Business Objects

Set Instance Limits in Business Objects XI R2

In a previous post I mentioned previously How to Manage Instances in Business Objects XI R2, well – that’s only half of the story – rather than having to go and delete old instances yourself it’s far easier to have Business Objects tidy up after itself.  This is one of the few scheduler-related features that can be found in the CMC under “Settings” and “Limits” allows you to specify…

• The maximum number of instances to be maintained (anything above will be deleted). 
• How many instances should be kept for each user/group (I choose Everyone). 
• How many days these instance should be kept for each user/group (again, I choose Everyone). 

Please note that the excess instances do not get deleted immediately after the change, in my case it seemed to happen by the next day so do be patient!  This will really help to cut down the size of your FRS and may disaster recovery much easier. 

View Failed Scheduled Jobs in Business Objects XI R2

When I moved from old-school Business Objects (6.x) to XI Release 2 I was happy that the scheduler would now be available for all users directly in WebI, unfortunately along with the improvement came the down-side – no Broadcast Agent Console (BCA Console) to monitor jobs. The lack of an admin tool for the scheduler becomes a pain every now & again since without very careful control you’ll quickly lose track of which documents have been scheduled by whom and whether they’re still running, if they’re failing, etc. Additionally if you experience some sort of system failure (e.g. FRS goes down) you may need to see what failed in order to make sure it is re-run manually.

Well, despite the fact that the CMC doesn’t include a great deal of schedule management functionality, there is a tool bundled with the standard BO installation which can at least help with the issue. The tool is called the Instance Manager and it’s part of the SDK ‘use case’ examples provided and can be accessed via the Administration Launchpad URL as so…

http://bo_server_name:8080/businessobjects/enterprise115/adminlaunch/launchpad.html

… effectively this redirects you to here…

http://bo_server_name:8080/businessobjects/enterprise115/adminlaunch/instancesByStatus/main.jsp

After selecting the Instance Manager on the left-hand menu (see image above), log in with the relevant credentials…

Select the status you want to look for and a username for which you want to search (username optional for everything but ‘All Statuses’) and hit Go!

You then see a list of instances (note, this is Instances not Jobs)…

Please be careful since this screen will allow you to delete instances (up to 100 at a time) if required and whilst there is a message box asking “are you sure you want to delete…” I don’t know of any way to roll back the changes.

How to Kill Sessions in Business Objects XI (R2 + R3)

There are some areas of functionality that the Business Objects admin tools really lack, one is the ability to view who is currently using the system and then kill a particular session.  These tools can be useful if you’re diagnosing problems and whilst BO don’t provide them the SDK does allow for them to be created and thankfully Johnny Ye has created the Business Objects XI Session Removal Tool, it’s fairly rudimentary in design terms but it does the job perfectly and I’m sure if you know Java and HTML you could customise it easily.

Out of the box it should work with XI R3 but I’ve not tried it since I’m currently running XI R2, in order to get it running on Release 2 you’ll need to modify Johnny’s instructions as follows (instructions taken from this thread on BOB)…

1. In the new_utilities.jsp file change the
line: result += "Delta " +conn.getDelta()+"<br>";
to: result += "Delta " + "na"/*(conn.getDelta()*/+"<br>";

2. Copy the library files from:
~\Tomcat\webapps\businessobjects\enterprise115\adminlaunch\WEB-INF\lib
to:
~\Tomcat\webapps\KillSession\WEB-INF\lib

NOTE: ~ = “C:\Program Files\Business Objects”

3. Restart Apache/Tomcat (I’m using IIS so this didn’t really impact my users).

Business Objects: Change FRS Location without the CMC

Today I had a situation where the one of my file servers went down, the failure was significant and required a rebuild so it was going to be down for a good day or so and whilst most of the data on that server wasn’t regarded as critical for my department it did host the Input and Output File Repository for Business Objects (where all of the reports, etc. are stored).  Obviously we had a backup of the data so I had it restored to another location then went to re-point Business Objects to the new directory, normally this is done using the Central Management Console (CMC) as follows:

Unfortunately, because the original server was unreachable the IFRS and OFRS services had failed and despite my attempts in the Central Configuration Manager (CCM) to restart them they would not come back up.  After putting in a support call with our BO partner one of their consultants found the answer, there is a command-line switch that can be added into the service properties (viewable via the CCM or services.msc) that allows you to over-write the location stored in the CMC which would look like the following…

“\\BusObjServer\C$\Program Files\Business Objects\BusinessObjects Enterprise 11.5\win32_x86\inputfileserver.exe” -rootDir \\NewFileServer\Filestore\Input -service -name Input.BusObjServer -ns BusObjServer -restart -requestport 40020

It was quite a weird problem but that worked a treat.

Business Objects Fails to Retrieve Rows from MySQL Table

I just encountered a very odd situation in Business Objects where a Web Intelligence (WebI) report running against a MySQL v5 database suddenly stopped returning any rows, even though the data had not been deleted.  It turned out that the underlying application had been upgraded and the database schema had undergone a few minor changes but essentially the table structure appeared not to have changed at all. 

After investigating for a few hours I managed to rule out the change in storage engine (from MyISAM to InnoDB) and had managed to narrow the issue down to a single varchar column.  I didn’t have access to the pre-upgrade table scripts so I couldn’t see what had changed but I did notice that in the new table definition the character set of the column was being specified as ASCII like so…

`column` VARCHAR(10) CHARACTER SET ASCII DEFAULT NULL

To fix this I went back to the Universe level and amended my object definition to CAST the column values…

CAST(`column` AS CHAR CHARACTER SET UTF8)

That solved the problem, if you have the same issue – I hope this helps you on your way.