On Feb 1, 2008 12:47 PM, Niki Kovacs contact@kikinovak.net wrote:
Les Bell a écrit :
Policy. It's a drag, writing policies, but without policies, you're in the "Ready! Fire! Aim!" school of security. The top tier of policy is the "Enterprise Security Policy", which establishes the security function, roles, responsibilities, budget, etc. It also gives the power to enforce penalties for breaches of policies. At the next tier, you have system- and issue-specific policies, such as the "Use of corporate email" policy, the "Inappropriate content in the workplace" policy. You may then move down to standards (platforms, SOE, etc.) and procedures (e.g. for provisioning user accounts, resetting passwords, etc.).
<snip>
Thanks for your very detailed response. Though I can't help feeling a bit like having asked for an identity photo... and getting a 10-foot oil painting :oD
Basically, all I'm concerned about security-wise is a modest Apache/PHP/MySQL server running a single public library management software, and interconnecting eleven (small) public libraries, with a total of 60.000 database entries. No (very) big deal.
The configuration is supposed to run on a dedicated server, so my question will be more practical:
- Is it worth the hassle to bother with SELinux?
Must be your last concern. Use permissive. If you have time switch to enforcing at release time.
- Is the standard firewall configuration enough, or do I really have to
fine-tune the thing?
The problem is not the tools, It is its usage, and its user here. Drugs can heal, but can kill too! Yes this is a good start, but try toi understand what you are doing. But the best is to put a cheap router/firwall in front of your server and forward _only_ the required ports. Dont give your server a public IP.
- Basically, what auditing tools besides NMap can you recommend for such
a thing?
nmap :-)
cheers,
Niki
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos