- Pivotal Greenplum Database (GPDB) 4.3.x
- Operating System- Red Hat Enterprise Linux 6.x
The following steps led to the symptom of this issue:
- The database was recreated using backup and restore.
- gpperfmon database was also reinstalled.
- When attempting to start, the old Pivotal Greenplum Command Center (GPCC) console for the newly created cluster results in the error, "ERROR: schema 'gpcmdr_instance_<instance_name>' does not exist".
gpadmin@greenplum-mdw /data/master/gpseg-1/gpperfmon/conf> gpcmdr --start Starting instance gp_cc... Failed to start Command Center instance dca_cc: Error executing query select oid from pg_filespace where oid not in (select distinct(oid) from gpcmdr_instance_gp_cc.fsmapcache): ERROR: schema "gpcmdr_instance_gp_cc" does not exist LINE 1: ...space where oid not in (select distinct(oid) from gpcmdr_ins...
When the command center console GUI is first set up, it creates a schema in the gpperfmon database which stores most of the information about the cluster. The schema is created with the naming format "gpcmdr_instance_<instance_name>". An example is shown below:
gpperfmon=# \dn List of schemas Name | Owner ------------------------------------------+--------- gp_toolkit | gpadmin gpcmdr_instance_gpcc_4380_200_1459944426 | gpmon information_schema | gpadmin
During the gpperfmon reinitialization (i.e. drop and recreate), the schema information is lost.
Create a new GPCC console instance using "gpcmdr --setup".
If you wish to choose the old GPCC console instance name, then first move or rename the directory for the instance name that was created during that last installation.
and recreate the console schema using "gpcmdr --setup" using the old GPCC console instance name.
Note: Reinstalling the GPPC console instance (i.e. GUI) using "gpcmdr --setup" does not need a database restart.
Refer to the documentation on how to install the GPCC Console.