Reviewing boot logs on Debian

Recently, I installed a debian system as a server from the netinst CD. From the debian website:

A network install or netinst CD is a single CD which enables you to install the entire operating system. This single CD contains just the minimal amount of software to start the installation and fetch the remaining packages over the Internet.

The install was straight forward and i only opted to install standard utilities and nothing else. Everything installed successfully and the system booted without a single problem.

It is good practice to familiarize yourself with what you see on the console when your system boots up. Therefore in the event that something goes wrong and services aren't being started at boot, you would already be familiar with what your system usually does when it boots up. Typically, messages will indicate that your services were started "OK", something has "Failed", "Warning" notifiers, example: Starting periodic scheduler: Cron or some other service has failed. The issue that you will encouter is that these messages that you see on the console scrolls by very quickly and you usually only have about a second to recognize something useful.

By default, my new debian system did not log these boot messages anywhere. It turned out that bootlogd, the daemon that is responsible for logging these messages to "/var/log/boot" was not installed by default. Runnign # aptitude install bootlogd, was all that was needed. Once you reboot your system, boot messages will be logged to the file, "/var/log/boot".

You can simply cat the contents of the log file to view its contents but there is an issue where the colored strings are not escaped and you end up with bash color codes. These colored strings are what you see at boot with strings like "failed", "ok","warn", etc. To fix this you can view the logs with the following command

# sed 's/\^\[/\o33/g;s/\[1G\[/\[27G\[/' /var/log/boot | more

You can set up an alias for this in your .bashrc file to make life easier.

Authenticating an Stunnel server to an Stunnel client

Stunnel can take care of our network encryption needs for any TCP connection. I blogged about setting this up here. Stunnel makes use off SSL certificates which enables us to achieve these 3 main things, 1. Authentication, 2, Confidentiality and 3, Integrity. If you referenced my first post about stunnel, you may or may not have realized that i did not make use of the "authentication" feat of SSL. In this post, i aim to achieve that.

Essentially, the client will be able to connect to the server and verify that it's indeed connected to the correct server. By default the client will communicate with any stunnel server and accept the server's certificate without taking extra steps to verify it. What this means is that a malicious person can potentially setup a rogue stunnel server with their own certificate and pose as a ligitimate server. Because your client accepts everything blindly, it will make the assumption that it is connected to the right server, when indeed it might  not be. Lets fix that.

Step 1. Generate private key, certificate and PEM file.

# openssl genrsa -out server.key 2048
# openssl req -new -x509 -nodes -days 365 -key server.key -out server.crt
# cat server.key server.crt > server.pem
# chmod 640 server.pem
# chmod 640 server.key

Step 2. Run stunnel server.

# stunnel -f -d 443 -r 127.0.0.1:80 -p server.pem -s nobody -g nogroup -P /tmp/stunnel.pid

Step 3. Copy ONLY server.crt file to clients that would be connecting to the stunnel server. The server.pem and server.key file MUST be kept secret at all times.

Step 4. Run client with extra option -v 3 to enable certificate verification via locally installed cert.

# stunnel -c -f  -d 443 -r remote.server.com:443 -v 3 -A server.crt

Thats it.

If a malicious person tries to impersonate your server,it will fail to pass these checks:

1. When the initial connection between client and server is made, the server sends its certificate to the client. The client would only accept this certificate if it is within it's white list of certs "-A option".
2. If for some reason the malicious person was able to have the server send the correct cert, the client would challenge the server to prove that it has the matching private key for the cert. Because the malicious user does not have the private key and is unable to prove that it does, the client will terminiate the session.

Here is a screenshot of what things look like on the client under normal working conditions.

Here is a screenshot of what will happen on the client if a malicious server is used to impersonate a legitimate one.

The way we have setup stunnel we achieve the three main goals of SSL certificates.

1. Authentication. Server authenticates itself to the client before communication is established.
2. Confidentiality. Session is encrypted using symmetric cryptography.
3. Integrity. Ensures that the encrypted data has not been tampered with via a negotiated hashing algorithm.

SSH tunneling made easy with sshuttle.

It amazes me how many things you can do with ssh. They say netcat is the swiss army knfie networking tool, but ssh would be like the swiss army knife networking protocol. We know ssh is mainly used for remote administration via a terminal or command prompt, but it can also be used for other things like the tunneling of TCP connections. If you are familiar with this process, then you know how neat this can be, but configuring each port to be tunneled can get ugly. However, there is a cleaner and simpler way for tunneling your tcp connections over ssh using sshuttle. What you ultimatly get is a poor mans VPN. Once sshuttle is running, all your traffic will be proxied through an ssh connection (dns traffic can also be proxied if told to do so).

Here is an example of how i use sshuttle.

# ./sshuttle --dns -r username@remoteip.com 0/0

Resources / Good Reading:

sshuttle github

Configuration management with puppet

Scenario:

How do you manage the configuration and management of multiple servers (25 plus nodes) in an effective manner?. Say we want to setup the ssh service on 20 servers, configure the service to meet our needs, then reload the service for the changes to take effect. This would involve you doing this manually for each system (can't be done remotely with ssh cause its not installed yet ).What if it was for another service like an ftp server. You can configure one server then scp the config file(s) to each of the 20 machines, then issue the commands to restart the service for the changes to take effect. OK  that works, but its only 20 machines. This clearly doesn't scale up very well when you need to mange hundreds of systems. Clearly there must be a better way. My solution to such a problem is puppet.

Puppet is an open source configuration management written in ruby. It allows you to automate many task across many unix-like systems (windows integration is still in the works) including the installation of packages, configuration of services, creation of users, execution of bash commands and more. It is based on a client-server model. The server (which is called the puppetmaster) will contain the definitions that essentially determines the end state or how things should be on the nodes, for example: what services should be installed, what config files should be downloaded to these nodes, etc. The client (callet the puppet) will poll the server every 30 minutes by default, for these "state" definitions and process them. The client will compare its state with these definitions, and make the necessary corrections where necessary.

I came across some useful tutorials and resources that will get you up on running quickly. The configuration syntax does take some getting used to but if you are like me, you can adapt quite easily after looking at some examples.

Resources / Good reading:

http://puppetlabs.com
http://en.wikipedia.org/wiki/Puppet_(software)
www.linuxjournal.com
http://bitfieldconsulting.com/puppet-tutorial
http://www.debian-administration.org/articles/526

Protecting your services with fail2ban

There is no one solution that takes care of your network security. Protecting your services from the bad guys is quite complex and involved. Lets say we wanted to protect our openssh service for example; there are quite a few things that can be done in order to improve its security.
The first thing you would want to do is ensure that you have the latest version of openssh. You should also setup a mechanism that will check for updates regularly. Next is configuration. Server software can be quite complex to configure securly and openssh is no exception. Once your configs are all polished, there is still one attack vector that still could impact us, bruteforce attacks. Unfortunately, openssh doesn't have a feature that we can configure or enable that prevents bruteforce attacks. However there is a project, fail2ban, that can protect us from such attacks.
Fail2ban has been around for almost 5 years and simply works. It works by monitoring log files (in openssh's case, the auth.log file found on debian bases systems in /var/log/) for signs of failed login attempts. When it detects a number of failed attempts for a given duration of time, it can block the offending IP address using iptables and also send an email to the system admin. Fail2ban doesn't just work for openssh but can be configured to work for any service that logs fail logging attempts. See the links below for examples and details for configuring fail2ban for your services.

Resources / Good reading:

fail2ban.org
https://help.ubuntu.com/community/Fail2ban  
http://www.howtoforge.com/fail2ban_debian_etch
http://www.the-art-of-web.com/system/fail2ban/

Raid and High-Availability linux solutions

I've been doing a lot of research lately on Linux solutions on Raid for custom/home built NAS solutions, and high availabilty for networked services. As usual, Linux has some of the best solutions on everything. In my research, i was able to test different solutions and technologies; btrfs, mdadm with LVM, glusterFS, drbd and heartbeat. I won't be exlpaining how to get up and running with these systems but will leave some of the references that i used to get up and going. In the end, i learned alot about raid in general and found mdadm with LVM to be the best solution for me. Btrfs was excellent in my tests but is still somewhat experimental and lack raid 5 support currently :(. Lets briefly get acquainted with these individual software systems.

Btrfs is an advanced filesystem that enables you to aggregate disks and partitions into usable raid arrays that can be quickly and easily mounted. The filesystem is capable of error detection and correction through check sums, snapshots and sub volume management for volume/storage management (as with LVM). As of today, btrfs is still in development and not considered fully stable. Here are some references to get you up and running. Link1, Link2.

Drbd is a mirroring technology that has the ability to sync two disks or partitions from two distinct machines. This is useful for setting up a backup/secondary server that should each have identical data. This is essentially raid 1 over IP. Here are some resources. Link1, Link2

Heartbeat is a high-availability solution that allows for system fail-over. What this means is that heartbeat will provide access to the resources on machine A and in the event that machine A goes down, it will notify machine B to wake up and take over the duties of machine A. This transfer of resources is seamless in most cases and there may be only a 2-6 seconds downtime depending on how things are configured. Here are some resources. Link1, Link2

GlusterFS is a rather unique filesystem that enables a network of machines to combine there storage resources (disks, partitions, folders) and add raid like capabilities with these storage devices. This is similar to drbd, however, i find it to be more flexible. While drbd operates at a lower level (block level), glusterFS is at the filesystem level and operates on the files. You can mirror folders to multiple peer machines (raid 1 over the network), stripe data (raid 0 over the network), combine the storage capacities of all peers and more. Resources: Link1, Link2

Mdadm is the traditional Linux software raid solution. Its been tried and tested and is still the goto solution amongst linux sys-admins. It has support for many raid types including, raid-0, raid-1, raid-5, raid-10 and more. It is common to use LVM for volume management on top of the mdadm raid solution. LVM allows for control and management of the storage pool, which makes it easy to grow or shrink volumes (which are like partitions on a hard drive), take snapshots, etc. Although I loved using btrfs, as it works great for a raid solution, it did lack raid 5 support at the time (mdadm has raid-5 support. Support for raid-5 is due out in a later kernel release for btrf). I found the combination of mdadm and lvm more involved in its setup than btrfs but still relatively easy. I spent quite sometime simulating failed disks in both raid 1 and 5. I was playing a video from the local system when i simulated a disk failure on one of the disks. While replacing the failed disk (still, we are simulating) and rebuilding the array, the video kept on playing with no lag/downtime.

Mdadm Resources: Link1, Link2, Link3
LVM resources: Link1, Link2, Link3

System Monitoring for home users.

I recently covered a system monitoring tool called icinga. Its features and capabilities are very powerful but may be overkill for a home network. In this post, i will show an alternate setup which will allow you to pull system information from networked linux machines and send then back to a central server.

Here is an overview of the pieces that make this work. We will connect to client machines over ssh, execute a local script that will retrieve system information and then send this information back to us on a custom port using netcat. I would also use a utility called expect to aid in automating everything. You can find out much more about expect from google or watch this excellent tutorial on hak5.

The script file will need to be copied to each client machine. Here are the contents of the script. Give this script file executable permissions with the chmod command. Its a modified version from this:

Filename would be sysstat.sh

#!/bin/bash
CPUTIME=$(ps -eo pcpu | awk 'NR>1' | awk '{tot=tot+$1} END {print tot}')
CPUCORES=$(cat /proc/cpuinfo | grep -c processor)
echo "
System Summary (collected on `date`)

 - CPU Cores             = `echo $CPUCORES`
 - CPU Usage (average)       = `echo $CPUTIME / $CPUCORES | bc`%
 - Memory free (real)        = `free -m | head -n 2 | tail -n 1 | awk {'print $4'}` Mb
 - Memory free (cache)       = `free -m | head -n 3 | tail -n 1 | awk {'print $3'}` Mb
 - Swap in use               = `free -m | tail -n 1 | awk {'print $3'}` Mb
 - System Uptime             = `uptime`
 - Local IPs             = `ifconfig | grep -B1 "inet addr" | awk '$1 == "inet"{ print $2}'|awk -F: '{print $2}' |grep -v "127.0.0.1"
`
 - Public IP                 = `dig +short myip.opendns.com @resolver1.opendns.com`
 - Disk Space Used           = `df -h|awk '$6 ~ /\/$/ {print $1 ": percentage used: " $5 " out of " $2 " total on " $6}'
`
################################################################
"

The expect script file's contents are as below. Give this file executable permissions afterwords with chmod:

 Filename would be ssh.exp

#!/usr/bin/expect

spawn ssh root@127.0.0.1

expect "?assword"
send "test\n"
expect "root@"
send "sleep 5\n"
expect "root@"
send "./sysstat.sh|nc -q 1 127.0.0.1 4444\n"
expect "root@"
send "echo $?\n"

expect {
"0" {send "echo 'Success !!!' \n"}
"1" {send "echo 'Something went wrong !!!' \n"}
}
send "exit\n"
interact

We will setup a netcat listener on our machine that will receive the client system information. I used ncat, a similar utility to netcat but has the -k option that will allow us to accept multiple connections instead of one.

ncat -klvp 1234

 Now, with the ncat program listening for connections, we only need to run the expect script file. This script will initiate an ssh connection with the remote client system, logs you in with the proper password, sleeps for 5 seconds , execute the sysstat.sh shell script we created, which will gather information about the system (like CPU load, RAM usage, Hard disk space, etc) and output the results to netcat. Netcat will then send this information to our ncat listener. The expect script then exits the ssh session and finishes.

Here is a sample of the output you can expect from our ncat listener|:

System Summary (collected on Tue Jul 24 18:43:53 EDT 2012)

- CPU Cores = 2
- CPU Usage (average) = 17%
- Memory free (real) = 477 Mb
- Memory free (cache) = 221 Mb
- Swap in use = 0 Mb
- System Uptime = 18:43:53 up 43 min, 4 users, load average: 0.07, 0.09, 0.12
- Local IPs = 192.168.2.15
- Public IP = 123.45.678.90
- Disk Space Used = /dev/sda5: percentage used: 41% out of 39G total on $6
################################################################

Resources / Good reading:
question-defense.com
hak5.org: expect tutorial
thegeekstuff.com