LAE

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: spalluru

Hi ,
I have implemented 2 Done pools one for normal BRE users and one for Project as below , when the batch graph is running through laeController
I am trying to monitor the total drones ( we assume it will be java threads on server) present on the server.
I see that there are 841 java threads running on server and lae server has spawned 137 threads on server. As per my understanding BRE pool has got 20 +
project pool has got 80 which means there should be total 100 threads should be running on the server. But may I know why we have got 841, 137 threads
shown on server . Please note this is Linux server and I got these thread counts using Geneos Monitoring tool.

Can you please comment on this behaviour? correct me if my understanding on drone pool/threads is wrong ?iMAFA_laeserver_thread_count_137.jpgiMAFA_totaljavathreads_count_841.jpg

# An unlimited pool for BRE users
pool:bre
size:20
defaultgroup:users
end:bre

# A pool for running ABC graphs

pool:ABC
size:80
defaultgroup:ABC
defaultuser:abcuser
end:ABC
-Sreekanth

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: ejones

I have not pooling very much, so I hope that someone with more experience will comment. I tested it by counting the number of connections I could make to the server in BRE. And I've always understood that the drones used by an executing BRX would also be included in the pooling counts which I believe it is what you are attempting to count as well using the Geneos Monitoring tool. And the number of processes you are seeing is clearly more than the 100 limit you've defined as the absolute maximum.

The counts there may be higher because there are other processes involved for managing the threads. When a BRX runs there has to be a controller process that decides which nodes to run and kicks off the threads that run each node. I'm sure there are more processes than this, especially in version 5.0 with the web server.

Here I'm starting to speculate hoping to be corrected, but it would not surprise me to learn that when process finishes running a node it may not immediately exit because it might be more efficient to stay around waiting for another assignment.

When I counted processes in my drone testing I did it all from BRE. I mainly used the counts shown at the bottom of the window. But in BRE one could also go to Tools -> Server -> Status... and see what threads are running, which graphs, and user logins. It also shows threads that are allocated and not doing anything. And most importantly, it gives you the total number of threads and total number of drones which would be closer to what you are wanting to monitor.

Note: This was originally posted by an inactive account. Content was preserved by moving under an admin account.

Originally posted by: spalluru

Hi ,
Thanks for the update. So If I understand correctly for counting total number of drones, we should see the number of threads shown for lavastorm server process.
and can you also comment on java processes and threads which are spawned by these drones do we have any control on this ? is there any parameter to control
spawning of java threads for these drones ?
-Sreekanth