Why does CPU usage spike when running Checkpoint Add-On version 2.1?

[redman1138]

We were running the Checkpoint add-on v. 2.02 and recently upgraded to 2.1 and we notice that when we have several connections to CLM's and CMA's the CPU load on the HF spikes up. We have 8 connections to CLM's and 8 connections to CMA's and one a CentOS system with 6 CPU's, we average around 88-93% CPU load. This is the case on all 8 of our HF's with this app. If we disable the app, cpu utilization drops to 5-8%. I have tried the no_nagle as well as various options for conn_buf_size. On the older version (2.02) we had to set SPLUNK_REST_STATUS_COMMIT=1000000 for it to collect the logs in a timely manner. Right now, that has been disabled.

Any suggestions or assistance would be greatly appreciated.

[jcoates_splunk]

just posting an answer to get this out of my filter... Probably because you're asking for a lot of work to be done, and the software is then doing it, and the hardware isn't robust enough to balance that load.