I have installed and configured Postgres-XL 9.5 (revision 1.6) on 4 separate virtual machines and have gotten to where I am able to control each node from the other (start/stop etc.) However, when I attempt to run a Createdb on the coordinator node I receive the following
PGXC Createdb testdb
Selected coord1.
createdb: database creation failed: ERROR: Could not begin transaction on Datanodes
CONTEXT: Error message received from nodes: datanode#1 datanode#2
So it looks like there's an issue on those nodes I log into either Datanode and check the log in the pg_log file and see the following
ERROR: node "coord1_3116" does not exist
STATEMENT: SET globalsession TO coord1_3116;SET parentPGXCPid TO 3116;
As you can see the Createdb command is trying to use the PID as part of the connection string to reach the coordinator, it increments as you would expect if I open something else and try again I'm having difficulty finding where pgxc_ctl specifies the port to find out why it's attempting to plug this in. Here is the conf file so that you can see that what I have specified:
#===================================================
# pgxc configuration file updated due to GTM master addition
# 20170920_15:12:27
gtmName=gtm
gtmMasterServer=10.212.221.141
gtmMasterPort=6666
gtmMasterDir=/pg/gtm
#----End of reconfiguration -------------------------
#===================================================
#===================================================
# pgxc configuration file updated due to coordinator master addition
# 20170920_15:15:49
coordNames=( coord1 )
coordMasterServers=( 10.212.221.142 )
coordPorts=( 5432 )
poolerPorts=( 16667 )
coordMasterDirs=( /pg01/coord )
coordMaxWALSenders=( 5 )
coordSlave=n
coordSlaveServers=( none )
coordSlavePorts=( none )
coordSlavePoolerPorts=( none )
coordSlaveDirs=( none )
coordArchLogDirs=( none )
coordSpecificExtraConfig=( none )
coordSpecificExtraPgHba=( none )
#----End of reconfiguration -------------------------
#===================================================
# pgxc configuration file updated due to datanode master addition
# 20170920_15:17:45
datanodeNames=( dn1 dn2 )
datanodeMasterServers=( 10.212.221.143 10.212.221.144 )
datanodePorts=( 15432 15433 )
datanodePoolerPorts=( 16668 16669 )
datanodeMasterDirs=( /pg01/data /pg01/data )
datanodeMasterWALDirs=( /pg01/data/xlog /pg01/data/xlog )
datanodeMaxWALSenders=( 5 5 )
datanodeSpecificExtraConfig=( none none )
datanodeSpecificExtraPgHba=( none none )
#----End of reconfiguration -------------------------
I can get this running on one node without issue, but for some reason when I try to use another the connection string comes out wrong, I've tried re-downloading the source and reinstalling etc. and even deleting the VM's and creating new ones to figure out what I'm doing wrong, but at this point I'm kindof stuck.
Any help would be greatly appreciated (particularly if you could tell me where the string is calling for the coordinator port so that I can see why it's attempting to put the PID there.)
If it's pertinent I'm running Ubuntu 16.04.3 as an OS on these VM's and the VM host