centos-release-7-3.1611.el7.centos.x86_64

Can you check the um1 install log for any errors: /tmp/um1_binary_install.log or add it here? Also what does getSystemStatus and getProcessStatus show in mcsadmin?

This is the problem output in the install script log:

Error: can't locate mariadb-columnstore-1.0.7-1 Package in directory /root

Please place a copy of the MariaDB ColumnStore Packages in directory /root and press <enter> to continue or enter 'exit' to exit the install >

Please have the binary install tar gz (or rpms if installing that way) available in the root directory. This is then used to scp to the um1 node for installation.

If this is the case then there is still a bug that the install should really fail at this point rather than later.

I can not start mcsadmin:
'# mcsadmin
bash: mcsadmin: command not found...
About Error: can't locate mariadb-columnstore-1.0.7-1 Package in directory /root
I copied binary package to /root and passed this step.

um1_binary_install.log

Did it install successfully after this? Mcsadmin is added as an alias by the columstoreAlias script or you can fully qualify the path to the bin directory or you can add bin directory to your path.

This is result from mcsadmin:
'# /usr/local/mariadb/columnstore/bin/mcsadmin

WARNING: running on non Parent OAM Module, can't make configuration changes in this session.
Access Console from 'pm1' if you need to make changes.

MariaDB ColumnStore Admin Console
enter 'help' for list of commands
enter 'exit' to exit the MariaDB ColumnStore Command Console
use up/down arrows to recall commands

mcsadmin> getSystemStatus
getsystemstatus Wed Mar 8 13:34:54 2017

System mymcs1

System and Module statuses

Component Status Last Status Change
------------ -------------------------- ------------------------

• • • • printSystemStatus Failed = API Failure return in getSystemStatus:MessageQueueClient-Error API
        mcsadmin> getProcessStatus
        getprocessstatus Wed Mar 8 13:35:12 2017

MariaDB ColumnStore Process statuses

Process Module Status Last Status Change Process ID
------------------ ------ --------------- ------------------------ ----------

• • • • printProcessStatus Failed = API Failure return in getProcessStatus:MessageQueueClient API:
        API Failure return in getProcessStatus:write API

Also please check the server firewall settings on all nodes. This type of error generally points to a problem communicating between nodes.

https://mariadb.com/kb/en/mariadb/preparing-for-columnstore-installation/#firewall-considerations

Also the Trouble shooting guide has steps to help diagnosis the issue, which includes checking the firewall settings

https://mariadb.com/kb/en/mariadb/system-troubleshooting-mariadb-columnstore/#multi-node-install-problems-and-how-to-diagnose

David Hill

1. Firewall status:
'#systemctl status firewalld
● firewalld.service - firewalld - dynamic firewall daemon
Loaded: loaded (/usr/lib/systemd/system/firewalld.service; disabled; vendor preset: enabled)
Active: inactive (dead)
Docs: man:firewalld(1)
2.SELinux:
'# cat /etc/selinux/config
SELINUX=disabled
SELINUXTYPE=targeted

CAn you confirm you ran postConfigure again before doing this? Is there anything new in the log file output that indicates an error?

If still stuck, have you been able to do a single server install on one of these servers - might be a place to get started?

I ensure that I disabled firewalld before I ran postConfigure.
My installation a single server is good.
Result:
mcsadmin> getSystemStatus
getsystemstatus Thu Mar 9 14:44:57 2017

System columnstore-1

System and Module statuses

Component Status Last Status Change
------------ -------------------------- ------------------------
System ACTIVE Thu Mar 9 14:44:02 2017

Module pm1 ACTIVE Thu Mar 9 14:43:57 2017

mcsadmin> getProcessStatus
getprocessstatus Thu Mar 9 14:45:25 2017

MariaDB ColumnStore Process statuses

Process Module Status Last Status Change Process ID
------------------ ------ --------------- ------------------------ ----------
ProcessMonitor pm1 ACTIVE Thu Mar 9 14:43:21 2017 2891
ProcessManager pm1 ACTIVE Thu Mar 9 14:43:28 2017 3000
DBRMControllerNode pm1 ACTIVE Thu Mar 9 14:43:34 2017 3408
ServerMonitor pm1 ACTIVE Thu Mar 9 14:43:36 2017 3428
DBRMWorkerNode pm1 ACTIVE Thu Mar 9 14:43:36 2017 3450
DecomSvr pm1 ACTIVE Thu Mar 9 14:43:40 2017 3511
PrimProc pm1 ACTIVE Thu Mar 9 14:43:42 2017 3537
ExeMgr pm1 ACTIVE Thu Mar 9 14:43:46 2017 3579
WriteEngineServer pm1 ACTIVE Thu Mar 9 14:43:50 2017 3663
DDLProc pm1 ACTIVE Thu Mar 9 14:43:54 2017 3706
DMLProc pm1 ACTIVE Thu Mar 9 14:44:00 2017 3740
mysqld pm1 ACTIVE Thu Mar 9 14:43:38 2017 3331
script_install_single.txt

That's encouraging. Does it work ok on both servers? If so that will eliminate each individual server being a problem in which case it must be a problem with server to server but you've eliminated firewall and remote ssh access which are the most common setup problems.

I saw something like this recently where the issue was due to performing the install with a non login root shell (ie. using su rather than su -). Could explain this.