Hi, You can notify users when you concurrent managers gets crashed for various reasons and it all depends on you , to how to notify users . When Conc Manager is crashed all the running requests will fail as completed Error . You can identify with this . By default any concurrent program you define , there is a option called *Restart on System* *Failure* this will take care of restarting failed concurrent programs. Once you start all the services , it automatically pics the concurrent programs to run , if this option is checked (Restart on System Failure ). Regards Bala On Sun, Oct 4, 2009 at 10:03 PM, Kumar Madduri <ksmadduri@xxxxxxxxx> wrote: > Hi All > Let us database crashed because the node was unresponsive or for some other > reason. In Oracle Apps, concurrent manager also crashes . In my > understanding, when the database is back up, and you bring up the concurrent > manager the jobs that were running at the time of the crash should be able > to recover and continue without any data corruption (unless you have custom > programs that are commiting parent tables without inserting data in to child > tables. Even these should be fine theoritically because > fnd_concurrrent_Requests should maintain the status of the jobs and continue > from where it left and I have not checked but most Oracle concurrent > programs (given by Oracle) would have the logic of checking which tables are > updated/deleted/instered an in what order ). > So my question is should we not just bring up the concurrent manager after > a crash and not have to worry about which programs were running just before > the crash and notify users that such and such program was running before the > crash.. > Since fnd_concurrent_requests maintain the status of the jobs , the > concurrent manager should be able to pick up the jobs that were in this > table and their statuses and we should be good to go. > > What is your opinion on this? > > Thank you > Kumar >