Subject: vxW errors
From: "Roderick A. Anderson" <raanders@acm.org>
Date: Thu, 08 May 2008 09:48:10 -0700

System: CentOS 5 with all current updates from the CentOS 
{base,update,etc}, dhozac-vserver, and rpmforge repositories.

I have been noticing the following type of errors in the host log files 
coming from several guests.

May  8 09:01:01 admin kernel: vxW: [�crond�,6875:#1195|1195|1195] 
messing with the procfs.

1195 is the context for a guest but I see similar messages for other guests.

Researching, I see it is happening when crond runs the entires in 
/etc/crontab that call run-parts and where the corresponding 
cron.{hourly,daily,weekly,monthly} directory has no entries.

Looking a bit more I see entires in the guest's /var/log/cron file like 
this:

May  8 09:01:01 rptas crond[6874]: Cannot make/remove an entry for the 
specified session
May  8 09:01:01 rptas crond[6874]: CRON (root) ERROR: failed to open PAM 
security session: Success
May  8 09:01:01 rptas crond[6874]: CRON (root) ERROR: cannot set ecurity 
context

Each of the files in the hosts /etc/pam.d directory have had the 
following sed script applied to them.

sed --in-place -e  "s/^session.*required.*pam_loginuid.so/# 
session\trequired\tpam_loginuid.so/g" /etc/pam.d/*

This came from one of the install HOWTOs and as I reflect on it now I 
wonder what it really is accomplishing.   I know I've had problems with 
getting a ssh login to a guest before applying this but do I need to do 
it in the guests also?

Any clues as to why I'm getting the entires?  And possibly the fix?


Thanks,
Rod
--