corosync is killed by oom-killer when starting cman in RHEL 6

Issue

• When manually start cman, OOM happens soon
• corosync gets killed by oom-killer when starting cman

Nov  8 17:20:01 localhost kernel: DLM (built Jun 13 2012 18:26:45) installed
Nov  8 17:20:05 localhost kernel: corosync invoked oom-killer: gfp_mask=0x280da, order=0, oom_adj=0, oom_score_adj=0
Nov  8 17:20:05 localhost kernel: corosync cpuset=/ mems_allowed=0
Nov  8 17:20:05 localhost kernel: Pid: 27405, comm: corosync Not tainted 2.6.32-279.el6.x86_64 #1
Nov  8 17:20:05 localhost kernel: Call Trace:
Nov  8 17:20:05 localhost kernel: [<ffffffff810c4971>] ? cpuset_print_task_mems_allowed+0x91/0xb0
Nov  8 17:20:05 localhost kernel: [<ffffffff811170e0>] ? dump_header+0x90/0x1b0
Nov  8 17:20:05 localhost kernel: [<ffffffff810e355e>] ? __delayacct_freepages_end+0x2e/0x30
Nov  8 17:20:05 localhost kernel: [<ffffffff812146fc>] ? security_real_capable_noaudit+0x3c/0x70
Nov  8 17:20:05 localhost kernel: [<ffffffff81117562>] ? oom_kill_process+0x82/0x2a0
Nov  8 17:20:05 localhost kernel: [<ffffffff8111745e>] ? select_bad_process+0x9e/0x120
Nov  8 17:20:05 localhost kernel: [<ffffffff811179a0>] ? out_of_memory+0x220/0x3c0
Nov  8 17:20:05 localhost kernel: [<ffffffff811276be>] ? __alloc_pages_nodemask+0x89e/0x940
Nov  8 17:20:05 localhost kernel: [<ffffffff8115c2da>] ? alloc_pages_vma+0x9a/0x150
Nov  8 17:20:05 localhost kernel: [<ffffffff8113f83b>] ? handle_pte_fault+0x76b/0xb50
Nov  8 17:20:05 localhost kernel: [<ffffffff81048ac7>] ? pte_alloc_one+0x37/0x50
Nov  8 17:20:05 localhost kernel: [<ffffffff811765a9>] ? do_huge_pmd_anonymous_page+0xb9/0x380
Nov  8 17:20:05 localhost kernel: [<ffffffff8113fe04>] ? handle_mm_fault+0x1e4/0x2b0
Nov  8 17:20:05 localhost kernel: [<ffffffff8113fffa>] ? __get_user_pages+0x12a/0x430
Nov  8 17:20:05 localhost kernel: [<ffffffff81141a90>] ? __mlock_vma_pages_range+0xa0/0x1f0
Nov  8 17:20:05 localhost kernel: [<ffffffff81141c78>] ? mlock_vma_pages_range+0x98/0xa0
Nov  8 17:20:05 localhost kernel: [<ffffffff811458d2>] ? mmap_region+0x302/0x590
Nov  8 17:20:05 localhost kernel: [<ffffffff8119579f>] ? destroy_inode+0x2f/0x60
Nov  8 17:20:05 localhost kernel: [<ffffffff81145e9a>] ? do_mmap_pgoff+0x33a/0x380
Nov  8 17:20:05 localhost kernel: [<ffffffff8113574a>] ? sys_mmap_pgoff+0x7a/0x2d0
Nov  8 17:20:05 localhost kernel: [<ffffffff810104e9>] ? sys_mmap+0x29/0x30
Nov  8 17:20:05 localhost kernel: [<ffffffff8100b0f2>] ? system_call_fastpath+0x16/0x1b
Nov  8 17:20:05 localhost kernel: Mem-Info:
Nov  8 17:20:05 localhost kernel: Node 0 DMA per-cpu:
Nov  8 17:20:05 localhost kernel: CPU    0: hi:    0, btch:   1 usd:   0
Nov  8 17:20:05 localhost kernel: CPU    1: hi:    0, btch:   1 usd:   0
...
Nov  8 17:20:05 localhost kernel: Out of memory: Kill process 27405 (corosync) score 345 or sacrifice child
Nov  8 17:20:05 localhost kernel: Killed process 27405, UID 0, (corosync) total-vm:8254276kB, anon-rss:7721036kB, file-rss:14660kB