The Linux Router Project
By design, LRP is meant for low-level networking applications. To this extent, the stock Linux 2.0 LRP kernel is compiled with almost all networking options and several enhancement patches (for example, ipportfw). Next is a short list of default functionality available in the base. Understand that these features apply to any interface type Linux can use, from modem to wireless T1.
Full IP routing with numberless IP and multiple IP interfacing
Complete firewalling and IP accounting
Traffic shaping (low-speed links)
IP and Port (TCP/UDP) redirection with transparent proxy and IP load balancing
Extended Common NAT (IP Masquerade) as firewall rules
Media insensitive interface load balancing (EQL)
RIP, OSPF, BGP and other routing protocols (via gated and other packages)
IPX and AppleTalk support
Past these core functions, a Linux Router unit is quite viable for many server applications. One of the most popular is using the Portslave RADIUS client software (pslave.lrp) to host inbound Internet access modem pools, which give you functionality like the popular Livingston Portmaster series. Bind is packaged (bind.lrp) and does a fine job as a secondary or even primary DNS. Boa (boa.lrp) and thttpd (thttpd.lrp) are both small web servers that will fit on a 1.44MB LRP disk. Apache, though not packaged right now, could be hand fit.
On the client side of things, the VNC package will give you a single-disk Windows terminal. XFree86 is not yet packaged, but it certainly will be, allowing easy creation of X terminals. Linux Router also makes a good base for workstations that mount their file system from remote NFS or SMB (Samba) servers.
Understand that Linux Router is not the solution to all your computing needs. You wouldn't want to use it as your primary web server if you are an Internet Service Provider, or as a kernel hacker's development machine. But for backup services or where feature requirements are very specific (and administration skills scarce), it makes an efficient and sound choice.
While a “minimal” install of Debian Linux may have well over 5000 files, you would be lucky to break 500 with a very feature-rich LRP. The impact of this on administration and security is obvious; the entire system can be backed up and restored in less than a minute.
This minimal footprint coupled with RAM disk operation adds up to a large performance increase in operation and decrease in required hardware. The most mediocre 16MHz 386SX with 8MB of RAM, a 1.44MB floppy and no video card can handle most people's needs for light routing jobs. Few will have a need to climb past a 486 class CPU; however, with low-end Pentium class machines being so inexpensive, it certainly makes sense to do so.
Furthermore, with this default RAM disk approach, it is very difficult to be left with an unusable system. File system get corrupted? Flash the power. Everything just crashed? Flash the power. Cracker break in? Install your backup (in 30 seconds) and flash the power.
The other advantage of running from a RAM disk that people seem to overlook is, generally, it cannot break. Let's say you have 250 workstations with one path to the outside world. Do you want that path to depend on a mechanical device like a hard drive? Probably not. How about if that router sits between you and your game of Quake? Definitely not.
Some of the many reasons to use LRP over other Linux systems: fully administrable via high-strength encryption using secure shell and secure copy (ssh, scp), a huge routing feature super set and open source. These are things most commercial solutions cannot touch.
By now, I am sure you are probably excited to give Linux Router a try. Since it is now easier than ever to make a disk, you should do it.
All that needs to be done to get started is creating a file system and installing a boot loader on your boot medium. The most common choice for this is MS-DOS (FAT) and Peter Anvin's SysLinux. GRUB also makes a good boot loader. LILO is generally a poor choice, because unlike the former two, it cannot dynamically search for the kernel by name, and LRP does not have LILO available if you change kernels. Default syslinux.cfg and syslinux.dpy files are available at the LRP ftp site.
An LRP-capable kernel can be found in the kernel*.tar.gz archive. Choose an FPU (floating-point unit) or non-FPU kernel as needed and copy it to the disk as “LINUX”.
Copy the base and any additional packages; root.lrp, etc.lrp and log.lrp are required. modules.lrp is also needed if you are using a kernel with modules, as the provided LRP kernels do.
Finally, you must update the options passed to the kernel at boot time, commonly known as the boot loader's “append” line. A sample syslinux.cfg file looks like this:
display syslinux.dpy append=load_ramdisk=1 initrd=root.lrp \ initrd_archive=minix ramdisk_size=4096 \ root=/dev/ram0 boot=/dev/fd0,msdos \ LRP=etc,log,modules
The options after append= will be common to any boot loader you use. The LRP-specific options are:
initrd_archive=minix: initrd_archive is a kernel feature that lets you use tar.gz archives instead of raw images. This kernel patch is required in order to use LRP.
boot=/dev/bootdev[,fs]: this is the device name you are booting from. It is the device which linuxrc will try to mount to install any remaining packages. Optionally, you can specify the file system for the mount attempt. The boot line is semi-optional. linuxrc has a back-up list of devices to try to mount. It looks at /proc/filesystems for available file systems. Using boot= speeds up the boot process and ensures a mount if you are using an odd boot device.
LRP=etc.lrp,log.lrp[,modules.lrp][,package1] [,package2]...: these are the packages linuxrc should try to load at boot time. Remember that root.lrp is loaded by the initrd facility of the boot loader, so etc and log are the first packages listed here.
With your base prepared, you must add the needed kernel modules. Normally, one does this by booting the LRP disk, mounting a second floppy, copying the modules from that floppy to /lib/modules, editing /etc/modules and then using lrcfg to back up the modules package. Sound too hard? See http://www.linuxrouter.org/modmaker/ for a modules generation system. Just select the modules for the features and hardware you want; it will create a modules.lrp with the modules, their dependencies and an /etc/modules file configured to load all of them.
After boot and login, lrcfg will be started to help you get around to the files you need to edit. The configuration files for the base and packages are self-documented. You should be only a few minutes away from a working system.
|Speed Up Your Web Site with Varnish||Jun 19, 2013|
|Non-Linux FOSS: libnotify, OS X Style||Jun 18, 2013|
|Containers—Not Virtual Machines—Are the Future Cloud||Jun 17, 2013|
|Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer||Jun 12, 2013|
|Weechat, Irssi's Little Brother||Jun 11, 2013|
|One Tail Just Isn't Enough||Jun 07, 2013|
- Speed Up Your Web Site with Varnish
- Containers—Not Virtual Machines—Are the Future Cloud
- Linux Systems Administrator
- Lock-Free Multi-Producer Multi-Consumer Queue on Ring Buffer
- Non-Linux FOSS: libnotify, OS X Style
- Senior Perl Developer
- Technical Support Rep
- UX Designer
- RSS Feeds
- Reply to comment | Linux Journal
1 hour 29 min ago
- Reply to comment | Linux Journal
5 hours 29 min ago
- Yeah, user namespaces are
6 hours 45 min ago
- Cari Uang
10 hours 17 min ago
- user namespaces
13 hours 10 min ago
13 hours 36 min ago
- One advantage with VMs
16 hours 5 min ago
- about info
16 hours 38 min ago
16 hours 39 min ago
16 hours 40 min ago
Free Webinar: Hadoop
How to Build an Optimal Hadoop Cluster to Store and Maintain Unlimited Amounts of Data Using Microservers
Realizing the promise of Apache® Hadoop® requires the effective deployment of compute, memory, storage and networking to achieve optimal results. With its flexibility and multitude of options, it is easy to over or under provision the server infrastructure, resulting in poor performance and high TCO. Join us for an in depth, technical discussion with industry experts from leading Hadoop and server companies who will provide insights into the key considerations for designing and deploying an optimal Hadoop cluster.
Some of key questions to be discussed are:
- What is the “typical” Hadoop cluster and what should be installed on the different machine types?
- Why should you consider the typical workload patterns when making your hardware decisions?
- Are all microservers created equal for Hadoop deployments?
- How do I plan for expansion if I require more compute, memory, storage or networking?