Running "pxf [cluster] register" command fails with the following error message
[gpadmin@ltvidagpma01 ~]$ pxf cluster register
Installing PXF extension on coordinator host, standby coordinator host, and 6 segment hosts...
ERROR: Failed to install PXF extension on 8 out of 8 hosts
..... ==> install: usage [-1cDlnNpqsv] [-C config] [-H links] [-S links] [-g group] [-m mode] [-o owner] files destination
install: usage -d [-nqrv] [-C config] [-g group] [-m mode] [-o owner] directory...
:::
This leads to subsequent failure to move pxf.control file to the Postgres extension path.
pxf script is executing incorrect version of the "install" command.
Ensure correct "install" (/usr/bin/install) is being referenced through running the following command:
[gpadmin@gp6-m ~]$ which install
/usr/bin/install
[gpadmin@gp6-m ~]$
Following command can be used for enabling bash tracing to further troubleshooting this error:
$ which pxf && cd $(dirname $(which pxf)) && bash -x pxf register
Example output snippet for successful run:
:::
+ install --verbose --mode=0644 /usr/local/pxf-gp6/gpextable/pxf.control /usr/local/6.28.0/greenplum-db-6.28.0/share/postgresql/extension/pxf.control
removed ‘/usr/local/6.28.0/greenplum-db-6.28.0/share/postgresql/extension/pxf.control’
‘/usr/local/pxf-gp6/gpextable/pxf.control’ -> ‘/usr/local/6.28.0/greenplum-db-6.28.0/share/postgresql/extension/pxf.control’
+ [[ -d /usr/local/pxf-gp6/fdw ]]
+ exit 0
[gpadmin@gp6-m bin]$
Example output snippet for failed run:
:::
+ install --verbose --mode=0644 /usr/local/pxf-gp6/gpextable/pxf.control /usr/local/greenplum-db-6.28.1/share/postgresql/extension/pxf.control
install: usage [-1cDlnNpqsv] [-C config] [-H links] [-S links] [-g group] [-m mode] [-o owner] files destination
install: usage -d [-nqrv] [-C config] [-g group] [-m mode] [-o owner] directory
install: usage -h
install: usage -R [options] target
install: usage -V [-C config]
+ fail 'cannot install pxf.control to '\''/usr/local/greenplum-db-6.28.1/share/postgresql/extension/pxf.control'\'''