Hi "yarn.resourcemanager.am.max-attempts" is at global level configuration for all the app master. And it is maximum application attempts by any application can launch AM.
Similarly, individual application can specify max-attempts using ApplicaitonSubmissionContext#setMaxAppAttempts while submitting application and this value should be in the range [1,globalMaxAttempts]. This works in HA scenario also. In your case, RM is able to restart the attempt but it is not able to launch AM and dies immediately. For finding reason, you need to look into application master console log for why AM launch is failing? Thanks & Regards Rohith Sharma K S -----Original Message----- From: Henry Saputra [mailto:henry.sapu...@gmail.com] Sent: 04 June 2015 02:31 To: yarn-dev@hadoop.apache.org Subject: Question about effect of yarn.resourcemanager.am.max-attempts config for already running ApplicationMaster Hi All, I would like to know if "yarn.resourcemanager.am.max-attempts" config parameter will make the already running ApplicationMaster (AM) to have HA mode in YARN once it is already running? Meaning that if the running AM process dies (though permgen, OOM, or kill JVM with kill signal) then ResourceManager (RM) should be able to restart the number of times specified by "yarn.resourcemanager.am.max-attempts" config value ? I was trying it and it seems like the there was an attempt to restart the AppMaster but dies immediately. Thanks, Henry