Background story: Sudenly, our monitoring system showed that RS went down. I've checked replication server log file and found following error:
F. 2016/09/28 11:08:10. FATAL ERROR #7036 REP AGENT(ASE_PROD.testdb) - /packapp.c(79)
Additional allocation of '16368' bytes to the currently allocated memory of '524281559' bytes would exceed the memory_limit of '524288000' specified in the configuration.
T. 2016/09/28 11:08:10. (288): Exiting due to a fatal error.
Solution: Solution to this issue is simple - just run Replication Server again. But, what this error means and what cause it?
This error means that Replication Server wanted to allocate more memory then it is allowed to it by parameter memory_limit. In this situation, memory_limit parameter was set to 500MB.
But what caused that RS went out of memory? I've checked some sites comments and found out that some versions of Replication Servers are having memory leak. This picture, which represents Replication Server memory usage can prove that. It showed that Replication Server is steadilly progressing towards memory_limit and that it will crash again.
Conclusion: To prevent this version of RS to crash, do the regular, wanted, RS restart, which will put RS memory to the lowest size.
