Need some odd Linux Client behavior explained

Usage and Config Patch
1

Our Linux support group used IEM to patch the glibc bug this weekend. They used a Baseline to target Redhat 5 & 6 systems, ~250 of them.

They configured the action to remain ‘active’ for 5 hours, from 1/31/2015@11PM until 2/1/2015@4AM.
The configured a Post Action restart, but they inadvertently left the Post-Action reboot deadline at 1 day.
Some percentage (as yet undetermined) of the machines are now refusing SSH sessions with the following message …

The system is going down on Fri Jan 30 23:06:09 2015

IBM Endpoint Manager Restart (Force count:2) from ActionID xxxxxx

Connection closed by nnn.nnn.nnn.nnn

This message was generated circa 2/2/2015 @ 8:37am when the Unix admins started to check the status of their systems. Given the fact that the Reboot Deadline extended beyond the Action Deadline, is this ‘expected’ behavior?

All the services on the machines are function, but nobody can SSH into a large number of the systems. Restarting the systems resolves the problem, but simply cycling the BESClient service doesn’t.

2

If the client picked up the task towards the tail-end of the window, the task runs to completion regardless of the window that is defined. It won’t stop in the middle of its assigned tasks if it breaches the other side of the window.

I know from a previous interaction with IBM that there is sometimes a bug with the BESClient and RHEL5 where if you use the restart function on the post-action tab, sometimes the clients will create the “nologin” file and, after rebooting, don’t clear that file. Create an analysis with this relevance to determine if it exists.

exists file "nologin" of folder "/etc"

Next, create an action to remove it (written in ActionScript but you can use your script of choice):

delete "/etc/nologin"

Here is a link describing the problem, an action to remove it on boot using the cron, and a comment giving further explanation. I guess it’s a bug with RHEL5.

http://bigfix.me/fixlet/details/3870

3

Is anyone else still experiencing issues like this on RedHat/OEL 5 ?