Improve CAS Cluster Performance
I would like to know if someone has some documentation about how to improve the performance of CAS when working with Cluster Farm.
I already did some configurations like increase the value "DMI_AGGR_ROWS_TO_SEND" and I also know that is possible to increase performance enabling the cluster to work with cache (more details https://community.compuwareapm.com/community/display/DCRUM122/Enabling+DMI+Cache+Support+in+Server+Clusters).
Besides that, are there any best practices to follow that will improve the performance of CAS Cluster?
PS: I know that my performance is poor because the DMI Application Health Status takes more than 20 seconds to load on the CAS Master when it is aggregating data, but if I select to see data only from one of the Slave CAS'es, the DMI loads in about 2 seconds.
Adam, I was searching and I found the configuration "DELTA_THREADS" in the common.properties file.
Searching in the community, I found that this configuration is the number of threads that the Delta user is able to execute in the database, correct?
So, if I increase this default value, and my database is good enough to handle with more threads from this user, will I increase in performance, once I will be able to get more data with more connections in a shorter period of time?
The only thing has left is going to 12.3 for Report Server Load Balancing ...
Thanks for the reply Adam.
Just one more question, there is a limit to increase the value "DMI_AGGR_ROWS_TO_SEND" or it depends of each environment?
I searched about it, but I haven't found any "warning" about it.
Thanks!
Increasing this params specifies how much data can be transferred from one CAS to another in a Farm. It was designed to preserve CASes performance and not allow them (or rather users) to transfer to much.
So increasing would be rather a response to not receiving all desired huge set of data but decreasing would rather speed up the farm ...
