Qlik Community

Qlik Support Techspert Thursdays

Here you'll find the recordings of our Qlik Support Techspert Thursdays sessions.

Highlighted
Not applicable

The QlikView Server Management Console from a Support expert perspective 3/15

Follow Brett Bless (Principal Technical Engineer) in discovering the QlikView Server Management Console.

Deep dive into functions, tuning and dependencies. Brett will explain how to avoid configuration issues and share best practices from more than 10 Years of QlikView support experience.



For best viewing, we recommend you set the highest video resolution.


Q&A - The QlikView Server Management Console from a Support expert perspective

Q: Shall we assign named cal and document cal users in authorization of document?

A: CAL and Authorizations are really two distinct things, users must have either a Named or Document CAL in order to be able to open a document on the server, but when QVServer is in DMS Security mode, you must also ensure the users that require access have also been specified in the Authorization tab settings and/or you set All Users or All Authenticated in the Authorization settings.  To state things another way, the CAL is what allows a user to have access to the actual QVServer, the Authorization is what determines which apps the user may actually open, although Document CALs are app specific, so it can be a bit confusing keeping things in their correct definitions in those cases.

Q: We are a little confused about the use of hyphens in and the length of a server name. Can you explain the max length and the allowed characters of the server name, e.g. the use of hyphens?

A: I was able to find the following Microsoft Support KB link that I believe will answer both questions quite nicely.  Hyphens are not restricted, so they should be fine, and the article explains the NetBIOS name is limited to 15 characters, as the 16th is a reserved space for the suffix information.

https://support.microsoft.com/en-us/kb/909264

Q: How can I automatically remove the "$LASTKNOWNSTATE" elements stored in the ".shared" files?

A: I was not able to find any means to automatically remove the $LASTKNOWNSTATE bookmarks, the only current option would be to turn off Session Recovery option in the QlikView Server resource Documents tab settings in QMC and then have the users in the AccessPoint view click the ‘view details’ link on the thumbnail they wish to then remove the state bookmark and then click the ‘Remove last document state’ button there.  I did confirm these bookmarks are not visible in the QMC Server Objects under User Documents either.  I would reference Article 4042 – Bad Performance Opening a Document… for additional information.

Q: Anyway meta file is related to DMS authorization?

A: Yes, the DMS Authorization information is stored to the .meta file.  One means to see what is actually in your .meta file is to use the Shared File Viewer power tool and open the .meta file in question, then go to the File menu in the Shared File Viewer power tool and choose Export to XML, which will then prompt for store location and name, but you can then go open the XML file and see what the contents are, which may prove helpful in troubleshooting some issues.

Q: What is the QVPR viewer tool called again?

A: To easily view the QVPR files, use the XMLDBViewer Power Tool.  This tool needs to be launched using the right-click Run as Administrator option as well, but it will allow you to search across all files/tables and save changes you make.  Please note that when changing entries, sometimes it is necessary to click out of the field you edited/changed in order to be able to save that change, something I learned the difficult way that may save you some trouble, so if you make a change and wish to save, but the save option is not clickable, then try clicking out of the field in the table you changed and the save option should become live then.

Q: The only way to keep it up-to-date currently is a manual excel file, but if we were able to manage SA in the QMC all AD changes would automatically flow through the reports.

A: You do have the ability to use/configure Section Access via the QMC.  This is done via the Users tab and Section Access Management sub-option there, but this is only available if you have a licensed Publisher, the Reload Engine does not allow this functionality.  You still need to manage the lists though no matter how you do things with Section Access.  Using Loop and Reduce/Loop and Distribute via Publisher can be a much better approach than Section Access and Dynamic Data Reduction if the date for each user/group is not the same.

Q: For clusters we had some problems changing settings.ini to xml for pgo, most likely synch. problem or?

A: Once the PgoAsXmlAlso=1 setting has been added to one of the cluster nodes, the setting should propagate to the shared Inidata.pgo file in the Root Folder path of the QlikView Server resource in the QMC, which will in turn be picked up by the other QVServer instances in the clusters.  The XML files are only written to the Root Folder path as well.  If things do not seem to be making sense, I generally check the Root Folder pgo file timestamps with those on each of the QVServer nodes in ProgramData\QlikTech\QlikViewServer.  If anything is out of synch by more than a day, you may need to stop the QVServer service on all nodes, and restart the services.  The one thing to check in this case is the node you start first, has a good Settings.ini file, i.e. all the options are correct set in that version.

Q: In a qvw-file I added the script line "execute cmd.exe /c del <anyfile>". My problem is that the publisher which executes the qvw-file does not execute the delete command. Why?

A: I assume it runs when you run it from the Desktop Client?  If so, that means you have the ‘Can Execute External Programs’ option set in the Script Editor on the Settings tab.  The only thing to which I could point is checking that the paths are identical in the Desktop environment and the Server environment, and if you have spaces in the path, that things are double-quoted as well there.  I am not aware of any reason this would not work other than the delete command is not allowed to run on the server etc., which could be if domain security policies are in use etc.  If nothing is checking out, I would recommend opening a Support Case on the issue, so we can look into things further with you.

Q: How the Chunk size impacts/works in QDS?

A: Chunk size is part of the Management Service, not the Distribution Service, but it impacts the ‘Workorder’ information that is passed from the Management Service to the Distribution Service.  Here is some more detailed information I received from R&D regarding this area, which will hopefully help explain things:

The QMS sends Tasks to the QDS where they end up in the following folder: C:\ProgramData\QlikTech\DistributionService\Tasks

QMS and QDS have a common hash value used in communication between the services, when the value differs on either service, QMS resends all tasks to the QDS.  This may also be accomplished by clicking the "Send Workorder Button" in the QMC under

System->Setup->Distribution Services->QDS resource->Advanced tab settings

The QMSChunkSize controls how many tasks are sent to the QDS at a time.

Q: In a cluster server environment. (With 4 nodes)

Let’s say a first user accessed an application and the document is loaded in Node 2.

How do I configure to make sure that the same loaded document is used for the second user also who is coming in and is there a place where i can configure this to happen to all the documents and not do this at a task level?

A: The only really best practice method here would be to set the Load Balancing method on the QVWS instance to use Loaded Document to ensure once the app is loaded on one of the servers, all sessions go to that server.  You can use that in conjunction with the QMC\Documents\User Documents\click on a document after expanding the list\choose the Server tab to the right and then the Performance submenu and use the Document Control settings to dictate which server the app loads upon.  You can also preload larger apps this way too.  The only way to have a mixed environment would be to have two QVWS resources using the same QVServer cluster, but one resource would use Loaded Document and the other CPU with RAM overload, but the problem with this scenario is the need for two web addresses, which is generally problematic for users.  Using a portal to provide a link based upon some criteria could be one method of redirection in that case.

Q: Which scenarios do you recommend the tag ClearCacheTimesPerDay setup in settings.ini?

A: I do not recommend setting any additional settings unless you are directed to do so by one of the Support staff to do so for the particular instance.  Using Settings options without fully understanding why they were created in the first place can lead to assumptions that may prove to be performance issues etc., so I always recommend caution when using non-standard setting options.  This being said, there is an Article on this, number 3061 – How to Clear the Cache Used by QlikView Server.  This option is designed to clear the cache without restart of the QVServer services, but again, unless there is an issue in the environment that is requiring the cache to be cleared ‘manually’, I would recommend letting the QVServer manage things.  Many times an underlying issue is the Working Set Limits of QVServer are what may need to be adjusted and/or Publisher may need to be moved to its own server as well if they are sharing a server and the load is unpredictable.  This is the best guidance I can give on this.

Q: But sometimes, you have many "LASTKNOWNSTATE" for the same userQ: and when there is some "INPUT FIELDS" in the application, the ".shared" file gain many MB?

A: This is working as expected due to the Input Fields, as the input field data must be captured into the bookmark.  You need to be very careful of the functionality in use when determining when to use or not use other features in the product, as there are performance related caveats with many features.  I would reference Article 4042 – Bad Performance Opening a Document… for additional information.

Q: In a Document when I grant authorization to a user by looking them up from AD, it's not working that why I am adding the user manually, any suggestion?

A: I would utilize the Active Directory DSP under the DSC resource in QMC to configure your Active Directory connection, so the DSC reads in all the users and groups.  Once that is done, then in the Authorization tab, you can click the green + button to add a new access rule and then click the User Type dropdown and choose Named Users option, then you will see under the Users and Groups the ability to click the ‘people’ icon and search the directory directly and add users/groups that way.  If you have Publisher, you can do this in the Source Documents area as well when performing a QVServer distribution.

Q: We use some parameters in the URL to set some selections for our users. It seems that when the reload schedule reloads the qvw, this URL parameter is ignored for the current logged-on user. Is there any way around this?

A: One thing I see a lot of companies doing is attempting to automate selections, and this is not really necessary, as any selections that are made and saved in the Source Document are maintained by Publisher after it reloads the app.  Publisher opens the source documents, clears the selections after noting them, then reloads and then reapplies the selections and saves the app, so if making the selections directly in the source document is an option, I would advise the developers to do that before posting the app to the Source Documents folder.  The only way this becomes problematic is if you do Loop and Reduce in Publisher and a result does not have that value available after the reduction.  I hope I got close to what you were asking.  Here is another link by one of the consultants that might potentially be helpful as well to check the URL: http://community.qlik.com/thread/96800.

Q: Can you say more about how to get the overview over where which log files goes where and how to control it?

A: Please check: Video - QlikView Server log files explained - http://community.qlik.com/docs/DOC-3869

You will also find recordings of past TST Sessions in the Support Community like "Troubleshooting failed reload tasks in QlikView Publisher 06/14"

In addition to the above, you can also check Article 4112 – How To Change QlikView Service Logs/AppData to an Alternate Location.

Q: When using Security Matrix and give user (*) it's just getting values exist in security matrix if we have US country and don't have manager for US then will not appear with *

A: This is working exactly as it is designed to do, as this is the meaning of *.  If you wish all values to load, you must ensure the service account or logged in user have ‘blank’ in reduction field, as this will ensure it loads all values etc.  The only other method is to ensure you have an ‘all’ reduction value that ties to all possible values in the data structure, but you will need a link table in that case to link the all value to all the other values etc.  Hopefully this makes a little sense, but the thing most often missed is ensuring the service account has no value set for the reduction fields, as this is the best practice to ensure Publisher completes a full load.  The service account will not be able to open the application via the AccessPoint in this case, as the blank value in the reduction field and ‘strict exclusion’ in the Document Properties will prevent it.

Q: Ideally it’s suggested to get reduced but I am not sure what the reduced numbers actually do...i.e. if it’s set to 50 then what does the 50 do?

A: I assume you are referring to the QMS Chunk Size setting, please refer to the prior question and my answer there, hopefully that will make sense.  The value is basically the number of tasks being passed each communication round, so the underlying issue is finding the value at which things can be transferred without any timeouts etc.

Q: I was wondering if you could show section access in the QMC...because its connected to the Server and we use AD groups to manage access, putting section access into the apps by AD groups for tab/object access is our next step

A: I am not the person to discuss this one, as I do not do a lot of development work.  I would recommend putting up a post on Qlik Community to see if you can find something there, but I can give you some Design Blog references from Henric Cronstom that may be helpful too, and there are many other posts on the Design Blog area as well that may be very helpful to you:

A Primer on Section Access

http://community.qlik.com/blogs/qlikviewdesignblog/2014/05/26/section-access

Data Reduction Using Multiple Fields

http://community.qlik.com/blogs/qlikviewdesignblog/2014/06/02/data-reduction-using-multiple-fields

Data Reduction – Yes, but How?

http://community.qlik.com/blogs/qlikviewdesignblog/2014/06/09/methods-for-data-reduction

Q: So in place of 100 we need to reduce that?

A: Again, I am assuming this is in relation to the QMSChunkSize setting, and it really depends, but what I normally do is try 75 first, if that works, you can leave it there, or if you wish to be more precise, you can adjust back upward say half the distance to 100 and try again etc.  Be sure to leave yourself some room for growth in tasks etc. or you will be right back where you started sooner versus later.  The later SRs of version 11.20 track should not normally need to have this setting adjusted as well.

Q: Can you touch a bit on active directory?

A: Please check - Video - Section Access best practice - http://community.qlik.com/docs/DOC-3871

There is also a recording of TST Session that covers section Access for Qlik Sense in the TST Section

Q: Can old Category names be removed?

A: That is a very good question, and a bit tricky to answer properly as well.  There is no means to do this directly in the QMC, it would have to be done directly in the QVPR, and this is one of those instances I would recommend using the XML DB Viewer power tool, as what you will want to do is find the GUID for the Category(ies) in question and search all tables for that GUID and be sure if the Category is still in use, that the GUIDs in the other tables get repointed to other Category GUIDs.  The trick to avoiding issues in the QMC when modifying the QVPR is to be sure you do not orphan GUID entries in other tables.  What I always do when attempting to clean things up is take a backup of the QVPR before I start and make sure nobody else is going to be in the QMC changing anything, which you can prevent by stopping the QMS as well.  Make your changes, bring things back up and test the QMC to be sure everything is working properly on all tabs and submenus etc., and once you are comfortable things are good, do another QVPR backup to have in case anything happens before the next automated backup occurs.  For those that do not know where this is, it is in the QMC\System\Setup\Management Service\Repository tab and the Backup Now button.  Please note if you do not have a licensed Publisher, the Repository tab will not be there, which would require you to manually back things up etc.