1

My process gets killed on linux server without manual intervention. I have verified the following scenarios.

  • No manual intervention done to kill either by user or admin
  • RAM and SWAP are not exhausted
  • 'strace' gives me the message that "+++ killed by SIGKILL +++" - This makes me to assume that Kernel has killed the process.

The process occupies 16.5GB of virtual memory where the RAM is 16GB and the SWAP being 50GB.

Tasks: 393 total, 2 running, 387 sleeping, 4 stopped, 0 zombie

Cpu(s): 12.8%us, 0.5%sy, 0.0%ni, 86.7%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st

Mem: 16015M total, 8700M used, 7315M free, 202M buffers

Swap: 51199M total, 150M used, 51048M free, 5906M cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND

31466 xyz 20 0 16.4g 1.3g 6704 R 100 8.3 0:16.99 test*

Could you please tell me the other reasons for Kernel to kill the process..

Sobhan
  • 21
  • 1
  • 10

1 Answers1

1

It is likely that kernel out-of-memory (OOM) killer killed your process. You need check dmesg output to see what was the exact reason for the kernel to kill your process.

Maxim Egorushkin
  • 131,725
  • 17
  • 180
  • 271
  • Maxim, I have examined dmesg output while the process is being killed. But no info related to killing of process could be seen from it. Is the procedure I have followed is correct ? – Sobhan Jun 23 '14 at 13:52
  • @Sobhan Depending on the distribution, old kernel log messages may be archived, so you need to check those archives. – Maxim Egorushkin Jun 23 '14 at 14:26
  • I couldn't observe any such messages in /var/log of the machine – Sobhan Jun 24 '14 at 08:36