DSM-G600, DNS-3xx and NSA-220 Hack Forum

I don't see anything wrong with your assumption on the size - but - if you're assuming that RAID1 removes the need for backup, that would be an incorrect assumption.

What do you define as "stability or recovery issues"?

This is why I pointed it out - many people have the impression that RAID1 is there to protect against data loss - and when it does not deliver on their expectation they get upset and trash talk the device and it's RAID implementation.  RAID1 has nothing to do with preventing data loss - it's purpose is simply to reduce the downtime impact of a failed drive.

Regarding stability & recovery - there is no recovery process, if a drive in a RAID1 array fails, the end user is typically unaware of the event, life goes as normal, and the data remains accessible just as it was before the drive failure.

What should happen is that the system administrator should be alerted through some mechanism - in the case of the DNS-323 it's an amber LED (which in my experience does not work) and, if you so choose, email alerts - the administrator would then remove & replace the failed drive and the unit should rebuild the array - you'll find that some people here have reported problems with data loss during the rebuild, one possible cause is rebuilding to a drive that already contains data - personally I have had no problems in this regard.

Here's another plug for my script...  b8b, what I would recommend is changing back to "standard" setup where both drives are independent.  Then have a script automatically copy from one to the other only the things that need backing up.  Here's a thread related to that:

http://dns323.kood.org/forum/t1150-Tuto … night.html

You might also want to check out my scripts:

http://backupnetclone.sourceforge.net/

b8b wrote:

It seems that RAID also prevents data loss since it is redundant, no?  Perhaps I don't understand your statement.

Well, I guess this depends on the type of the data loss event you want to protect against.  You are correct that
the data is redundant; however, the redundant data is also being operated on, in real time.  So, in the situation
that you have a power failure (assume you don't have a UPS) and a file you had open gets corrupted.  Both
redundant copies get corrupted and you lose the data. Same situation if a user deletes the wrong file (or directory),
both redundant copies are deleted.

RAID 1 (mirroring) is designed to allow you to continue to access your data in the event of a single hard drive failure.

If the only data loss event you want to prevent is a single hard drive failure, then RAID 1 is a pretty good solution. 
However, RAID 1 is not a very robust backup strategy.

From http://en.wikipedia.org/wiki/Backup
"In information technology, backup refers to making copies of data so that these additional copies may be
used to restore the original after a data loss event."

Last edited by mig (2008-02-01 22:04:03)

b8b wrote:

Two questions: 1) is my assumption on the sizes wrong?  2) will I create stability or recovery issues with this mixed-mode format?

These are really good questions!  It shows you are thinking. My advice to you, is to 'practice' your disaster recovery
plan.  Think of all the data loss events you want to recover from, and develop a procedure of how you would do that. 
Prove to yourself, by simulating the data loss event, that you can restore the data you care about. 

For starters, consider, what would you do to recover your data if....

  1) a user deletes a file (or directory)
  2) single hard drive failure
  3) both hard drives fail
  4) the DNS-323 fails (and what if the DNS-323 is not manufactured anymore?)
  5) catastrophic loss of building (fire, earthquake)

You don't have to be able to perform a complete data recovery for every situation,  only you can determine what
costs are acceptable to prevent the data loss.

Last edited by mig (2008-02-01 22:32:02)

So to go from a Raid 1 setup to JBOD, can't you simply:

1. Power down the unit
2. Pull out drive A
3. Reformat drive B using JBOD
4. Plug in drive A and copy data over to B
5. Done?

No I haveen't, hence I put that list as a question. Come to think of it, the safest way might be to back up all the Raid 1 data to another drive on the network, reformat the 323 using Standard, then pull the data back. Can anyone who has done this migration jump in here?

Sorry, I meant to say standard instead of JBOD. ok, I promise I won't come looking for you if things break

So let me double check just to make sure.

Option #1:
1. Backup data to another drive
2. Power down unit
3. Remove Drive A from the unit
4. Power up unit and push reset button at the rear
5. Power down & reinstall drive B
6. Power up and use

OR

Option #2:
1. Backup data to another drive
2. Power down unit
3. Remove Drive A
4. Power up unit
5. Format Drive B (using the unit's GUI wizard)
6. Power down unit
7. Install Drive A
8. Power up and use

Are both of these methods properly sequenced? Which method is better?

Again, thanks in advance!

Add up:

1. Come to think of it, in Option #1, what does the reset button do at the rear? Does it kill the Raid format?
2. in Option #2, when you replace Drive B, wouldn't the unit try and rebuild the Raid array? Or would it not because the last drive left in the unit has been reformatted (meaning the first drive you pulled out does not contain any Raid array info)? Not sure if my lingo is right so please bare with a newbie here

Last edited by mealto (2008-02-08 21:47:38)

fordem wrote:

The sequencing listed is basically correct, although there is a mistake which I believe may be a typo in option #1 - item 5 should say drive A rather than drive B - as it stands you're removing drive A and then replacing drive B.

Edit: Never mind, I got Option #1 & #2 mixed up. Looks like I need to think about this before the format change. Thanks so much!

Last edited by mealto (2008-02-09 03:06:38)

mealto wrote:

... 3. Remove Drive A ...

Which drive do you call Drive A?
Left for right hand side (as you look at the DNS-323's power button?)