MQ Tables COLD restart

Document ID : KB000088395
Last Modified Date : 14/04/2018
Show Technical Document Details
Issue:
MQ Tables COLD restart
Resolution:
Detailed Description and Symptoms
 
We are preparing for a Disaster Recovery test and want to know what we might need to prepare the database. 

Investigation

MQ stands for "Message Queue".  The Message Queue tables basically act as a work space for all of the information traveling between Automic components. 

There are 14 MQ tables in v9 and v10

  • MQCP001
  • MQCP002
  • MQCP003
  • MQCP004
  • MQCP005
  • MQDWP
  • MQLS
  • MQMEM
  • MQPWP
  • MQSRV
  • MQWP
  • MQQWP
  • MQOWP
  • MQRWP

15 MQ tables in v11.1 (one more table because the JWP has been introduced)

  • MQCP001
  • MQCP002
  • MQCP003
  • MQCP004
  • MQCP005
  • MQDWP
  • MQJWP
  • MQLS
  • MQMEM
  • MQOWP
  • MQPWP
  • MQQWP
  • MQRWP
  • MQSRV
  • MQWP

26 tables in version 11.2 and v12 because of the Zero Downtime Upgrade:

  • MQ1CP001
  • MQ1CP002
  • MQ1CP003
  • MQ1CP004
  • MQ1CP005
  • MQ2CP001
  • MQ2CP002
  • MQ2CP003
  • MQ2CP004
  • MQ2CP005
  • MQ1DWP
  • MQ1JWP
  • MQ1OWP
  • MQ1PWP
  • MQ1QWP
  • MQ1RWP
  • MQ1WP
  • MQ2DWP
  • MQ2JWP
  • MQ2OWP
  • MQ2PWP
  • MQ2QWP
  • MQ2RWP
  • MQ2WP
  • MQLS
  • MQMEM
  • MQSRV


 
Sometimes, old messages can be left in the MQ tables causing the Automic system to not be able to start correctly, or freeze.  The most common reason for this is if the Automic system has been brought down ungracefully.  The best example is if the server that Automic is on has been rebooted or crashed unexpectedly as in a Disaster Recovery (DR) scenario.
When the machine stops suddenly, the messages in the MQ tables are left.  When Automic attempts to start the system in "Normal" mode, it will try and process these messages.  Because they may no longer have valid information, the WPs or CPs will not connect properly and the system will be unable to start.


Solution
In most cases, a "Cold" restart is done either by setting the startmode=COLD in the ucsrv.ini file or by right-clicking the WP process in Service Manager Dialog and selecting "Start Service with" > "Coldstart"
 

This should clear the MQ tables.  However, because of potential database issues (locked tables, etc.) you still may not be able to start your system.  You can select from the MQ tables to confirm that there are records left in these tables.
If this is the case, you will have to truncate from the MQ tables. Please contact support for the commands to perform this action.