D-Link Forums
The Graveyard - Products No Longer Supported => D-Link Storage => DNS-323 => Topic started by: joejiz on June 15, 2014, 04:36:56 PM
-
Hi
Everything was fine (raid 1) 2 TB Seagate drives. Up for months no problem.
E-mail alerts setup and worked just fine.
My dad reports his outlook 2003 email inbox is missing information. ( his pst file is saved on the dlink shared drive). . .
I rdp into his machine and check his shared drive mapped drive. Sure enough I only show 2 folders instead of a bunch of folders. I go into the folder where the pst file is a bunch of other files in there are missing.
I check the properties of the mapped drive and it says its only 60 GB. ?
Ok that is weird.
I log into the webserver for the device and see that hard is blank for left bank. Ok so no email alerts or anything.
Nice job of saving the data should a hard drive die!?!? Maybe his outlook file was open during the time the hard drive died or something.
PS no hibernation is set on drives. So I am not in front of the device. I am province away.
I choose systems - restart
upon reboot I get two emails
1- raid degraded
2- left hdd failed
Ok, I browse the shared drive via mapped drive on windows 7 machine and all the folders start showing up. And files in the mail folder start showing but the main pst file is 4 gb less still.
When the system came back up DLINK 323 web server said the current hard drive is to small to resync or essentially fix the raid 1.
"The replacement hard drive does not have enough capacity to support the RAID 1 volume. Please insert a hard drive with more capacity."
Any ideas anyone?
Even when I put a new drive and it finishes resyncing the mail file has been used so it's not like it magically will bring back the older file and overwrite the new pst.
We're screwed aren't we.
-
To answer your last question not necessarily, it depends on what you have done since the drive failed.
Assuming you don't have a backup (any form of RAID is not a backup) and you have not messed about with the good drive you should have all of the data safe on it.
Read through the information contained in the Sticky Topic: DNS-323 - Data recovery (Windows PCs).
It will require that you have physical access to the good drive.
If all goes well you should get everything back. I know it works because we tested the recovery that way.
The only problem I can see is that by your remote access without checking the RAID 1 status you may have forced a partial RAID rebuild from the failed drive to the good drive which would result in the 60GB size report. If that is the case then R-Studio bay be able to recover more of your data than a simple copy using Ext2IFS.
-
Hi Ivan,
I don't have a recent backup of the file, the file size should be 4.5 GB and its only 87 MB.
I will review the Sticky as you suggest.
This is what I believe happened.
1. Unit running fine, hard drive died
2. Unit tried to auto rebuild on the fly(as set) but the bad hdd identified itself as 60 GB only.
3. Windows 7 pc browsed to mapped drive and only showed two folders out of 15.
4. I right clicked on mapped drive, holy crap what is the 60 gb crap?
5. Loaded web server ( remember no dlink alerts for failed drive, rebuild or anything ) - never seen this before
6. Went to status found out on drive failed, degraded raid status.
Actions Taken:
1. loaded web server , chose restart
2. Logged back in and the unit on FIRST page said ""The replacement hard drive does not have enough capacity to support the RAID 1 volume. Please insert a hard drive with more capacity."
3. From windows pc browsed to mapped drive and found more folders being rebuilt. Quick assessment it appears like "everything else came back" however I haven't truly verified. I could dig deeper - used space versus free space formula.
Option 1: Potential new steps:
1. Shut down device
2. get my dad to remove dead hard drive
3. Turn on device
4. Inspect pst file to see if it's fully there ( why wouldn't it be ?)
Option 2:
1. Shut down device
2. Get my dad to remove dead hard drive
3. Insert replacement hard drive of exact model if possible if not choose same manufacture but with more GB.
4. Turn on Device.
5. Log into server
6. Choose auto rebuild and make it happen - somehow not choosing delete both and make raid 1
What do you guys think? I'll go read sticky.
BTW - > With this exact model type another customer of mine got the alert of dead hard drive and I went in and choose option 2 and it rebuilt perfectly. So I know the product works.
-
Correct your Option 2 will work if one of the hard drives is dead. If on the other hand the the drive is not completely dead (some sectors may be trashed but not all) you will see the drive reporting a smaller capacity as you are.
One thing you should do is to get your father to remove the failed disk and restart the unit. You then should be able to see what is on the remaining disk and if it has all the data there and it is accessible and the full capacity is shown then it should rebuild the array correctly when a new disk is installed.
If it does not show full capacity and if nothing has been written to that drive there is a very good chance that you will be able to recover it all but it does require you to have physical access to the drive unless you are certain that your father can do the necessary work on his own.
If ther is any doubt I recommend that you need to try and do the recovery first mainly because we see and have to fix the sort of problems toy are seeing and I always feel that we should have as much of a clients data stored safely should anything go wrong with the RAID rebuild.
The decision is now yours - do you try for a rebuild with a new same size disk or do you try recovery first.
-
I will follow the steps of turn off and have him remove the bad drive.
Turn it on then we'll see if the file is 4.5 gb somehow.
The 60 gb disk size was only there the day I first looked at it. Upon restart it shows the 2TB free space and what appears to be all the data. However lets try breaking the raid by removing the drive and turning it on.
Quick question, when looking at the unit with the front door in front of you, the left drive is the left drive and the right drive is the right drive?
-
To answer your question yes, and if the drive has completely failed the LED in front of it will be a sort of pinkish colour and not the standard blue.
-
The drive LEDs are not a particularly reliable indicator of drive condition or status, and, based on D-Link's documentation, a failed drive will be shown by an amber/yellow LED - experience has shown that with a failed drive you can have no LED, a blue, yellow or pink/purple LED - depending on the nature of the failure.
The LED actually has two sections, one blue, one amber - the pink/purple color is produced by running both sections simultaneously, the actual color seen will depend on the duty cycle - this condition is not documented by D-Link and is NOT indicative of a failed drive as it can be caused by a "full drive".
-
Thanks fordem for the information on the colours since we have never seen anything other than blue here (no failed drives in the NAS) and our servers use individual LEDs to indicate status of the individual drive.
If I remember correctly you have first hand experience with this and I hope everything is working well for you now.
-
I do have some personal experience garnered over the roughly six years I owned a DNS-323, it's still sitting on a shelf behind me, even though it no longer boots (either corrupt firmware or failed hardware) - I did have at least one, "live" disk failure, but before actually deploying the unit, I spent a number of weeks, simulating various scenarios, which included inserting "known defective" disks, to see if the unit would detect them as defective (it rarely did) and also "hot unplugging" disks to test RAID rebuilds.
Alas, those tests showed the unit was not ready for "prime time", which disappointed me immensely as I had a number of SOHO clients who could have benefitted from it.
-
Thanks Ivan, I will be getting my dad to remove drive when it is off. Turn back on and we'll see if the .pst file grows in size. I have a feeling it won't but one never knows.
-
Update:
Had the unit off for awhile, he turned it on as he needed a file.
the system started rebuilding, I checked mapped drive(only two folders there) + small pst file there
Systems says sync'ing. So this is a bit new to me but must of happened previously as this "feels" just the same as a couple times ago.
Next step finish of sync, turn off and remove the broken hard drive.
-
which drive is this: HDD2 Hard Drive Has Failed ?
-
HDD2 is the drive in the left bay when facing the unit.
-
joejiz, don't be at all surprised if you find most of your data missing when you remove and replace the defective drive.
This is at least the second time the drive has been rebuilt with a defective drive in the unit. In such cases it is time to put in a new drive and do a full format of the RAID 1 array and restore from backup.
Anyway, get that damaged drive out of the unit before you loose everything on the good drive.
As long as the damaged drive stays in the unit and tries to rebuild you are losing data!
-
Ok.
I will shut unit down, have him pull the left drive (hdd2 facing it).
Turn it on and see if the file is back.
Will report back after.
-
Ok, completed task mentioned above and two folders show , then in next 3 mins the rest of the folders show and the .pst file is still not the right size. ( just like before ).
So all I can really do is shut it down, wait till new hard arrives at house. turn device off and put in drive. Turn it on, connect to web answer questions - It should resync with out formatting all data and maybe file size increases?
The unit appears to be recovering from the raid problem but why this one file is cut down from 4.5 gb to 100 mb just has me stumped. It's like my dad left outlook open when the crash happened or something.
Or it has nothing to do with this crash, gosh it didn't get archived, I can tell you that!?
Are you stumped too? ??? Ever heard of this?
-
Backup whatever data you can before replacing the drive - strange things have been known to happen.
-
There are many reasons that the file is smaller than what you expect.
1) It could be that it was being copied to the array when the disk failed and so ended up truncated.
2) Because of the several rebuild attempts the incorrect size has been reported.
The above are the most likely but there could be others.
Download a copy of Kernel Outlook PST Viewer and use that to see what you have in the PST file. That way you will know if you need to delete it and start again.
The other very important thing is to follow fordem's advice and backup everything you can as soon as possible. That way you don't lose anything else and, should the need arise, you can do a complete reformat and restore (which is what I recommend because of the uncertain state of what remains of the data on the good disk).
-
Update:
Hi, well after leaving my dad's computer on last night while the folders starting appears my dad called me today and said ," all but 3 folders are remaining , the rest disappeared " Back to day 1 when he first called me.
I went into his pc to mapped drive the drive reports as 60gb. But this is the good drive? Something is very strange. If this drive was good then this wouldn't of happened. I think both drives might be messed up?
The unit isn't responding to the web server request either. I'm going to have my dad turn the unit off and turn it back on. I hope the unit brings all the folders back so we can copy off data.
I realize the rebuild and break thing has happened several times now, I know I need to copy all data off ASAP and then when new drive arrives , erase both drives, recopy data, re-do raid1. There is also firmware v1.10 to update as well.
Thanks for your help on this, I know the product is out of warranty and this is best effort, I really appreciate checking back on this thread and you helping out.
I'll keep you posted on what happens next.
Part of me thinks once we get the data off the drive (praying) - which I should of done as soon as I saw all folders - we'll turn it off and insert the broken drive back in and remove the RIGHT DRIVE and have it boot. See if that drive has the larger file on it. (what have we got to lose at that point?)
Crazy. I think the unit has recovered to many times from power outages and done the rebuild to many times that the raid 1 ability combined with a flakey drive(s) has caused this!
Talk about fun times, once the data is off I will investigate the PST file utility you mentioned- THANK YOU AGAIN FOR ALL YOUR HELP...it means a lot.
-
This is a long shot and will depend on you having enough space somewhere local to store a n image of the good drive.
Rather than try and copy files from the drive it would be a better idea to make a sector by sector image of the drive first. If you have an image you can recreate the drive as many times as necessary to try and get as much of your data from it as possible. It also gives you something to work on directly rather than at a distance. It should also work on the faulty drive if it is possible to read from it but they MUST be in the NAS as single drives, NOT TOGETHER.
Good luck and let us know how you get on.
-
Thanks Ivan, oh I see use that software and plug the drive into another working computer. I may do that after we get the files off if we can. we are almost ready to turn it back on.
-
Update: Turned on device, 799 gb copying off as we speak. 9mb/sec to a pc's external USB drive. eta 23 hours but I suspect hopefully tomorrow morning.
-
8 hrs remaining in copy 9.83 mb/sec
-
Finished copying. Next step is remove old hdd(one that was shown as working since day 1)
Put defective one in for kicks in it's previous bay(2)Right hand
boot it
See if pst file happens to come back - doubtful
Will report back.
-
Did the above and drive came up no problem but pst still small.
Gonna try your utility for pst and point it to shared drive and see if really is bigger than it is.
doubtful.
Will report back
-
Well from the windows 7 machine I ran the scanpst.exe file on the pst file in question.
It didn't expand as expected.
I went to another machine on the network (xp) and viewed the folder with the file and sure enough it was showing 3.5gb and a date on it from when this whole fiasco happened.
So back on the windows 7 machine it didn't show it more than 100mb. On the xp machine I copied the 3.5gb file to another folder and then looked at it from the win 7 machine . it shows 3.5gb.
So something happened to the folder that houses the it and somehow windows 7 is reporting wrong file size. The heck I know.
I'm just glad it appears we have the mail back.
Next steps:
1 - label the broken hdd (2) as bad, get my dad to mail it to me.
2 - Put new hdd in.
3 - upload firmware 1.10
4 - Format whole thing RAID 1
5 - Copy back all files to device.
Call it a day.
-
Once you get the new RAID 1formatted and ready the long tedious job begins - checking and restoring your data.
As you have found there are problems caused by a bad disk and attempted recovery of the array. To be certain that all the data you have recovered is 'good' will require you recreate the folder structure of the original array on the new array and do a file by file (or group of files) copy back using the command line rather than something like windows explorer (the command line will flag bad files, explorer will not).
If you had the backup disk a better way of remote restoring would be to use FTP by activating the FTP server on the DNS-323 and using something like FileZilla client at your end. With careful setup of parameters - like a set number of tries to read a file from the backup - you should be able to transfer most, if not all, files in a folder at a time (I would still not try to transfer the folder structure because you have shown there are problems with it).
It should be possible to use FTP at your father's end but I think it will require you to talk him through what he would need to do.
-
Yah when I was copying files using Windows explorer an error came out about to long of a file name for copying so I had to skip 320 files. My dad made file names to long I guess. 256 character limit with folders and file names is the max I THINK. What can one do? Not even sure what folder that was in. We may just have to abandon those files. File system limitation I guess from Linux to NTFS.
I am able to RDP to my dad's network and control dlink and his win 7 pc and window xp machine.
I will format the whole array so it's blank after I upgrade firmware(need to download, hopefully sticky has link).
After that I was going to use Hoffman rich copy but I do like your idea of enabling FTP, creating an account and using file zilla to copy back all the info from USB DRIVE we backed up onto using THE WINDOWS XP machine back on to DLINK.
I would hazard to "guess" the only file messed was his pst file because he left it open during the fail.
The finish line is near.
I guess I should be running scandisk once a year for safe measure.
-
Just wondering how the saga is going.
-
Sorry for delay, we were able to use the pst and run from the bad drive.
tonight I am removing bad drive.
inserting old (good drive)
inserting new drive
updating firmware
making new raid 1 format
then restore data using miscorsoft rich copy
-
updated firmware to v1.10
put new hard drive in left bay
put old good hdd in right bay
Turned on
unit prompted me to format new drive and sync
I did this.
Unit put out a log saying left hard drive dead which was odd as it brand new
Rebuild completed. so far been working ok.
I haven't formatted the raid and copied back the 1 TB of data yet.
-
so far working ok without reformat.
-
If it rebuilt the RAID array it would automatically do the format as part of the process.
The fact that everything is working points to success.