Hi All,
In this post i'll discuss the role played by web server and batch server while transferring the reports to report repository specified. I'll be assuming http/https transfer throughout the post.xcopy/ftp i'll discuss in another post.
I find in many documentation that distribution agent process in process scheduler server transfers the output files generated by a process to report repository and deletes the file from log/output directory.
Well i find this a very rough idea as it fails to mention the role played by web server. In fact files are posted to repository and removed from log/output directory completely by web server !!!!!
Did it ever occur to you where is the repository path picked from ? Consider a situation, i have two batch servers configured on two different machines. and a web server running on third machine .Now the report repository path given in web profile is c:\psreports.
Now if transferring the files was the job of process schedulers, they will read the repository path as local directory since nowhere else do we specify the server address and they should post the files on the respective physical server those are running .
But this does not happen, reports are posted on the local drive where web server is running. Now lets understand the process how it happens :
On web servers two servlets, SchedulerTransfer and psreports owning the responsibility to transfer the reports and to view the reports from front end. All distribution agent does is send a request to web server with the details of processes for whom reports are to be sent. Report node setting plays a crucial role in this.
Report Node Settings :
In report node, URI host URI Port is the address of web server where SchedulerTransfer servlet is present .in URI Resource we give SchedulerTransfer/<site name>.Site name is used to pick the web profile it uses and from there report repository path is taken (if nothing given then configuration.properties it take value from). A request to transfer the files is sent to this servlet on this web server configured which picks the repository path from web profile/configuration.properties and transfers the files.
psreports servlet is used to view the reports from front end which we mention in the URL field of report node .Again here site name is mentioned as a hint to tell the repository path from where report has to be picked to view .
Understood this concept, you can setup two report repository such that output of process run by one process scheduler are posted on one server and reports of processes run by other batch server are posted on another server.However, reauthentication is required in that case. check my other post on re authentication .
Try these settings !!!!
In this post i'll discuss the role played by web server and batch server while transferring the reports to report repository specified. I'll be assuming http/https transfer throughout the post.xcopy/ftp i'll discuss in another post.
I find in many documentation that distribution agent process in process scheduler server transfers the output files generated by a process to report repository and deletes the file from log/output directory.
Well i find this a very rough idea as it fails to mention the role played by web server. In fact files are posted to repository and removed from log/output directory completely by web server !!!!!
Did it ever occur to you where is the repository path picked from ? Consider a situation, i have two batch servers configured on two different machines. and a web server running on third machine .Now the report repository path given in web profile is c:\psreports.
Now if transferring the files was the job of process schedulers, they will read the repository path as local directory since nowhere else do we specify the server address and they should post the files on the respective physical server those are running .
But this does not happen, reports are posted on the local drive where web server is running. Now lets understand the process how it happens :
On web servers two servlets, SchedulerTransfer and psreports owning the responsibility to transfer the reports and to view the reports from front end. All distribution agent does is send a request to web server with the details of processes for whom reports are to be sent. Report node setting plays a crucial role in this.
Report Node Settings :
In report node, URI host URI Port is the address of web server where SchedulerTransfer servlet is present .in URI Resource we give SchedulerTransfer/<site name>.Site name is used to pick the web profile it uses and from there report repository path is taken (if nothing given then configuration.properties it take value from). A request to transfer the files is sent to this servlet on this web server configured which picks the repository path from web profile/configuration.properties and transfers the files.
psreports servlet is used to view the reports from front end which we mention in the URL field of report node .Again here site name is mentioned as a hint to tell the repository path from where report has to be picked to view .
Understood this concept, you can setup two report repository such that output of process run by one process scheduler are posted on one server and reports of processes run by other batch server are posted on another server.However, reauthentication is required in that case. check my other post on re authentication .
Try these settings !!!!
No comments:
Post a Comment