High CPU Usage - RPC Errors - Excessive Log Writes
Unknown Causes
There are a number of potential causes for this sort of problem. If you are unable to determine the cause the fix in most all cases is to reinstall the exact same version of FRP on the machine using the instructions in section 1 for Major upgrades and installs found here . This method will preserve the jobs and settings. Do not reinstall over an old installation!
Known Causes
- Versions prior to 6.5.0 had a bug that could cause excessive log writes. An upgrade is needed to at least version 6.5.0 or higher. If you have annual support and upgrade contracts in force for your servers contact support@filereplicationpro.com for instructions on how to upgrade.
- Environments that block multicast on UPD port 9110. This results in RPC errors being written to the logging system to the point it can bring down the service, and or prevent replication while filling up the error logs found in the /logs folder of the installation. Check firewalls to be sure this port is open. Sometimes turning off the autodiscovery feature of FRP will help with this problem, that is done by changing the flag for autodiscovery in the /etc/s2s.properties file on each server to "false" and restarting the service. If a different port can be used for multicast you can change the port used by editing the /etc/properties.xml file with a plain text editior. Change the port from 9110 to the port that you desire, restart the FRPRep service. Repeat on ALL machines.
- It is known that some hosting solutions such Amazon cloud do not permit multicast and you may be forced to consider another provider.