Mango Performance Tuning
Are there any steps to increase the Java and H2 memory from the defaults? I have large number of points and API use and think increase in these to areas would be helpful. Running on linux Centos 7 with Java 7 and the H2.
I edited the ma-start.sh and changed the memory to 2048m. After doing so, the H2 cache was automatically increased from 54m to 115m. Will monitor for stability, but CPU use seems to be lower now. Below is the config entry:
$EXECJAVA $JPDA $JAVAOPTS -Xmx2048m -Xms2048m -server -cp "$MA_CP"
hussam last edited by
how many points you used, more than ten thousand?
we have about 14,000 currently loaded, with more on the way. The box is on a VmWare with 2 cores, 64-bit CentOs 7, Java 7, and 4 gig of memory. The TOP looks like this:
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1352 root 20 0 3567912 2.229g 22884 S 33.6 60.2 246:32.96 java
jeremyh last edited by
It sounds like your application of Mango could be really interesting. I would love to hear more details, such as the industry you are in, the types of equipment you interface to/data you collect, and the Mango modules you are using. Also how you are using Mango - whether you are using it for control as well as data collection, and what you are doing with your data.
If you can share any details, I'd love to hear them! You could make a post over in the 'user stories' forum section.
@jeremyh I'm swamped at work right but but will be happy to write up when I get a break, probably around 1July. Short story, we are a datacenter company and using the product as a backend data acquisition tool to provide energy and environmental (Critical and Essential) such as room temperature/humidity/air handler information. We have a various set of protocols ranging from ModBus, BacNet, SNMP, SQL, and custom SOAP calls and needed a one stop package that Mango provides. We have a front-end website that makes API calls to the Mango for data point retrieval.