It looks like under some instances, when a user logs in
via ssh, sshd starts spanning bash processes in an infinite
loop. The whole system will be blocked until the kernel runs
out of memory and kills sshd:
kernel: VM: killing process sshd
After that everything is fine again.
What's the users .bash* files look like?
.bash_profile = RH default
# Source global definitions
if [ -f /etc/bashrc ]; then
# User specific aliases and functions
Can you be more specific about "some instances"? Does it happen regularly for
some users? for all users? Does adding "set -x" to the top of the .bashrc file
shed any light on this?
Never occured again.