I was configuring GoldenGate in HA mode by following this document. Everything worked ok but in the end while running agctl config goldengate to view the configuration of GoldenGate resource, it was failing with the following error:

[oracle@exadatadb02 ~]$ agctl config goldengate GG_TARGET
 Failed to execute the command ""/u01/app/xag/bin/clsecho" -p xag -f xag -m 5080 "GG_TARGET"" (rc=134), with the message:
 Oracle Clusterware infrastructure fatal error in clsecho.bin (OS PID 126367_140570897783808): Internal error (ID (:CLSB00107:)) - Error -1 (ORA-08275) determining Oracle base
 /u01/app/xag/bin/clsecho: line 45: 126367 Aborted                 (core dumped) ${CRS_HOME}/bin/clsecho.bin "$@"
 Failed to execute the command ""/u01/app/xag/bin/clsecho" -p xag -f xag -m 5081 "/u01/app/oragg/product"" (rc=134), with the message:

If you look at the error in bold it sounds kinda obvious that it is not able to figure our where the ORACLE_BASE is. But somehow it didn’t strike me at that moment. So started looking around. If we look at the command it is running, it runs clsecho. This is simply a shell script which in turn calls $CRS_HOME/bin/clsecho.bin . In the script, it sets various environment variables and that is where the problem was. There are lines like:

ORACLE_BASE=
export ORACLE_BASE

Nowhere in the script, it is setting the value of ORACLE_BASE. That was causing an issue. I changed the first line to set the ORACLE_BASE location and it worked fine after that. There was another issue i faced with ggsci after doing xag configuration. Will do another blog post on that.