EP: Monthly KM and WPC Update – 11/03/2016
Welcome to the first of my monthly KM and WPC updates. The purpose of the blog is to highlight and review the latest SAP Notes and Knowledge Base Articles related to the KM and Web Page Composer components and to highlight any new commonly reported bugs and coding issues. I’ll also occasionally discuss unusual issues which might have been encountered in the previous month and offer advice on some of the more common configuration issues being reported.
New Notes and KBAs
KM – This area outlines a selection of Notes and KBAs which have been released in the general KM component area in the past couple of months
This KBA outlines an issue with accessing KM folders whereby a system error may be returned. The root cause of the issue is the failure to configure the System Landscape Service to run in clustered mode although it’s a clustered portal installation. Correcting the configuration and performing a system restart resolves the issue.
This note describes an error which may be observed in the trace if the time-based publishing scenario is being implemented in KM. It’s due to a coding issue and a patch fix is required to correct the issue.
2264817 – Some documents cannot be published
Also related to time-based publishing, this note outlines a recently discovered coding issue which may result in documents not becoming published when their ValidFrom property is reached. It only effects resources which are stored in a repository on which the ‘properties’ repository service is not activated. The note not only describes which patches contain the fix but also a workaround which may be implemented.
The third and last of three time-based publishing related notes in this blog update, this issue does not affect the publishing scenario itself but rather the timebased publishing report which may be used to see an overview of all documents on which TBP is active. Instead of returning only documents within the specified ValidFrom and ValidTo properties, the report returns all documents which have lifetime enabled. The fix is provided in a patch.
2265575 – Unable to upload ICE packages in portal
When using Content Exchange service to perform an offline transport, an error may be returned when attempting to upload the ICE package to the destination system. The error is due to a coding issue and is corrected with a patch fix.
This note describes an exception which may be generated in the default.trc logs when using the KM Transport Service (Export/Import). Due to a coding issue, a patch fix to resolve the issue is available.
This cause of this issue actually lies in a series of updates for Java 8. If using a JRE on a version between 1.8.0_60 and 1.8.0_71, a known issue with these JRE versions will prevent the ‘Multiple File Upload’ functionality from working in KM. Instead, the upload process will freeze on ‘Uploading Files. Wait.’. As outlined in the KBA, the fix is included in update 72 and higher for Java 8.
The symptom of this issue is that the Collaboration Room Administration returns a portal runtime error but the cause of the issue is actually outside of KM. As described in this Knowledge Base Article, such behaviour may be encountered due to inconsistencies in the UME. The article also outlines an exception which should be located in the default.trc if this is indeed causing the issue. Running the UME consistency check tool and correcting any inconsistencies should resolve the issue.
This outlines another issue which can be encountered within KM but the cause of which might be difficult to locate due to it being caused by code from another component area. In this case, the issue is a startup error with the application com.sap.netweaver.coll.appl.room. The root cause is actually a UWL coding issue and the fix is included in a patch for the JWF (Java Workflow) component. The KBA describes which note should be followed to apply the fix.
This note describes a coding error with the Deleted Items Quota Check report. An exception (specified in the note) will be returned both in the default.trc and the UI when attempting to run the note. A patch fix has been provided for the issue.
WPC – New Notes and KBAs relating to the Web Page Composer Application
If you restructure a WPC area which contains one or more WPC pages on which regionalization is activated, an error message will be returned when attempting to publish the Area – ‘Exception caught during event invocation’. This is due to a coding error and a patch fix has been released. However, this note also outlines a workaround which may be easily implemented to resolve the issue.
Due to some missing translation, the WPC editor panel in only partially translated. This fix is provided in a patch.
FORUMS/WIKI – New Notes and KBAs relating to the KM Forums and KM WIKI applications
The note describes a patch fix which corrects a security vulnerability which may allow information such as installed products and landscape configuration to be discovered through the KM Forums application.
2279587 – Wiki Spell Checker stops responding
There is an issue which causes the Wiki Spell Checker to freeze when run. The cause of the issue is that when opened, the spell checker popup runs in Standards mode when it should in fact be running in Quirks mode. This results in a coding issue. A fix has been provided to ensure that the spell checker runs in Quirks and the note also outlines a workaround for the issue.
Monthly Hot Topic – Portal Restart
Recently I’ve observed an increase in the number of issues reported which were either being caused by a failure to perform a required system restart, or the required fix for an issue had not taken effect for the same reason. “Have you turned it off and on again?” may be an old cliché but when it comes to Knowledge Management I can’t overstate just how important this step is. Failure to do so may result in changes made within the KM configuration not taking effect or worse, can cause the often encountered ‘Hot-Deployment’ issue to occur.
Hot Deployment is especially important to avoid as the range of issues and symptoms which might be caused by this are erratic and wide ranging. These can include Out Of Memory and locking issues which may cause the portal to crash. If any component or application is deployed on the system (including patch fixes), it is imperative that a full system restart is performed afterwards. Failure to do so will result in Hot Deployment. If it occurs as a result of applying a patch fix, you should note that the fix deployed will not take effect until the restart is performed either.
Hot Deployment can be observed by checking the component monitor (System Administration > Monitoring > Knowledge Management > Component Monitor). SystemInfo will be displayed in red and show ‘startup failed’ as follows –
Clicking on ‘properties’ and selecting the ‘Miscellaneous’ tab will display the exceptions indicating the exact component which is causing the hot-deployment.
The good news is that Hot-Deployment can be resolved by performing yes, you guessed it, a full system restart!
Aside from the risk of Hot-Deployment, there are also a wide range of configuration settings within Knowledge Management which will not take effect until the system is restarted. The KMC starts up during the portal startup sequence and it is at this time that the KM configuration is loaded. It’s not possible to restart the KMC independently of the portal and therefore many changes within the KM configuration require a portal restart before they have any impact. Some of the more common changes which require a system restart are the following –
- Creation and Deletion of repository managers
- Changes made to repository manager configuration
- Changes made to the System Landscape Service
- Changes made to the System Landscape Definitions (eg. the deletion of invalid CM systems)
- The configuration of scheduler tasks including their assignment to specific CM systems
Unlike Hot Deployment, making any of the above changes will not result in errors being caused by the failure to restart the system. However, any changes you do make will not have an effect until the restart has been performed. You might create a new repository manager and then wonder why it’s not showing up in the KM Content – did you restart the system after creating it?
Fortunately, when performing such a change, the system does return a message advising you whether a system restart is required or not. For example, a change to my ‘test_fsdb’ repository will result in the following message –
It should also be noted that in many cases, it’s mistakenly assumed that restarting only the server node on which the change has been performed will be sufficient. This is not the case. Any change which will affect the entire KMC, whether performed only on a specific server node, will require a full portal restart.