Clustering Support for ON Demand Read

On Demand Read, as read only

The default behavior of on demand read is that filebc just read the content of the file and sent the response back to the corresponding component which made the original read request, Note: this is only a read only operation so any number of thread can be reading the same file , i.e filebc can serve any number of request without worrying about synchronizing access to the file , similarly if filebc is cluster , i.e more than one instanse of filebc may be accessing the same the file without any synchronization.

On Demand read with delete

In this option the file is deleted once the file is read, So in this case the access to the file needs to be synchronised both within and accross the process. i.e within the process the filebc ensures that only one thread can have access to the file during the reading of the content, and across the process only one process can access the file at any given point of time

Archive option

This is the default option while using the delete option, In this case the file is archive either in a default location or in the archive dorectory based on the user configuration, with this option user can keep a history of the all the file that has been processed , usually during archiving the filename is appended with a uuid to keep the name unique

Without Archive

User can choose not the archive the prcessed file , in this case the file simply deleted once the file has been processed

This page (revision-2) was last changed on 31-Mar-09 13:42 PM, -0700 by sujit.
This page was created on 31-Mar-09 13:41 PM, -0700 by sujit.

More info...