ITCAM for Transactions: Robotic,RPT,best practices,pitfalls Part2
ericmtn 1000009W88 Visits (5869)
Part 2 of this blog
9) kt6agent jvm and RPT execution runtime jvm, max heap size increase (KT6_JVM_OPTIONS - kt6env - managed_jvm.xml)
If you start to see that your robotic script playbacks are running fine and produced monitoring data in TEP robotic views for some times, then suddenly there is no more monitoring and response time data reported in TEP, it's possible that too many scripts are played back and the T6's JVM or the RPT execution runtime's JVM got out of memory. Review and apply the following steps, as it could help....
9.1) Check in TEP under robotic response time if you see "overrun" playback status messages, which is in indication that you should either reduce the number of robotic script playbacks scheduled on this RRT t6 agent player; or increase the playback frequency to a higher value, like for example from every 5 min to 10 min.
9.2) Attempt to increase the max heap size of KT6 agent's jvm, or RPT runtime's jvm :
KT6 agent JVM:
You can control the behavior of the KT6agent JVM, to a certain extend only, by modifying the file kt6env
(for example : C:\I
Remark:Some parameters like:
JVMJ9VM081W Malformed option value, option "-Xmx512m" contains trailing characters " " which have been ignored
Either amend the existing "Xms" or "Xmx" entry, or add the following line:
Remark: Windows OS maximum setting:
Note: We have found that anything > 1260m crashed T6 agent on Windows. If you set the maximum (-Xmx) too high, the JVM process runs out of memory.
T6 agent only use 32-bit JVMs (so far) so the process limit is 2GB.
There is no lower limit to the minimum (-Xms), but if you set it too low and memory resources are low, the JVM takes a long time to start.
Please note: Do not exceed a 1200 MB max heap size on 32-bit Windows
The default settings are:
RPT runtime's jvm :
So for example, you can increase the amount of java memory available for RPT script playbacks .
The default settings are::
The maximum value for:
is determined by the machines physical memory and "bitness"
Physical Memory 32bit OS 64bit OS 1GB 512m 512m 2GB 1024m 1024m 3GB 1200m 1536m 4GB 1200m 1900m 6GB 1200m 1900m 8GB 1200m 1900m 16GB 1200m 1900m
Using these parameters and increasing the JVMs has been helpful sometimes when you are experiencing symptoms like:
- T6 agent going offline in TEP
- kt6agent crashing
- "Exception 0xC0000005 (ACCESS_VIOLATION)" message found in kt6agent.ras file
Remark: Other RRT/T6 agent possible parameters to consider. The following parameters can be modified by reconfiguring the RRT / T6 agent:
1. Robotic monitoring Configuration > "Maximum number of concurrent robotic playbacks"
Increasing this value will allow more playbacks to run concurrently but places a higher need for machine resources including memory \ cpu \ java memory allocation for the agent and RPT
2. Data Analysis Configuration > "Maximum number for processing threads"
The "Maximum number of processing threads" option purpose is to limit the number of threads that are spawned at one time to process arm data (.dat) files.
Without this parameter, on startup if there are a large number of files to be processed, it is possible that the number of threads spawned at once could mean a big hit on machine resources ultimately leading to the T6 agent ceasing to run.
There shouldn't be any need to amend this parameter unless you experience problems with large numbers of "dat" files causing problems when an agent is restarted or during the ongoing processing of high volumes of data that result in problems when viewing data in one or more TEP workspaces or views. This might be needed if you have increased "Maximum number of concurrent robotic playbacks".
Consider adding only + 10 to one or the other of these 2 above parameters, restart the T6 agent, and then assess behavior of the robotic agent over a period of time.
Subscribe and follow us for all the latest information directly on your social feeds: