Yes, sometimes I wonder who made these instructions
A couple of comments:
- One profile for each Server/Publisher (or just Server). So, if you have 2 QVS (non-clustered) running, you will have 2 profiles. You can put the Publisher with one of the servers. Below is what I think you;ll need.
- For each profile, you need to create a profile folder. You have one (RCT). But you will need one for the 2nd QVS profile, one for the Aggregator, and one for the global aggregator. Copy the template profile folder for this and rename it. (You will end up with 4 profile folders). For this post, I will call the additional profiles: SERVER_No2, Aggregator, Global. You will need to use the same filepath in the template as the name of the profile folders.
- Currently each profile must include QVW/QVD files to scan (i.e. you can't just scan QVS logs in a profile.). Due to this, I point the 2nd profile to scan the RCT profile folder (please note the bullet below about using UNC) and I exclude the QVX_LastRun and QVD_Historical folders -- no need to scan these files.
- You are referencing C:\ for the profile folder -- is this the local C:\ of the machine that will reload the Dashboard? If not - or if the Governance Dashboard might reload on different Distribution service servers, please use UNC path instead of C:\.
- Other housekeeping: remove all 'unused' Server and aggregator variables, as described below.
- Publisher DateTime format - refer to doc on this -- you can find this in the TaskResults xml files.
//Here's what the script will look like:
// 1. How many distinct "Servers" (Source Document folder(s) and standalone or clustered Server/Publisher) will be scanned?
SET vHowManyServers = 2; // two distinct QVS (not-clustered)
// 2. How many Aggregators are needed?
SET vHowManyAggregators = 1; // 1 aggregator per 5 profiles
//3. How many years of server log history is needed?
SET vHistoricalYears = 1; // Years back from today of server log history.
//// "SERVER" Configuration
// Server 1
SET vMultiCluster_ProfilePath_1 = 'C:\ProgramData\QlikTech\Governance_1.1\profiles\RCT';
SET vMultiCluster_FilePath_1a = '\\hbeu.adroot.hsbc\dfsroot\GB002\corep_mi_qlikview\QlikView_AccessPoint';
SET vMultiCluster_ServerPath_1 = '\\gbw05041.hbeu.adroot.hsbc\QlikViewServer';
SET vMultiCluster_HasAudit_1 = 1; // Do you have audit logging set in QMC?
SET vMultiCluster_PublisherPath_1 = '\\gbw05040.hbeu.adroot.hsbc\QlikTech\DistributionService'; // Has Publisher?: Log Path
SET vMultiCluster_QVPRPath_1 = '\\gbw05040.hbeu.adroot.hsbc\QlikTech\ManagementService\QVPR';
SET vMultiCluster_PubDateTimeFormat_1 ='CCYY-MM-DD H*24:MI:SS'; // What's your datetime format?
// Server 2
SET vMultiCluster_ProfilePath_2 = 'C:\ProgramData\QlikTech\Governance_1.1\profiles\SERVER_No2';
SET vMultiCluster_FilePath_2a = 'C:\ProgramData\QlikTech\Governance_1.1\profiles\RCT';
SET vExcludeFiles_or_Folders_2a = 'QVX_LastRun|QVD_Historical';
SET vMultiCluster_ServerPath_2 = '\\gbw05042.hbeu.adroot.hsbc\QlikViewServer';
SET vMultiCluster_HasAudit_2 = 1;
SET vMultiCluster_PublisherPath_2 = '';
SET vMultiCluster_QVPRPath_2 = '';
SET vMultiCluster_PubDateTimeFormat_2 ='CCYY-MM-DD H*24:MI:SS'; // What's your datetime format?
// REMOVE all the other servers in the template -- Server 3 - Server 6
/// AGGREGATOR Configuration
//// Start with 101 and increment sequentially (do not skip numbers!)
// Aggregator 1
SET vMultiCluster_ProfilePath_101 = 'C:\ProgramData\QlikTech\Governance_1.1\profiles\Aggregator';
SET vMultiCluster_FilePath_101a = '$(vMultiCluster_ProfilePath_1)';
SET vMultiCluster_FilePath_101b = '$(vMultiCluster_ProfilePath_2)';
SET vClusterName_101a = 'gbw05041'; // Add unique name for each "cluster"
SET vClusterName_101b = 'gbw05042';
// REMOVE the other Aggregator 2 from the the template
//// GLOBAL Aggregator Configuration
// GLOBAL Aggregator
SET vMultiCluster_ProfilePath_201 = 'C:\ProgramData\QlikTech\Governance_1.1\profiles\Global';
SET vMultiCluster_FilePath_201a = '$(vMultiCluster_ProfilePath_101)';
Hi Tyler, thanks for that. You've explained it in an easy to understand way.
Yes "C" is the local drive of the server that the Governance Dashboard is installed on (not one of the servers in the cluster).
I created the relevant profile folders and updated the script accordingly:
When I try to scan it takes 2 seconds and I get the following error:
I don't understand why I get that error as I can copy and paste that path straight in to explorer on the server and it takes me to the correct please. It's also the same path I used to scan a single server instead of the cluster and that worked. Any ideas?
btw GBW05041 and GBW05042 are clustered (i.e. same QMC and AccessPoint).
I hit reload instead of scan and it almost worked. It certainly ran longer and scanned both servers but came up with an "error in hidden file" which I ok'ed and it continued to finish. However the Application data got a bit corrupt, it was missing the File Name. Gonna mess about with the settings and retry.
I recommend using the Reload (or CTRL+R) instead of the Scan button. The scan button tries to validate the folder paths which sometimes is a hard thing for QlikView for some reason.
When you try again, you can remove the files in the QVD_Historical folder of each profile. Not necessary, but cleans things up a little.
If you get the Error in Hidden Script message, open up the Governance Dashboard.qvw.log file and search for "Error:" (note the : at the end) and see where it is failing -- I included trace messages in the log that should help you see if it is failing during "Server 1" or "Aggregator 1", for example, as well as which tables were just loaded into memory.
Tyler, I would like some clarification please.
How can I tell if my environment is “clustered” or “non-clustered”?
My environment seems to be the same as Shane has described in his post. I’ve got QVS log files (Audit, Events, Performance, Sessions) being generated on 2 servers (QV1, and QV2). The QV1 server has everything, while the QV2 server has only QVS. There is a common Documents share being used.
In your response you make reference to “having 2 QVS (non-clustered) running”. So based on what Shane has described, dose he have a “non-clustered” environment?
If my environment is clustered what type of scan profile do I need to use?
Thanks for your time.
rasnyder - You can tell in one of two ways:
1) In QMC > System > QVS folder, click on the QVS in question and click on "General". You will see an entry for each QlikView Server
2) See how many QVS log folders there are. There is just one location (folder) per QlikView Server (single node or clustered). So, if you have two separate QVS folder (server A and server B), then I would say you have an non-clustered, two QVS environment (which would require the "multi-cluster" script in GovDB).
Per point 1) of your response, see the screen print below.
It would appear that I have a clustered environment. Yet I have each QVS writing to separate locations, local to each QVS.
Is is possible that my "clustered' environment was/is setup wrong?
Should the "Log Folder" in the screen print below be set/changed to a location common to both QVS's?
As it is now, each QVS has a local Log Folder as is set in the screen print'
I still plan to use the "multi-cluster" script as described above in this discussion unless you now advise different.
Interesting. I would have said to write the QVS logs to the same folder on a share somewhere easily accessible by both QVS servers.
If you are getting the logs sent to two separate servers then I would use the Muli-cluster config script.
Your other option would be to update where the two QVS logs are written. (You'd probably want to move the existing logs from both servers to the new location.) Talk it over with the Qlik admin of course!
Hi Bob, my clustered environment is the same set-up as yours:
The logs are written locally to C - I believe that's the default location.
I suppose if it had been set up to store logs from both servers in to a single share I would not have had to use the "MultiCluster_UserConfig_Template.txt" which is a lot simpler!