We have a Business Objects XI Cluster of 2 nodes. Note that we only use BOE as a host for Crystal Reports. In our use case users of our own custom web application access reports through BOE and then a CrystalReportsViewer is populated in the browser to interact with the crystal report.
The issue we’re having is that when a user runs a reports every other function of the report fails with a “Document could not be opened” message. Example of this would be paging - one click will fail then the next click will go to page 2. This pattern repeats indefinitely.
The theory here that seems to line up is that the persistent sessions we have configured at the load balancer level are not working on some level since it seems like the secondary BOE report is trying to fulfill each intermittent request but does not have any session information in order to do so. The load balancer should be keeping the session on the node it was originally given to.
Any tips? I have been fiddling with this for over a week and I feel like I have scoured the entire internet for tips or answers but nothing seems to resolve this issue
It’s been a while since I worked with a XI 3.1 system but you might want to look at your Caching configuration. By default the Crystal Reports Cache Server will cache locally on each server. If you change your Cache Files Directory to a shared location, both Crystal Reports Cache Servers will have access to it. My theory is that the Crystal Reports Viewer is going back to the server cache to get the additional page views and it may not be cached on the second server.
Also, make sure that the option to Share Report Data Between Clients is checked. (I can’t remember if that option was available in XI3.1.) This setting should be available for both the Crystal Reports Cache Server and the Crystal Reports Processing Server.
The load balancer persistent sessions only refer to which server you login is connected to, it doesn’t have anything to do with which Crystal Reports Cache Server the Central Management Console decides to send the request for the report cache to.
For reference, I am looking at a BI4.3sp2 environment so the names could be a little different.
Oh, and welcome to BB
Thank you so much for the reply!!
I’m looking at the 2 CrystalReportsCacheServers and they DO have “Share Report Data Between Clients” however the “Cache Files Directory” is “%DefaultDataDir%/CrystalReportsCachingServer/temp” which is the default value.
The part that’s getting me is I have no idea how to view or change the value of the %DefaultDataDir% variable? I’m embarassed it’s that difficult to track down.
We have a working environment in prod where everything is working fine and one of my TS resources said it’s not a system environment variable there
If you want to try changing the Cache Files Directory, you can just hard code a location there using Universal Naming Conventions, for example \someserver\directory. You don’t need to update %DefaultDataDir%.
It’s not uncommon to be frustrated with Business Objects when things work in one environment and not another. It’s rather finicky. At least your Production environment is working fine.
This “new” environment is intended to be our new prod environment (new servers built on a new version of Windows Server OS) but we can’t cut over until we get things working
interestingly - apparently our prod servers that ARE working have the same value for Caching Server - the D:\Program Files… value
I performed the update to the Cache Files Directory (to a location on same server as our IFR/OFR) and it didn’t seem to change the behavior - which in a very small way I’m happy about as I would ideally like our new environment to work identically to the currently-working-prod environment
one thought/question i had - the various “servers” within BOE - are they all executed under the credentials of the SIA “log on as” user?
The reason I’m asking is I wonder if maybe there’s some kind of permissions issue with accessing the IFS/OFS by the secondary node servers?
In the SIAs - the same domain account is set as “log on as” for both cluster nodes
Yes, all of the servers run under the credentials configured for the SIA.
Is your FileStore (IFS/OFS) located in a shared location and the Input and Output File Repository Server configured the same on both nodes? That could cause this issue as well.
yes indeed the IFS/OFS are both set and configured identically to a shared file server on the network