Partitioning drive

I believe that the read/write heads does not come in contact with the disk surface, therefore there is no such think as "reduce excessive wear on one location of the disk".  Disk positioners do wear as does the motor and bearings. 

Block size is fixed therefore this is impossible "matching block size with the file size"

You will find out after many attempts the you really can't separate application and system software for many reasons.  I will give you two.  When you install an application the registry on the system partition gets modified even if everything else gets loaded elseware.  Microsoft tends to load certains parts of applications on the system drive irrespective of where you want the application loaded.   I think that fragmentations effect is over hyped, I defragment when I think about and it has never appeared to effect the performance of my PC.  Of course my disk is 40 percent empty. and I have a pile of RAM, so I don't swap, and am running a 15,000 RPM SCSI U160 drive.

I my experience I have lost three drives at home an six at work.  All six at work were caused by bearings freezing on the motors.  One home drive was installed in 1989 and died 1997, and other drive dies at 5 years of age.  The last drive died this past summer unexpectlly at only 2 1/2 years  of age and I have been unable to figure out exactly what happened.

 

Message Edited by msgale on 12-12-2003 10:31 PM

There are some general practices that some people use that may help you. Some people will break up their drive like this:

Partition 1: OS
Partition 2: Personal data and software

or

Part 1: OS
Part 2: \Documents and Settings
Part 3: \Program Files
Part 4: Swap partition (usually twice the size of your RAM)

It's really up to you how to do it. The main reason I can think of to partition your drives is to keep your OS separate from your data. If you ever have to reinstall your OS then you don't have to worry about formatting that partition and loosing any personal settings. Or if you do have personal settings and data on your OS partition, you can easily dump it over to your second partition before reinstalling.

Message Edited by mikerl on 12-13-2003 05:06 PM

I see no problem separating data from programs as as mikerl said it has the benifit of improved data protection, however I stongly disagree about separating application from the system.   Let me explain.  When you install a program the application get installed somewhere, but also references to the application are placed in the registry.  Unless you have a good current backup of the system you will loose the registry, after you restore the system you will probably have to restore the applications to insure registry references are correct.  On the other hand you lost the application partition, you have a big time problem, uinless you have a good current backup of the application partition.  The registry which is undamaged has references to non-existant application.  These references may very preclude reinstalling the application, and since the application does not exist an uninstall may very well not work either.  Result you will have to manually edit the registry.   P.S. As I said in my earlier post many Microsoft Product will instal some files in the system partition no matter what you selected.

As I said... it's a matter of personal preferrence.

I could list reasons why separating os, applications, and personal data makes sense, but it really doesn't matter. I can tell you also that this type of configuration makes even more sense when used with an OS that is not bound to a 'registry', eh-hem *Linux*.

I use a two partition scheme for Windows and a four partition scheme for Linux.

Actually it does matter, or at least it should.  The more imformation given in a response the better the requester can judge the appliciable of the response.   About Linux, why muddy the waters, this is a Windows XP form.  P.S. four partitions seems too few for Linux to me.  Typically you need a root partition, a swap partition, a var and user are also handy to have, where are user's home files kept.

As I said... it is a matter of personal preference.

My computer is still running fine and I'm not running a configuration anything like yours. I have never had a major data loss due to drive failure. I can reinstall Windows and all of my applications in less than 2 hours if needed. As to the mention of Linux, it was nothing more than to explain why I use the second configuration. I never recommend any configuration, just provide reasons why I think a certain configuration is good. I would never dispute anyone elses choices in system configuration. A computer will run with only one partition. It will also run with twenty partitions. It really doesn't matter.

To quote you... " The more imformation given in a response the better the requester can judge the appliciable of the response."

Unless, of course, you meant only information provided by you applies.

One - I don't partition my windows disks, never have.  Two - I agree Linux is off topic, I only included to demostrate that I know some aboput Linux and what I didn't say alot about UNIX.   Three - I am not sure why I fixated on you last stetement but I did.  Your point is well taken about my post, you have given a reason for partitioning.  Four - I Still contend that you have not addressed the issue I did raise about the interrelationship of applications and the OS, and that is not an opinion.  Five - this whole issue would go away if everybody did daily back-ups and disaster recovery backups, but almost nobody does.  Six - I have not ever been able to rebuild a system, LINUX, UNIX, or Windowsit  to exactly match as it was before a crash if the system was more that a few days old.      

I can't believe I'm letting myself get drawn into replying to this thread again.

The reason I put my applications on a separate partition from my OS is because some of the apps I use will store information within the Program Files folder. The most important one is Adobe Photoshop. I record my frequently used routines into actions. These actions are stored in the program's folder. Another application is Trillian. All user configuration files, user lists, and history are stored in the program's folder. Adobe Photoshop Album stores the db files in the program's folders. Even though I still do have to reinstall the application after a reinstall, I don't have to go hunting for these files. It also removes the possibility of human error. The bulk of my applications do still reside on the same partition as my OS as they do not store any settings or data in their folders.

Incase the issue comes up, the Documents and Settings folders, I put on a separate partition because they contain the My Documents, Favorites, Outlook, Cookies, and many other user specific settings and files. Much in the same way that Linux (there's that evil word again) does with the /home/ directories. As you said before also, much of this information is bound to the registry so will not be usefull until the applications are reinstalled. But this removes the possibility of me forgetting to backup something important.

If I was not able to rebuild my system to the exact same functionality following a reinstall of my OS, I would re-evaluate my disaster recovery plan and data storage practices. There is no reason at all that a system can not be restore to do the exact same thing as it did before. It's hardware and software, there's nothing mystical about it.

For someone that never partitions his Windows drives you surely have a very strong opinion, and much information to offer on the topic.

I have provided my reasons for my configuration. I just feel like a fool for justifying it to you. It's the equivalent of justifying why I put my food away in the pantry in the order I do. I put my pasta on the same shelf as my vegetables, do you have a problem with that? :)

Take care,

Mike

Guys, chill out. I welcome discussions from all sides. There is no contest here. Just say what you believe and I think most readers can judge what may apply for themselves.

I thought the block size is based on the partition size. At least that is what I read with some of the partitioning related articles and yes, it is precomputed and fixed but a large partition results in large block size which make it inefficient to store most small files like regular documents or web pages. I was told that was one of the small flaw of big drives.

I didn't do much research on why a disk fails. Had 3 nonrecoverable disk errors and 2 corrputions (that requires total reload) in the last 5 years. One disk died in a little over a year. The other two lasted more than 5 years though. I do know there is no physical contact between the head and the media(or that is the end of it) but not sure if repeated read/write on the same area of the media surface would or would not shorten its life. Would welcome comments from people who know more about disk mechanism than I do.

I am thinking to separate out system from application because I'd like to minimize the fragmentation on the system drive. yes, I know registry get changed and many common library files are actually copied on the system partition. In general, I often installed and uninstalled applications to try them out and deleting things can often be the beginning of fragmentation. So, keeping the application from the system tend to slow down the fragmentation. Of course, my personal experience probably give me the bias since I had two idential systems one set up with everything on the C: drive and the other with partitioning. The C: drive system corrupted by unstable application doing stupid things and I ended up spending quite a bit of time to recover(and fail) the system and lost some of the data which sits on the same partition.

When the partitioning system had a physical drive problem, I was able to save the non-corrupted portions before the whole thing went dead. It could just be an coincidence.

Yes, if you have a good backup, you can be up and running quickly. My past crashes typically put me back 2 days. The first day trying various way to diagnose and fix what happened. If failed, try to save anything savagable. The second day reinstall everything ... I am afraid I can't do that in 2 hours. The quickest time I did took more than  5 hours because I had a lot of things on it and backup was not good.

The painful past leaded me to get the raid one this time and a decent daily backup program. Like I said before, I would like to find a sensible scheme to properly partition the drive to get reasonable efficiency in use, lower fragmentations. Actually, I used to use multiple raw partitions and/or disk drives in Solaris(variant of unix) to speed database application up as a DBA. The performance improvement is real and significiant.  And I do have data intensive applicaton on the XP box so I am trying to see if there is anything I can do from that angle. I was thinking about raid 5 but Dell doesn't have that for the consumer box(Dimension 8300) so I 'll probably do with what I have now and upgrade it later when disk access become a bottle neck.

Block size varies with partition size for FAT partitions, not for NTFS, as far as I know it's always 4KB for that. Also for FAT32, it isn't much of a problem (4 to 8GB, 8 to 16GB, 16 to 32GB, 32 above that)

This seems to happen frequently on the Delltalk forums, and is the reason I rarely frequent these boards. The 'regulars' seem to flame anyone that posts something that even resembles a conflict with their opinions or statements.

You, having worked with Solaris as a DBA, know far more than most (self included) about the reasons for setting up various types of partitions (and files systems) for different uses and data types.

I apologize to all for letting myself get lured into, what now appears as, a debate. I thought I could clearly state my opinions on this matter without having to explain myself. I was wrong. I'll go back to frequenting my usual boards so I don't bruise anyone elses ego.