Hi all,

I have some questions for you about RHEL 3.0 security. I have already found many documents about general UNIX security, but some points are still obscure.

Please take time to try to help me, I would be very grateful. Don't hesitate to explain things even if they seem obvious to you, my knowledge in UNIX's world being very limited...


1) How can we close any connection after a certain period of inactivity ?

2) How can I limit the amount of information given to the user when he is establishing a connection to a server ? (typically not giving the version of OS, etc.)

3) How to make a password respect a predefined policy ? (complexity, aging). I have seen the PAM cracklib module, but is it possible to do the same things without PAM ?

6) General question : if you had to make a RHEL 3 system very secure, which services would you allow and which ones would you make unavailable ?


Thank you for your help. Please answer even if you only know the answer to one question. It will help. And forgive me for my english (I am french).

Bye,

Aaclof

All of you r questions have many different answers depending on configuration and how the server is going to be used.

1). Most all of your firewall devices are going to timeout connections after an idle timeout of roughly 30 minutes. The default idle timeout on RHEL is 2hrs so the firewall times out way before linux gets involved. Here is some info from the kernel source tree.

tcp_keepalive_time - INTEGER
How often TCP sends out keepalive messages when keepalive is enabled.
Default: 2hours.

tcp_keepalive_probes - INTEGER
How many keepalive probes TCP sends out, until it decides that the
connection is broken. Default value: 9.

tcp_keepalive_intvl - INTEGER
How frequently the probes are send out. Multiplied by
tcp_keepalive_probes it is time to kill not responding connection,
after probes started. Default value: 75sec i.e. connection
will be aborted after ~11 minutes of retries.


2) How can I limit the amount of information given to the user when he is establishing a connection to a server ? (typically not giving the version of OS, etc.)

This is usually configurable within the daemon that they are connecting to.

3) How to make a password respect a predefined policy ? (complexity, aging). I have seen the PAM cracklib module, but is it possible to do the same things without PAM ?

Have a look at /etc/login.defs

6) maybe should be 4? General question : if you had to make a RHEL 3 system very secure, which services would you allow and which ones would you make unavailable ?

Only the services you need to accomplish the task the server was setup to do ie if you are running a webserver then you shouldn't have bind (dns server) running as well. Also only install the minimum package set that you need. netstat -anp should only have the services that you need listening for connections.

I am not sure why you would want to have users authenticate without using PAM. PAM is a great system for Linux. Care to elaborate?

Well, it is not that I don't want to use PAM. I just wanted to know whether it was possible to authenticate users without it or not.

If you tell me the answer is no, it's fine !!

I am quite sure you can, though I don't think I would want to. With PAM, you can setup different authentication types and complexities including secure ID, LDAP, and such. I was just curious if you had some motivation not to use PAM.