Community Pick: Many members of our community have endorsed this article.

Single partition, mirror raid, recursive backups and oncoming crashes.


I read in a discussion about a person who configured a very simple mirror RAID with two hard drives; the system and data were on the same partition. He asked how to repair the system as it was not booting up anymore.

In his case running chkdsk was enough to restore the master boot record functionality, but this will not always work. My suggestions below provide an escape route from this situation where your data is not recoverable.

First, frequently back up all your data to some external storage such as another PC or an external hard disk. It does not matter whether you use RAID; data can be lost anyway. If needed, save the drivers and configuration, a list of all programs you previously installed, and don't forget to reinstall tools and other things you might need later on.

The person was very absorbed by trying to make the broken hard drive bootable again, but my question is:
"Why should you make it bootable again?"

Anyone working with this setup should consider himself lucky at any time not to have lost it all. As a matter of fact, it seems the person did not save an image using a hard drive cloning tool; if he did, he would still have the the possibilty to restore the MBR, the first sectors, the partition table and/or the partitions/whole disk to an optimal condition. Those images are best created immediately after a fresh mint installation or after a restore on which are added major modifications.

When using a RAID it is also wise to save in some form (even just straight on paper as a reminder) the exact, detailed, configuration of the controller.

However, it is not really wise NOT to prevent these problems.

The above information and backups are the easiest way to rebuild your system in case of issues. Without them the only way to get back on rails is to repair the system, and this might be a big headache, very often a try and fail road with plenty of lost time and errors. It is even more difficult when there are multiple issues to clean up; this happens very often as, for example, many viruses open ports and let other malware install to the computer.

It is easy to see that the time and effort used to repair the system might be tremendously higher than just replacing it all with a fresh installation saved in a image.

You will realise that replacing the Operating System from saved images is very easy, whereas reinstalling from scratch is a much larger undertaking.

My suggestion for everyone who has an old, broken--not really pristine--OS, which may be heavily fragmented and on one unique partition, is to backup all data immediately, clean everything from the hard drive (take care to not delete hidden factory partitions if you wish to conserve them).

Now reinstall Windows in a 30-35GB "system only" partition (for Windows 7 and Vista you will want 50GB), then create a second "system data only" partition (maximum 32GB if you are using FAT32); this depends on the cloning tools you will use).

Create finally a third "user data only" partition with the remaining space.

If you want to have a mirror (or any kind of) RAID it is necessary to follow the extra steps to set this up accordingly to RAID rules, I will not explain how to do this as it goes beyond the argument I am writing about (read other specific articles or discussions if you need help on setting up a RAID).

When finished with the fresh mint new installation, make all backups possible with a disk cloning tool such as Clonezilla, Ghost or others, and check the backups are valid (many cloning tools have an option to do this).

Imaging the "sys data partition", the MBR, the first sectors, the partition table and the system partition, you will protect yourself from future crashes and you will have them always at hand in order to restore the PC.

Take care of the fact that not all cloning tools let you save in such a specific and fine way the various critical areas of the hard disk as described above.

Many just make a sort of big backup containing the system partition and maybe also other crucial areas, if you want to be sure to be able to solve also major problems, as a hard disk hardware failure can be for example, take care to choose a cloning tool that is simple but let you save the crucial areas in a selective way and in separate images too. This comes to be very important if you need to restore the system on a new hard disk, or when the MBR is ruined for any reason.

BUT... don't forget to backup the saved special files to a hard drive outside your PC too, you are less likely to lose it with two or three backup copies, if you are not in bad need of disk space leave the original backup in the FAT32 partition, so to be able to launch it fast and directly using the hard disk itself to restore the system partition, but as a security measure you MUST save it also outside the computer on other mass storage devices as a portable hard disk and a set of DVD.

Remember that the mirror RAID is merely hardware protection against Hard Drive failures. It is not a very clever system that understands something that is going on; if you introduce any problem in the disk, it will be mirrored on the other one too, so as long as the RAID works you can always expect to find an identical copy of your main disk, including all of your errors, problems and viruses.
In this cases if it does not boot from the main disk, it certainly won't from the other mirror disk either.

The "RAID 1, so a simple 2 identical disk mirror, and a single partition for the whole drive" is a far too simple configuration to give you any special security, you will get only, and brutally, an identical copy of the main disk. This will protect from a disk failure in the RAID set, but not any corruption of the data held on those disks.

If issues are somehow introduced in the system partition those will be included in the mirror disk automatically.

Once you learn to make recursive backups with disk cloning tools you will never stay without them again, guaranteed.

This is the real solution to the problem. The problem in reality is NOT that you might not reach a fix, but that even if you reach it successfully, you will have the problem again and again. Just as surely, sooner or later the trouble will be beyond your capabilities or will be just not solvable for some reason, ranging from major hardware failures to multiple complex software disruptions at various levels.

It is a must to consider that working in this way, with a unique partition and no restore images -- a sort of "naked in the deep jungle trip" -- is really dangerous and risky for your data and not practical at all.

Keep it in mind! Anyone can experience a disk or other hardware failure in any moment, being ready to restore on the same disk or a new one makes the big difference that goes beyond a reasonable comparison in relation to a reapration attempt.

I write "risky" for your data because there is always the random risk of losing both disks, losing the controller and one disk (with a mirror, this is not high risk, but if it happens you lose it all and get a "IT disaster"... you never know), or losing it all at once (lightning, severe shock, theft, motherboard/controller major destructive failures etc), so backup all your data frequently on storage outside your PC.

But most important, always create  images to recover the system with your favourite cloning tools and save them at least in three different places.

This is the very minimum to do in order to stay safe and reliable; further steps must be taken for higher levels of security, but this again goes beyond the basic matter discussed here.

Best regards

Comments (0)

Have a question about something in this article? You can receive help directly from the article author. Sign up for a free trial to get started.