Red Hat Satellite 6: oom-killer invocations make the Satellite server inaccessible.

Issue

• While satellite is performing Tasks and system process are consuming huge memory,then the OOM killer terminates the process hence satellite becomes unavailable. Following messages can be observed in /var/log/messages were OOM kills the high memory consuming process.

May 31 01:03:47 hostname1 kernel: [2379341.898683] Free swap  = 0kB
May 31 01:03:47 hostname1 kernel: [2379341.898684] Total swap = 8377604kB
:
May 31 01:03:47 hostname1 kernel: [2379341.898688] [ pid ]   uid  tgid total_vm      rss nr_ptes swapents oom_score_adj name
May 31 01:03:47 hostname1 kernel: [2379341.898830] [14227]   990 14227  5640069  4377444   10006   651556             0 ruby
May 31 01:03:47 hostname1 kernel: [2379341.898762] [31127]   184 31127 28502134     8568    7483    20205             0 mongod
May 31 01:03:47 hostname1 kernel: [2379341.898780] [31563]    91 31563  3049061   139278    1316   432188             0 java
:
May 31 01:03:47 hostname1 kernel: [2379341.899062] Out of memory: Kill process 15227 (ruby) score 612 or sacrifice child
May 31 01:03:47 hostname1 kernel: [2379341.899093] Killed process 15227 (ruby) total-vm:23552276kB, anon-rss:18609776kB, file-rss:0kB, shmem-rss:0kB