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

If the DNS-323 itself fails, the drive(s), assuming individual volumes or RAID1, can be pulled and installed in a linux system and the data recovered, they can also be installed in a Windows system with an ext2 filesystem driver installed and the data recovered - by installed I mean connected either internally or externally to whatever port the operating system will permit a hard drive to be connected to - SATA, eSATA, USB or firewire - if the system can recognise a drive attached to the port, you're good to go.

Now - since you're using the unit as a backup store - what's the worst case scenario?

The unit failing or being destroyed simultaneously with the primary data store - not impossible - but you need to consider the probability, I would not use RAID1 - RAID1 is designed to reduce the impact of a drive failure - what is the impact of a drive failure in a backup store?  Usually none, unless it fails at or around the same time as the primary store fails.

hmm, not sure how to install a ext2 filesystem driver so maybe I should get rid of Raid 1 and go to standard mode? Does this sound right?

I see. So this is a piece of software that enables Windows XP to see the DNS drives if I plug the SATA drives to a USB adapter?

Before I go ahead and install and test the recovery process, is there anything else that I need to know? ie. what are the down sides to installing this piece of software on a Windows machine etc...

Thanks in advance!

So let me see if I got this straight. If the DNS-323 is used as a backup file server (where daily dumps are done to it through the LAN), AND, if installing the FS Driver will allow Windows to recognize these drives once they are pulled out of the 323, then wouldn't it make sense to keep the Dlink in Raid 1 rather than doing a daily back up from Drive A to B? In other words, once the daily backup schedule runs, Raid 1 would duplicate the data across the 2 drives without requiring a 2nd backup schema to be run from Drive A to Drive B.

Reason for all of this is that Raid 1 is used to prevent single disk failures but I am using the 323 as a backup file server. So wouldn't Raid 1 be a better setup than standard since it is only being used as a backup server?

So I guess I am wondering if removing the Raid setup would lower the risk of losing the data on the unit? Any more thoughts?

So would Raid 1 bring any more risk to a back up server? I am thinking that if there is a power failure during a backup scheme running, both drives might end up with data corruption versus running in standard mode where a power failure during a backup might only result in 1 drive dieing.

Any more thoughts on this issue? We are trying to use the NAS for a backup server and recovery would be via Windows (either direct USB connection to the drive or through the 323 if it still works properly).

Exactly. After thinking about the situation some more, I am leaning towards running standard formatted dual hard drives. Backups would occur sequentially (LAN to Drive A then Drive A to Drive B). I think having to stick with a 500 Gig Raid Array even if bigger harddrives become more accessible in the future just does not make sense for a backup server. Not sure if this is correct in thinking this way but for backups, Raid would just be another factor that I have to deal with in the long run. WHy not simply move it out of the equation for a backup server. Sounds right?

blbrown wrote:

I agree.  Remove anything that adds complexity without adding benefit.  That way you have fewer points of failure.  For copying the backup data between disk A and B within the DNS-323, you should check out this thread:

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

Thanks for the vote of confidence and for the link. I read the first page of that thread and it looks like it deals with users who use Volume_1 as a real time file server. My situation is a little different in that the backup software is scheduled at 2 times in the day to copy files over from LAN'ed computers to the 323. I will simply set Volume_1 to sync with Volume_2, using the software that takes care of the original backups, right after the end of the backup sequences. I think that should work beautifully without having to do anymore to automate the process (ie. working with more code).