High Availability Linux with Software RAID

Turn your machine into an HA server after you test it out on a VMware setup.

Using a combination of software RAID and a recovery CD, a high availability Linux server can be set up and configured easily. It is conceivable that a drive could fail, putting the RAID into degraded mode, without the system administrator knowing immediately. A cron job could be created to periodically examine the /proc/mdstat file and send an e-mail or page if the RAID volumes were found to be in degraded mode.

After testing with VMware, I used the same technique outlined above to install Red Hat 8.0 on my physical machine. I also went through the same testing procedure on the physical machine, including failing a pure RAID drive, failing the first mixed RAID and native drive, restoring RAID volumes in degraded mode, booting of the recovery CD, rebuilding the native partition and installing the boot loader on the MBR. It worked flawlessly, and my system has been reborn as an HA server.


RAID Levels

Red Hat



Micah Silverman has been working in software development and computer security since the 1980s. He's been working with Linux since 1991. Micah is a CISSP (Certified Information Systems Security Professional) and a Certified Java Programmer.



Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

backup of /boot partition and MBR on second disk rather than CD?

Anonymous's picture

Is there any reason that the following wouldn't work as ab
alternative to using a boot CD to back up the /boot partition and master boot record in case of a failure on the first disk:

  • Maintain a copy of the boot partition on the second disk and a copy of the boot manager MBR in the MBR area of the second disk (presumably configured to use the boot partition on the second disk).
  • In case of a failure of the first disk, use the BIOS to switch to booting from the second disk by toggling "bootable" bits on the partitions?

    (Can BIOSes typically boot from a second disk?)

edit which /etc/fstab?

Anonymous's picture

In your second test and its recovery steps, you say to edit
/etc/fstab to comment out the /boot entry.

Does the boot fail after the RAID drivers/modules are
loaded, so that the volume containing /etc/fstab is available?

Re: High Availability Linux with Software RAID

Anonymous's picture

using soft raid for swap is waist of CPU,

linux can do the same without soft raid:

just append to all swap partitions "priority=1"

and linux will use them as they were a part of striped soft raid.

Re: High Availability Linux with Software RAID

Anonymous's picture

In case of _real_ drive fail, Linux can (and, imho, will in 99.99%) panic.


In our case drive didn't responded, stupid scsi driver tried to reset scsi adapter, then kernel died...

Certanly, this is far better that lost of _full_ filesystem, but..

Hardware raid is _only_ choise for servers ...

Re: High Availability Linux with Software RAID

Anonymous's picture

I've used RAID and forced failures in dozens of ways and NEVER had a kernel panic. This is with adaptec and Sumbios controllers, and with basically unplugging the drive from a hot shoe while the server was running and serving requests (test environment, as well as actual failures in the real environment)

I did however, know a coworker using a HW RAID controller who had it mark two disks bad because the cable to them had slipped off while the server was being moved. Guess who had to rebuild and restore his whole RAID array because his $1000 RAID card wouldn't let him restart the RAID5 in place due to two bad drives.

P.s. the CPU load on my dual PIII 750 running flat out accessing it's raid arrays is about 1% of a CPU. If you have to worry about 1% of your CPU you have a lot of things on your plate ahead of that.

Re: High Availability Linux with Software RAID

Anonymous's picture

I don't think thats what the author was intending to achieve max performance. But more guarenteed availablity.

If you use the partitions directly in the fstab with priority=1 and a drive fails then the mache will probally go down since a portion of the swap space is now corrupt. However if they are on a RAID 5 setup the machine will just keep on humming. Assuming you don't have a 2nd drive failure.

Re: High Availability Linux with Software RAID

Anonymous's picture

Yes, You could do that, but then You loose HA, because swap will fail,

as soon a disk with a swap partition fails.

Performance wise it would be better to use raid 1 than raid 5 for swap.

Re: High Availability Linux with Software RAID

Anonymous's picture

Anybody has info of how to do this using User Mode Linux?

Re: High Availability Linux with Software RAID

Anonymous's picture

UML is part of the kernel, so is not affected by the RAID subsystem underneath of it. You just need to set up the RAID Disk system as explained, and then install a UML kernel, and way you go.

Re: High Availability Linux with Software RAID

Anonymous's picture

thats not totally true ...
a bug in the ubd driver in uml prevents raidhotadd from working correctly. the bug is known, and a patch is available to fix it (it will be in the next uml release)


Re: High Availability Linux with Software RAID

Anonymous's picture

If one is looking to truely run a HA server, would it not be better to make /boot a RAID-1 array, and use a Ramdisk to boot the machine and allow access to the Software RAID. Also, for better performance of the swap partition, rather than creating a software RAID disk for swap, set all the relevant partitions to swap space and set them to equal priority in /etc/fstab so that they are used as a RAID-0 array, without the overhead of the Software RAID system running.

Re: High Availability Linux with Software RAID

Anonymous's picture

Having swap on RAID is a good idea, otherwise a single disk error

can make your machine crash.

I would tend to disagree with

Anonymous's picture

I would tend to disagree with the whole concept of placing your swap on a raid partition.

See line #18 in the link below for more information:


We're not talking about strip

Anonymous's picture

We're not talking about striping though, but mirroring, so if one drive dies, all the data written to swap doesn't go down with it, as that would be double plus ungood.

Geek Guide
The DevOps Toolbox

Tools and Technologies for Scale and Reliability
by Linux Journal Editor Bill Childers

Get your free copy today

Sponsored by IBM

8 Signs You're Beyond Cron

Scheduling Crontabs With an Enterprise Scheduler
On Demand
Moderated by Linux Journal Contributor Mike Diehl

Sign up and watch now

Sponsored by Skybot