On Wed, Jan 4, 2012 at 6:25 PM, Miroslav Grepl <mgrepl(a)redhat.com> wrote:
On 01/04/2012 12:31 PM, Nabeel Moidu wrote:
Hi
I'm trying to create an SELinux policy for an rpm software installation.
I've been getting sealerts in the var/log/messages but I am unable to view
them due to this error,
*[root@nmk-centos-60-1 policy]# sealert -l
6a6e02bc-23a7-4e55-adab-b06d0cdc2832
Error
query_alerts error (1003): id (6a6e02bc-23a7-4e55-adab-b06d0cdc2832) not
found
*
The problem is the alert has been already deleted from
setroubleshoot_database.xml.
Is there a timeframe for the xml overwrites ?
I believe this has to do with the setroubleshoot daemon not running.
setroubleshoot is DBus service in RHEL6.
OK. That explains it.
*
[root@nmk-centos-60-1 policy]# service setroubleshoot status
setroubleshoot: unrecognized service
[root@nmk-centos-60-1 policy]# service --status-all | grep setro*
I have the setroubleshoot softwares installed
*
[root@nmk-centos-60-1 policy]# rpm -qa | grep setroubles
92:setroubleshoot-server-3.0.38-2.1.el6.x86_64
425:setroubleshoot-plugins-3.0.16-1.el6.noarch
426:setroubleshoot-3.0.38-2.1.el6.x86_64
587:setroubleshoot-doc-3.0.38-2.1.el6.x86_64
[root@nmk-centos-60-1 policy]#
*
I don't see the setroubleshoot rpms creating any init script file in
init.d or elsewhere.
*
[root@nmk-centos-60-1 policy]# rpm -qa --list setroubleshoot-server |
grep -v ^/usr
1:/etc/audisp/plugins.d/sedispatch.conf
2:/etc/dbus-1/system.d/org.fedoraproject.SetroubleshootFixit.conf
3:/etc/dbus-1/system.d/org.fedoraproject.Setroubleshootd.conf
4:/etc/logrotate.d/setroubleshoot
5:/etc/setroubleshoot
6:/etc/setroubleshoot/setroubleshoot.conf
172:/var/lib/setroubleshoot
173:/var/lib/setroubleshoot/email_alert_recipients
174:/var/lib/setroubleshoot/setroubleshoot_database.xml
175:/var/log/setroubleshoot
176:/var/run/setroubleshoot
*SELinux is running in permissive mode with mls type on my system.*
[root@nmk-centos-60-1 policy]# sestatus
SELinux status: enabled
SELinuxfs mount: /selinux
Current mode: permissive
Mode from config file: permissive
Policy version: 24
Policy from config file: mls
*I am running Centos 6.0*
[root@nmk-centos-60-1 policy]# cat /etc/issue
CentOS Linux release 6.0 (Final)
Kernel \r on an \m
[root@nmk-centos-60-1 policy]# uname -a
Linux nmk-centos-60-1 2.6.32-71.el6.x86_64 #1 SMP Fri May 20 03:51:51 BST
2011 x86_64 x86_64 x86_64 GNU/Linux
[root@nmk-centos-60-1 policy]#
*
1) Did I miss anything with regards to the troubleshooting daemon
installation ?
2) How can I fix the query alert error and view the sealert output ?
I see that you use MLS policy. I would suggest you to use ausearch tool
rather than setroubleshoot in MLS policy.
I wanted to formulate the rules for a custom rpm. When using the targeted
policy, I
could not see any denials. So I switched to MLS to identify the
AVC denials. My approach is to log the AVC denials during rpm installation,
and apply the audit2allow on those denials and formulate the policy. Is
this workable ?
The policies for running the software can be different and I plan to have
that as a second stage. I just want to have the installation part getting
on fine with a targeted policy.
Another question, is MLS a namechange for the "strict" type used earlier.
Any links that explains the difference ?
For example:
$ ausearch -m avc -ts recent
$ ausearch -m avc -ts today
$ ausearch -m avc -su testdomain_t
This works, but I wanted to read the descriptive text about the denials
that shows
up in sealert.
All AVC msgs are located in /var/log/audit/audit.log.
Nabeel
--
selinux mailing
listselinux@lists.fedoraproject.orghttps://admin.fedoraproject.org/mailman/listinfo/selinux
--
Thanks and Regards
Nabeel Moidu