured.

UBE performance on the Enterprise Server will decrease due to anti-virus software. You will also experience performance degradation when building packages on the Deployment Server. Keep in mind that that anti-virus software scans all data that is transmitted and received over the network including data from the Enterprise and Deployment Servers, therefore performance degradation will occur in applications such as EnterpriseOne. Anti-virus software should be disabled during the package build process and also during the client installation. It will slow the process and in some cases can prevent a successful completion of the package or install.

11. Have a database audit performed.

Many times a database is installed optimized and after some time performance begins to suffer. There are items in every database that need to be checked. In addition, optimization jobs need to run occasionally. Make sure these jobs, such as “data integrity”, are running on a scheduler that is being monitored.

12. Have an EnterpriseOne audit performed.

Have a system audit performed at a minimum of once a year by an objective third party. The software is very complex and it constantly changing as a result of service packs, ESUs, planner updates, etc. It is impossible for anyone to be certain that the optimal settings are in place. Having an objective third party reviewing all of these settings periodically is a cost effective way to ensure that your company’s software is operating at its highest capacity.

13. Lock down Data Browser and “scroll to end” options for users.

Utilizing this functionality incorrectly puts system performance at risk. Data Browser and Scroll to End functionality allow users to view large amounts of data on a single click. This puts a major strain on the software and it should not be available to every user.

Scroll to end can be granted back to a user based on specific business cases that are shown not to jeopardize the system. Additionally, individual applications can be altered to pull a larger number of records at one time. Therefore, if it is common for a user to need 500+ records when they click find, this can be set in each application.

Furthermore, the default number of records pulled when a find is clicked is set in the JAS.ini using the GlobalPageSize. Data browser should never be used in a production environment. Users that require Data browser access should be given permission to access it through a non-production environment. They can then change the data source to point to production data.

Additionally, a separate environment and JVM can be setup that a user can access directly by going to a link such as http:\\server:88\jde\databrowser.

14. Web server and Enterprise server logs should be reviewed regularly.

Web server and enterprise server logs should be reviewed regularly to identify any current issues or trends that may lead to serious problems down the road. For example, some of the common errors you may see in the JAS logs may be related to “size mismatch” or “version not found” etc. In which case, perhaps a modification was made to an application and a new version needs to be created or recreated to resolve this issue. While some of these issues may not be directly related to performance issues, at the very least, it will assist in cleaning up the logs so performance issues can be easily identified.

15. Identify, resolve and clear any Call Object Kernels that are "Zombies".

A zombie is a kernel that is no longer responding. In theory, this should not happen as it indicates that the process on the enterprise server is no longer responding. Further troubleshooting is usually required via web server/enterprise server logs to identify the cause of the zombie.

16. Monitor the Event Viewer.

Monitor the Event Viewer regularly to identify any network/OS problems such as any network contingency issues that can have unpredictable results on the system.

17. Use Change Assistant to stay current on performance/memory ESUs.

Use Change Assistant to stay current with all performance/memory ESU installs to the system.

18. Ensure WebSphere’s HEAP settings are properly defined.

In the Websphere console, the best practice is to set the initial & max heap size memory setting to a smaller size, for example 64mb and let Java control the size. The initial 64MB recommendation comes from Denver Development. It has to do with the way garbage collection works and the smaller memory allows java to grow as needed which helps prevent the memory from getting fragmented.

19. Implement an appropriate security model.

Implement an appropriate security model (closed or open security) to manage the applications, actions, etc. that the users require.

20. Implement change management procedures/protocols.

Initially focus on the entire OMW object life cycle (DV, PY, PD, etc) then move to an over-all strategy to maintain a healthy system.

 

These are just a few simple tips we bring to the table. Our world class resources will not only increase the performance of your system but will partner with you to give you direction and insight to the latest technological advances.

Check out our other Tips & Tricks!