Five things that can improve your TADDM experience.

Five things that can improve your TADDM experience

1. Sensor timeouts
When checking discovery results, you can encounter sensors timing out on some target systems. A default sensor timeout is set to 10 minutes, but for some cases this is not enough time. The rule of thumb is: if a sensor times out you should increase timeout settings for the sensor. There are various reasons why some sensors run longer than others. For Example There might be a slow network connection or a huge amount of data to discover on a particular endpoint. The second example can be observed where you have large network devices with a lot of ports and connections. Increasing timeout can increase discovery length for sensors that were failing before and makes it possible for the sensor to finish a discovery and store discovered data.

2. Target systems’ configuration
The other thing you might see in your environment is sensors that end with warnings due to a target system configuration. Do not ignore these warnings unless you are sure what they mean. Some warnings show up when there are some prerequisites missing. Others may pop up because of a Simple Network Management Protocol (SNMP) timeout. In general, you can live with these warnings, but you might not be getting all the information that is available on the target system if you don’t take care of them. Taking care of these warnings is another way that you can improve your discovery depth and ensure that you discover all you can.

3. Database maintenance
The next area to look at is the TADDM database. Updating the database statistics is a very important activity. Without doing this, you may experience dramatic performance degradation. The database statistics should be updated after any major data change in the database, like after adding or deleting a lot of objects. The database should also be reorganized regularly.

The next thing to look for, after making sure the database is not a bottleneck, is TADDM scalability. You can improve storage rate by adding another storage server and you can improve discovery rate by adding another discovery server. Though, the first thing to do is to make sure the current servers are fully utilized.

4. Storage server configuration
If you notice that the number of sensors running is much greater than the “dwcount” property (com.collation.discover.dwcount), then TADDM is waiting to store sensor data. But before adding another storage server, you may want to increase the “topopumpcount” property (com.collation.discover.observer.topopumpcount). This property controls the number of concurrent threads that store data to TADDM database on a discovery server. The default value is 16 but can be safely increased to 24. Keep in mind that this will increase memory and CPU utilization on the storage server, as well as database load. If the overhead is too high, you can try a lower value like 8 or 10. In some environments, this is a better solution.

5. Discovery server configuration
If you notice that the number of sensors running is almost the same as “dwcount” property (com.collation.discover.dwcount), then it means that TADDM is waiting for new sensors to run. To fully utilize a discovery server in this case, you need to increase “dwcount” property. The default value is 32, but it can be safely increased two- to threefold depending on the discovery server capacity. Powerful systems can run more than 96 sensors at the same time. A good practice is to increase the value step-by-step and to monitor discovery performance as you do.

Having these five areas reviewed and configured properly should improve your TADDM experience in various ways. Sensors should no longer be timing out as much and should be able to discover as much as possible. Discovery time is shortened thanks to utilizing the discovery and storage servers up to their full capacity.