The QlikView Management Console displays Cluster is down or disconnected and the status of all tasks is listed as Unrunnable

[Sonja_Bauernfeind]

The QlikView Management Console under Status tab displays its Distribution Service as Cluster is Down under Started/Scheduled column and lists all tasks as Unrunnable for status. This behavior is unique to QlikView 12.10 and later and most often occurs after a full server restart or when all QlikView Services have been restarted. 
 

Root Cause One:

There is a hard coded timeout of 20 seconds for the workorder transfer from QMS to QDS service to complete. If a customer has a lot of old disabled tasks, or has had to lower their Chunk Size in the past, this can cause the workorder communication to take longer than 20 Seconds. It will attempt to start over and continue in this loop. This will cause the QMC to list tasks as “Unrunnable” , the QDS status to show as “Cluster Down” and errors in the QMS logs regarding “Quota size errors.” 

Other delays such as virus scans or network latency issues can lead to the same behaviour.

Root Cause Two:

Issue may have occurred due to Application Data content corruption.
 

Resolution:

Note: Make sure to perform the basic troubleshooting steps under Reload Engine disconnected task unrunnable cluster is down prior to steps below.

1. Upgrade to the latest available QlikView release. Improvements have been made across versions.
   
   
2. Verify that virus scans and other threat protection software has been set up as per: QlikView Folders and files to exclude from antivirus scanning 
   
   
3. Increase the "MaxReceivedMessageSize" setting found in C:\Program Files\QlikView\Management Service\QVManagementService.exe.config. For more information see Some distribution tasks not starting with 'error (413) Request Entity Too Large' in QlikView Managem...
   
   
4. The issue may also be resolved by adjusting the Chunk Size to a higher value, or cleaning out old disabled tasks in order to reduce the time the Workorder takes to complete. In some cases it may be necessary to decrease the chunk size (e.g: to value 75). You may verify if the chunk size needs to be adjusted by performing the following steps:
   
   
   1. Open the Qlikview Management Console (QMC) > System > Setup > Management Service >  General tab and set the Logging Level to Debug logging.
   2. Go to the Distribution Services and the QDS resource next on the same area of the QMC and go to the Advanced tab and click the Send Workorder button. 
   3. Go to the Management Service log, which is located in  C:\ProgramData\QlikTech\ManagementService\Log by default and open the current log and scroll to the bottom of the file.  What you are looking for is a message regarding the chunk size being too large there. 
      
      See example below. If similar is found, proceed to step d.
      
      The remote server returned an unexpected response: (413) Request Entity Too Large.
      Information    Failed to send message to distribution services. The message is removed from the queue. Error details: "(if (workorderresult.ODSeRotateResult) - SendSaveAllTasks) Failed to contact distributionservice at http://<servername>:4710/QDS/Service, chunkErr=QDS did not respond to request.Last exception (for http://<servername>:4710/QDS/Service): The remote server returned an unexpected response: (413) Request Entity Too Large." Message Type: "SaveAllTasks" Message Parameter: ""
      
      
   4. Open the Management Service config file located in C:\Program Files\QlikView\Management Service\QVManagementService.exe.config and look for the following lines and adjust the size lower (.i.e 75), restart the service and recheck the QMC\Status view.     
      
      <!-- Message size for QMS workorder -->
      <add key="QMSChunkSize" value="100"></add>
      
      
5. In order to increase the frequency of these communication attempts between QMS and QDS see QlikView Tasks status in the Management Console: Unrunnable
   
   
6. If neither above resolves the issue, How to recreate the QlikView Distribution Service folder may resolve the issue.