Skip to main content
Announcements
Have questions about Qlik Connect? Join us live on April 10th, at 11 AM ET: SIGN UP NOW
cancel
Showing results for 
Search instead for 
Did you mean: 
paultribe
Contributor III
Contributor III

QVPR - configuring in multi-server environment for GD

Hi,

Our QMS service is running on our publisher server.  we have two QVS/IIS servers. when setting up the MultiCluster_UserConfig_Template do we populate  this line in the config file for each server even if it is pointing to the QVPR which is only on one server? or do we leave it as an empty string on the QVS server configurations?

SET vMultiCluster_QVPRPath_1 = '';    // Repository Path

Thanks in advance

1 Solution

Accepted Solutions
paultribe
Contributor III
Contributor III
Author

Question no longer relevant. by moving log file locations we are now using the standard configuration on the configuration tab.

View solution in original post

15 Replies
Tyler_Waterfall
Employee
Employee

First, do you have more than one cluster or just one cluster with multiple nodes?

In other words, are the logs for both QVS servers being written to the same folder?

If you just have one QlikView cluster with multiple QVS (one QVS log folder) then you don't need to use the multi-cluster config.

paultribe
Contributor III
Contributor III
Author

we are set up as one cluster with two QVS servers in the cluster

Tyler_Waterfall
Employee
Employee

Is there just the one log folder for the QVSs? If so, just use the normal configuration.

Tyler_Waterfall
Employee
Employee

I should add, if you have to use multi-cluster config (because 2 separate QVS log folders), then you only need to use the Publisher log filepath and the QVPR filepath for one of the "servers".

paultribe
Contributor III
Contributor III
Author

okay. I switched configuration files. this is the one I am now using. no failure on reload and I am getting some data. I have to do an analysis on what is missing. a quick run through the sheets indicated user and publisher data has some gaps.


SET vQVGovHome ='G:\Qlikview\Governance_1.1';

SET vProfilePath ='G:\Qlikview\Governance_1.1\profiles\default';

SET vFileRootPath1='G:\QlikView\Documents\Public';
SET vFileRootPath1Exclude='*issue.qvw';
//Path 2
//SET vFileRootPath2=;
//SET vFileRootPath2Exclude=;


LET vServer1QVSFlag = 1;
LET vServer1AuditLogFlag = 1;
SET vServer1LogPath = '\\CCV003\QlikTech\QlikViewServer';
LET vServer1PubFlag = 0;

LET vServer2QVSFlag = 1;
LET vServer2AuditLogFlag = 1;
SET vServer2LogPath = '\\CCV004\QlikTech\QlikViewServer';
LET vServer2PubFlag = 0;


LET vServer3QVSFlag = 0;
LET vServer3AuditLogFlag = 1;
SET vServer3LogPath = 'G:\Qlikview\QDS';
LET vServer1PubFlag = 1;


SET vHistoricalYears = 1;

paultribe
Contributor III
Contributor III
Author

progress is being made however, the dashboard reload is generating errors and without access to the hidden script I cannot troubleshoot. Here is a listing by order in the log file using the user config file from the last comment. Any  ideas on what is generating the errors?

Log Action

Rows Returned

Querying File List

11

Querying write QVW File Entry QVX

Querying QVW tables

443

Querying QVWFields

2066

Querying QvdTableHeader

0

Querying wrtie QVX table QVX

Querying QvdFieldHeader

0

Querying write QVX field QVX

Querying Sessions

0

Querying ConcurrentSessions

0

Querying SessionHeatmap

0

Querying Performance

3876

Querying Events

1

Querying AuditLog

0

Querying PubSourceDoc

0

Querying PubDocTask

0

Querying TaskResults_Initial

0

Querying TableLineage

210

Querying Write QVXLineage QVX

Loaded

Querying QlikView Expressor Dataflow

records returned for QlikView Expressor Dataflow

Querying SOE

25226

File Scan complete

Events table found to have 1 Rows - extracting Node_TS info

Events table has 1 rows where Node field is not null

Performance table found to have  Rows - extracting Node_TS info

Performance table has 3861 rows where Node field is not null

Loading Expressions Table

1464 records returned for Expressions

Error: Error in hidden script. For details, open hidden script before reload.

Error: Error in hidden script. For details, open hidden script before reload.

Error: Error in hidden script. For details, open hidden script before reload.

Execution finished.

Thanks in advance

Tyler_Waterfall
Employee
Employee

I have never tried to do a multi-cluster reload using the UserConfig script. I don't believe that it will provide the needed multiple reloads (scan for each "server") that happens with multiclusterconfig.

However, you say that you have one cluster, so I am not sure why you have logs in two locations.  (I'm not a multi-node QVS configuration expert - but I thought you would or could use the same log folder location for all QVS nodes -- like on a shared drive.)

Useful items for troubleshooting are the following.

  1. Screen shot of the “configuration” screen of the Governance Dashboard
  2. Governance Dashboard.qvw.log
  3. fullscan.messages.log*
  4. fullscan.statistics.log*
  5. Description of how your environment is configured:
    1. Version of QlikView
    2. Where is GovernanceDashboard.exe installed?
    3. Where is the reload occurring?
    4. User account that is reloading it - does it have read/write access to the profiles and file folders?

* in [install directory]\governance_1.1\profiles\default\QVX_LastRun folder ==> Looks like you are using G:\Qlikview.

paultribe
Contributor III
Contributor III
Author

Hi Tyler,

Thanks for the detailed reply. All "G:" drive entries are on the QDS/QMS server. I have also tried using the UNC path to that server which I understand I have to use if trying to publish the file using QDS and the results are the same.

As for QVS log location you are probably correct. At this point our INT environment has to mimic our Prod/Staging environments which were configured by QlikTech consultants when we were not knowledgeable about QlikView so that is something we will consider. 

I will attach the following files as you recommended:

  1. Screen shot of the “configuration” screen of the Governance Dashboard
  2. Governance Dashboard.qvw.log
  3. fullscan.messages.log*
  4. fullscan.statistics.log*
  5. Description of how your environment is configured:
  6. Version of QlikView: 11.2 SR10 with IIS patch - 11.20.12361.0
  7. Where is GovernanceDashboard.exe installed? it is installed on the "G" drive of the QDS/QMS server. this is a SAN attached drive which due to firewall rules can be accessed by the QVS/IIS servers which are in a different security zone
  8. Where is the reload occurring? doing it manually on the QDS server in the 'default' profile directory.
  9. User account that is reloading it - does it have read/write access to the profiles and file folders? Using my account which is admin and created all the shares being utilized
Tyler_Waterfall
Employee
Employee

Just realized that I've been working on this same case with Qlik Support but with different information

Here's the latest response to support:

Some questions and observations:

  1. Although you have the UserConfig toggled on, the variables declared in the userconfig are not found on the configuration of the GovDB. Usually you put the variables for the filepaths that you declared in the userconfig txt in the configuration page. Otherwise the GovDB load script doesn’t really know what to do with the variables.
  2. So, the scan is going to take the values set in the configuration, namely:
    1. Qvws and qvds from the file share \\CCV752\Qlikview\Document\Public
    2. Server files from \\CCV003\QlikTech\QlikViewServer
    3. Reload G:\Qlikview\QDS
    4. No Task (QVPR) files – repository path is blank
  3. Remind what\where is G:\ ?
  4. My take is that, if the server files are found in two separate folders (non-clustered nodes), then you need to use the multiclusterconfig setup
    1. Server 1 could include:
      1. fileshare \\CCV752\Qlikview\Document\Public
      2. server = 003
      3. qds on g:\QlikView\QDS (can you use UNC to be clear?)
      4. qvpr on same g:\...
    2. Server 2 would then be
      1. Fileshare? (just point to a folder with one or more QVDs or QVWs – required)
      2. Server = 004
      3. No publisher
    3. Aggregator - see Create a MultiCluster configuration script.
    4. General folder structure for profiles > per documentation (copy template folder and rename for each “server1”, “server2”, “aggregator” and “global”

Based on what I understand of this environment, you could get away with just using the basic configuration (in the GovDB page) if the two servers 003 and 004 wrote their log files to the same folder. But if not, then Step 4 is needed to get all the files loaded into the GovDB.