On Linux guest operating systems communication between the agent and scheduler can be delayed if there isn't sufficient entropy available. One can verify the level of entropy available by executing the command below. A reading of > 2000 is recommended for normal operation of the agent.
cat /proc/sys/kernel/random/entropy_avail
Steps to identify the problem:
The agent's transmitter log with log level set to 8 is required to identify the problem. If there are any delays (than normal operation) between lines 'Socket factory has been obtained' and 'Communication protocol has been created', these delays might be due to low entropy levels.
11/10/2015 10:28:13.812 CST-0600 5 TCP/IP Controller Plugin.Transmitter pool thread <Regular:2>.CybTargetHandlerChannelLogHelper.logConnectionAttempt[:75] - Attempting to open conversation to [email protected]_development:7507 using plain socket
11/10/2015 10:28:13.812 CST-0600 7 TCP/IP Controller Plugin.Transmitter pool thread <Regular:2>.CybTargetHandlerChannel.constructConversation[:1138] - Socket factory has been obtained
11/10/2015 10:28:45.931 CST-0600 7 TCP/IP Controller Plugin.Transmitter pool thread <Regular:2>.CybTargetHandlerChannel.constructConversation[:1141] - Communication protocol has been created
11/10/2015 10:28:45.935 CST-0600 5 TCP/IP Controller Plugin.Transmitter pool thread <Regular:2>.CybTargetHandlerChannelLogHelper.logConnectionInfo[:109] - Opened conversation to [email protected]_development:7507 with partner at 10.10.111.111:7507 with timeout of 10000 from 10.110.110.110:44854
Solution:
Increase the entropy level available to the system.
One Red Hat based systems one can use rng-tools with --rng-device=/dev/urandom
rngd --rng-device=/dev/urandom
There are other hardware based solutions available as well, consult your Unix administrator prior to implementing the above solution.