Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[rh5000]

Total non-tech person here... Last night I was working on some drawings and all of a sudden the PC had a hard time saving.  It appeared to save the drawing I was working on so I then closed everything down thinking it was "tired" (like me) and time for a reboot.

When I rebooted, none of my drawings were available.  At first it couldn't find any of the files mapped drives (folders) on the NAS so I got the feeling I had an issue.

Long story, short, I left it off and restarted the NAS this morning and I had my folders back.  But then when I went to access some of my files, I realized many recent files were not there. As best I can tell, I don't have anything after 8/17.  So, somehow I have lost (let's say "misplaced" (please god, please)) 10 days of files. 

This is confusing to me because I have (2) 2TB drives set up as a RAID 1.  I thought that meant that everything on one drive was more or less instantly on the other drive.

When I check on System Status...Hard Drive Status...Only one drive shows up. (The "right" one) so I assume that means the left drive has taken a dive.  When I go into the SMARTest, both drives show up with a check next to the "L" drive and it says 0% in the progress and won't do anything.

My question, I think, is...where did the last 10 days go and why isn't it on the mirrored drive?

Next question:  if I put the working HDD in my computer with a recover program, am I likely to recover any more data?

Next question: if I take out the non working HDD and replace with a new one, the Raid 1 will put what's on the working drive back on to the new drive, right?  It won't somehow erase what's left will it?

What else should I be asking or doing?

Thanks!!!

[cable2]

Hi,
Unfortunately, I think you have learned the hard way what can happen with Raid 1.  Here is a link to the sticky above: http://forums.dlink.com/index.php?topic=52509.0 , where you will find out in great detail the falacy of thinking of Raid 1 as a safe backup.  What is not outlined are also the possibility of problems with rebuilding of a Raid array.  You will not be the first or last to make this mistake.  Frankly, I am not sure about putting your new drive in your 320 to rebuild will go smoothly or not, but at this point I would encourage you to backup the existing data to some other external hard drive not in your NAS and do it before you attempt to rebuild your Raid 1 in your 320.  You should never depend on one device as your sole backup source for important data.   Massive hard drives are really cheap these days consider them as an investment and you will not be disappointed.  Repeat, backup your existing data to another external drive or installed in your computer first then try to rebuild the Raid with another new drive, or consider other options like some form of cloud backup.  Good luck

[ivan]

A couple of questions first.

Were you saving your work just to the NAS or is it on the PCs hard drive as well?

Are you using the NAS as an external drive to your PC and assuming it will also act as a backup?  There is a very big difference between the two (we use several NAS boxes as direct backup to our PCs hard drives - backup every 10 minutes - those boxes are then backed up overnight to other NAS boxes and our storage system).

If you were using the NAS as an extended drive did you call up any of your work from the missing 10 days?  If you did and you could read it then it is most probably there unless you have the RAID 1 set to automatically rebuild in which case it MAY be 'gone' but recoverable.

Now to try and solve the problem, or at least not make it worse.

First, remove the possible failed drive leaving the 'good' drive in place (power off the unit to do so and while it is off give a visual check of the inside for fluff build-up or damage).

Now you use the 'failed' drive to practice on.  Get a USB/SATA caddy or powered adapter and install/attach the drive.  Assuming you are using a windows PC you can either download the software listed in the recovery sticky (Ext2IFS works best on older windows systems - R-Studio and NAS Data Recovery work on all versions of windows from XP but cost money) or follow the Additional References link and use a Live Linux CD.

Which ever route you decide to take you may, or may not, see any information on the 'failed' disk.  You can also download the disk manufacturers disk test tools and use them to test the disk to find out its actual state - good, bad or indifferent.

Because you set the drives up as RAID 1 you should be able to power up your NAS and use the files on the remaining disk as if both disks were present.  You may find your missing data.  If you have problems doing so there may be a problem with the NAS or power supply.

The next thing to do is backup all your data on the NAS to an external drive (it can be 2TB USB drive or an extra disk installed in your PC) that way should anything go wrong you have your data safe.

If you can read and use the data on the 'good' drive, get a new one and install it and allow the raid to rebuild.  You are then back at work but, for your own peace of mind, backup your work every night as the last thing you do.
  
Edit: Cable2 you beat me to it again

[rh5000]

The issue with relying on the RAID is pretty clear.

What I don't get is what would happen to lose 7 or 8 days of data on two drives within a matter of minutes.

I can see that one drive failed and I can see that with it failing the other drive might "mirror" the loss of some data just as if I had purposefully deleted it but would it do it that quickly and that...what's the word...comprehensively?  I have nothing after 8/19.  (I said 8/17 earlier...that was incorrect)  But, that just seems odd.  

For instance, I have a file on my drive currently working dated 8/19 but I modified and saved that drawing yesterday.  So, the file isn't gone...it's an old file.  That means that the drive on right that is still working had not been updated since 8/19.  Isn't that strange?  Does that mean my RAID had not been working since 8/19?   Seems like 2 different issues.  1) A HDD not working and 2) the NAS not copying things... no?

Just trying to reconstruct the issue so I can figure out how to try to recover some data.  Wondering if I'm incorrect about what is actually happening.

Thanks again!

[ivan]

Without technical knowledge it is very difficult to try and reconstruct what happened.

What might have happened. 
One disk fails or partly fails so that some files are not fully written to the disk, they may have been written to the other disk or not depending on if a write error was flagged (the failure could be a bad sector of the disk and the disk has used up all its replaceable sectors). 

Several days later you see a problem and power the system down. 

When you power it up again the suspect disk is flagged as having problems by the firmware and a raid array rebuild is started - this can cause loss of data if one of the disks is partly working but has bad sectors (the firmware tries to create two equal sets of data and if that is impossible it can go into panic and some files get flagged as erased - those can be recovered from the good disk if you do not write to the disk before attempting recovery).

If a disk is seen by the firmware as dead the files are written to the other disk as normal.  It is how the disk has died that causes subsequent problems.  The standard procedure is that notification is send that there is a problem with a disk - the problem disk is pulled and a new disk is used as a replacement - the firmware then rebuilds the raid array.  Depending on usage the rebuild may be fast or very slow (for example, our RAID 1 arrays in our HP Storage Array rebuild in a few hours but a NAS box can take over a day to do the same thing - depends on usage and processor power).

I assume you didn't notice writing problems until you did which means that there must have been problems in writing to the NAS, assuming you used the same file name which would cause an overwrite of the existing file, or at least a flagging of the space it occupied showing it could be overwritten.

Sorry I can't be more helpful but trying to diagnose problems by a forum post is rather more difficult than having the unit out on the workbench.  There are things my repair staff do without thinking about them but which mean success in the repair, something we can't duplicate here.

[rh5000]

I must have been replying at the same time as the second question/post there.  

Yes, all my work is on the NAS.  It's on the NAS so it can be accessed by more than just one PC. I thought it was being backed up to a separate USB drive but I was not too observant of that and apparently only "windows" files are being backed up.  Of course, now that I have an issue, I see the error of my ways.

Still confused though.  Why is one HDD dead and the other is stuck in the past?  

I tried to run the NAS with the one I suspect is dead (alone) and it will not show up so I think that one is the issue.  

I am running the NAS now with just the one that is working and that is working...but still only has files to 8/19.

I just now see another reply by Ivan and that does actually shine a light on a possible scenario.

I can see it trying to reconstruction the bad disk (1) from the good disk (2) but I'm surprised the good data was from 7 days ago.

So, going forward I have a new External HDD that will be a backup for the NAS but what, then, is the point of the RAID?  Why waste the extra HDD space for a mirror if, when there is a problem, that mirror is either useless or outdated?  And...not to be dense but can't the same thing happen with a backup...as the backup overwrites itself each day...if info is corrupted and you don't notice it for a couple days, haven't you just covered good information with bad in the same way?

The "dead" drive got warm while in the NAS and I think it is spinning...just apparently not reading.  Is that a sign I might be able to recover it?

Going to try to install the "dead" drive in the PC and see if I can find something.  Did this once before with a Linux disc but was totally clueless and remember it being difficult.

I really appreciate the input.

[ivan]

The time scale for good data possibly indicates when the problem first started to manifest itself.

RAID 1 gives protection if a disk fails because it allows you to still get at your data.  We use it and set it up for our clients use, we also have e-mail notifications enabled and have on a couple of occasions arrived at a clients office to replace a dead disk before they were aware there was a problem (that is part of what they pay us for).

In your case you need to evaluate exactly the amount of data security you want.  If your Nas is switched off on a regular basis then it might be better for you to have it setup a two single disks with one of them being your working disk and the being the backup of that disk - in that case I would still recommend that you also backup to an external disk as well for added security.

To check the 'dead' disk I would recommend using the USB/SATA caddy I mentioned before.  Mounting it in your PC is almost a guarantee that you will lose anything on there because the NAS uses the Linux file system and windows does nor recognise it at boot without a running IFS and as a consequence does nasty things the the MBR.  With it mounted as a USB drive you can use the manufacturers disk test tools to check if it might be possible to recover anything from it - in some cases you can, others not, it depends on what the cause of the problem is/was.   

[rh5000]

Thanks again.  I think I am going to leave the set up as it is but with the addition of the back up USB drive just to be safe!  Seems like that was more or less what you were suggesting.

Probably should add some sort of cloud storage too.  This NAS will do email notification, right? 

I did pull out a HDD enclosure and now I am in the throes of fighting with Ubuntu (seems to think I have a password I can't remember but doesn't seem to think I exist when I try to change it) but I don't think it's going to work anyway.  I have a little baby laptop that has Ubuntu on it from the last catastrophic NAS failure I had...disks were fine...the unit itself (Thecus) died.

The disk just spins non-stop in the enclosure and even logged on as a guest (in Ubuntu) it didn't appear to show up in disk utility.  Think I'll have to layout the jack to get it recovered by someone smarter than me.  If it's not a fortune it's probably worth it for a week's worth of work not to mention project pictures I took at a site 150 miles away.

Would love to have some real IT support but I work from home and still just trying to keep the lights on so I'm making do with my own meager efforts.

So I have both a 3TB external drive and new 2TB bare drive coming.  My plan is to copy everything to the external and then put the new bare drive in and see what happens.  If all looks good, then I'll delete from the external and then set up the external as a back up.  Hope that makes sense.

I'll follow up with what can be done with the broken drive and what it costs me.

-Ron

[ivan]

Your proposed setup with the USB disk and cloud storage as backup to the NAS should work.  I say should because it requires your effort to set it up and do periodic checks that it is working.

I can not emphasise enough the fact that if anyone can't afford to lose their data they must have a tested backup system in place and if it is very important they should have, if possible, an off-site backup as well.

How frequently you do your backups depends to a large extent on how often your files change (for example, we get the incremental backup of the accounts data from on of our clients every half hour and the full backup every 24 hours, that is stored in two places here which are backed up off-site every 12 hours - yes they pay for that service but it has payed for itself after they had a fire, arson, and they were up and running the morning after in temporary offices).  Note, I am not saying you need that amount of backup, I just gave that as an example.

I can feel your pain at the potential loss of data, I think we all have been through it at some time.  For me, it was when half the stack of punched cards I had prepared for the final run went missing and I had dumped my working notes so had to start all over again.

At this point your NAS should be working with the one drive in it.  If you have space on one of your PCs use that for saving any of your work.  The reason being, that good drive may hold your missing data (unseen because the space it is in has been flagged as usable) and if you write to that disk you may lose the data by overwriting it.  That does not mean that you can't read the data from that disk.

The next paragraph may help you or it may not, it does require some specialised knowledge, equipment and software.  I give it because I think your best chance of recovering anything will be from the good disk.

Our procedure in a case like yours would be to pull the good disk and make a sector by sector copy image of it, that keeps any deleted data safe and increases the probability of recovery.  Next a file by file copy of the available data is made, this is the backup and is stored on our storage system.  After that the disk is tested using the manufacturers test software and if it passes it is returned to the slot it was taken from.  The unit is then powered up and we check that the drive is seen and accessible.  If all is OK the replacement drive is installed, the unit is powered on, and rebuild RAID is activated.  While the rebuild is taking place the disk image is checked for deleted files and any found are recovered to a safe store - they will be copied back to the unit when the array is rebuilt.  After that the unit is checked again to ensure all files can be accessed then the unit is returned to the client.     

If you are prepared to spend some money you can use it as a description of what you need done and you could possibly get the the local college or university computer department to do it for you.

[Bulles30]

Hi,

Your issue seems to be the same as a well known which appeared with FW 2.0. And after upgrading my NAS to FW 2.0, I experienced the same problem as yours.
The solution was to connect one of the raid1 disk in a caddy and use ext2ifs. I saved up to 90% of datas correctly.
Fortunately, the more recent FW corrected this issue.

So, what version of FW is installed in your NAS ? And if it's not the last one, start an upgrade quickly in order to avoid the same erratic issue.

Good luck !

[ivan]

Bulles, did you notice that the OP had a dead drive that precipitated the problem?

All of our DNS-320 NAS boxes are at firmware v2.00 and we have not seen any strange writing problems here and neither have our clients.

[Bulles30]

Ok, maybe I can't remember exactly what version of FW was bad, but I wasn't alone having this issue. Same kinds of issues were also reported using Teracopy and DNS 320 (with loss of datas, or moving directory trees without reasons).
Which disappeared with the latest versions of the FW.
And it's usually a good practice to indicate versions of components when reporting problems, having them at the good level.

But I didn't want to open a discuss and I wish that rh5000 will retrieve datas with your advices.

Bye,