Hi All,
One of our servers stopped responding , unable to take any logins into it, the response is almost nil...later it resumed
Also during this time one of our application processes which was costly on memory got killed..is it an OOM kill?
Would like to know to know how to avoid such things in future.
Also any logs where to look for how the process got killed etc?
Thanks for your help
With everything that you have told us (a system was slow, a memory-hogging process terminated), you haven't given us any clue as to who or what might have terminated the process.
Did the terminated process have a log file? If so, does it indicate why it terminated?
Was your system running job accounting? If so, does it indicate why the job in question terminated? Does it indicate if anyone ran the kill command close to the time when the job in question terminated?
Who was logged in when the process terminated?
What else was running when the process terminated?
Why was the system running slow?
Also, you can check the syslog or dmesg to see if OOM Killer was invoked. The OOM killer is because of an out of memory condition, which is often caused by an application that uses to much memory, because of a memory leak (bad programming) for example, but there may be other causes..
The log file of the process shows when it was terminated but we want to know who terminated it.
there were many users in the system at that time but none with amdin rights and no fresh logins could be taken. The admin did something Iam not sure what he did.
Iam trying to find out how the process got killed.
dmesg give below output
Feb 22 09:53
Can't read kernel memory
You need to execute dmesg as root. Can you become root.? Otherwise, can you read what is in /var/adm/syslog/syslog.log ?