RAID 1 failure and confusion about how to recover..??

[angelleye]

In this setup I have a RAID 1 array setup as a storage drive. My system drive is a separate drive.

I am getting a Failed status on the RAID, and the drive is not detected at all by Windows at this point.

The company I bought the computer from (www.avadirect.com) is supposed to offer tech support, however, they are giving me very conflicting information from what I thought I understood about RAID1.

I expected to be able to keep running from the good drive, get the other drive replaced, and when I pop in the new drive it would rebuild and I'd be back to how I was originally. They are telling me this is not the case, and RAID 1 will not load properly without both drives in place.

Is there really no way for me to simply run off the good drive? I thought that was one of the benefits of RAID 1..?? If a drive fails, you keep working on the good one and then rebuild the array when you replace the bad drive. No..??

I see this option to "Reset Disks to Non-RAID". Will that format the drives or will that just leave them as they are but load them up as separate disks so I could then easily see which one is the bad one using disk manager..?? As it stands now I don't even know which one is the bad one.

In the RAID BIOS I see RAID Volume 0 has a status of Failed, and then it shows Physical Devices below that. It shows 1 of the drives associated with my RAID array, and then it shows another drive that is not associated with my RAID. It does not show the second RAID drive, though.

The thing is, the one that does show up there says "Error Occurred (0)" in red next to it. So is that the good drive (since it's at least showing up in BIOS) or is that the bad drive (since it says error next to it)?

If I just pull the bad drive would things load up like usual as a single, stand-alone drive?

I'm completely lost here so I'd appreciate any information I can get on this. Thanks!

 

[junkeymonkey]

First, if you lose the RAID card in theory you should be able to take either one of the drives in the mirrored set, cable it to a standard IDE port, point the BIOS to that drive if bootable and you should be in business. However many times this is not the case for two main reasons.

http://dtidatarecovery.com/converting-raid-1-to-single-drive/

looks like you got that now out of the way

none of this is covered in your boards raid manual ?


http://superuser.com/questions/677564/can-a-raid-1-be-created-without-losing-data

http://www.howtogeek.com/howto/36504/how-to-create-a-software-raid-array-in-windows-7/

thing is if your rais filed due to something out side just one drive failed that should of just been a hot swap with a new drive theres not telling ?

its not like were there to see for ourselves or hands on to know the full story all we can do is help guess so to say

 

[FireWire2]

Ok, so now with the smart error drive unplugged, the system boots with RAID Status Degraded, and the single drive that shows up there shows Member Disk 0 in green. As such, now when I boot into Windows everything loads and my data is indeed there and accessible on the single, working drive.

So at this point I should be able to pull the bad one, get it replaced, then simply plug in the new one when it gets here, and go into RAID bios to choose rebuild, right?

Your post is VERY confusing
1st " so now with the smart error drive unplugged, the system boots ..." , then " So at this point I should be able to pull the bad one, get it replaced" ?!!!! Does the RAID run with ONE drive?

Base on the info, my guess there is ONLY one drive in RAID1, correct? If so, now go and get the new HDD and just plug it in,
It should rebuild, the RAID engine of your MB need to initialize the HDD before it can rebuild - look at the MB manual it will tell you

 

[USAFRet]

This whole thread is a testament to needing an actual backup behind a RAID of any kind.
Or behind any data you feel is critical.

With a viable backup, this could have been fixed in an afternoon.
Recreate the RAID array with blank drives. Resurrect the data from your backup. Done.

 

[FireWire2]

This whole thread is a testament to needing an actual backup behind a RAID of any kind.
Or behind any data you feel is critical.

With a viable backup, this could have been fixed in an afternoon.
Recreate the RAID array with blank drives. Resurrect the data from your backup. Done.

It's a debate-able topic - He has back up, but just NOT understand nor use correct controller RAID controller

In RAID1, 10, 5, 6 - NEVER remove a HDD when power off, HOT SWAP it - That is all to it

 

[angelleye]

Your post is VERY confusing
1st " so now with the smart error drive unplugged, the system boots ..." , then " So at this point I should be able to pull the bad one, get it replaced" ?!!!! Does the RAID run with ONE drive?

Base on the info, my guess there is ONLY one drive in RAID1, correct? If so, now go and get the new HDD and just plug it in,
It should rebuild, the RAID engine of your MB need to initialize the HDD before it can rebuild - look at the MB manual it will tell you

Yes, this whole thing has been very confusing for me, too.

The drive that was showing the smart error is apparently the bad one. I can't get the RAID to to load at all with only that drive plugged in.

With that drive pulled, though, when I start the machine it shows a failed RAID array, but if I go into RAID BIOS and choose to rebuild, it switches to Degraded. Then when I continue and let Windows load from there, the drive is accessible, the data is there, and I was even able to then install things like Google Drive, which I use this storage drive to sync to along with all the other data actually stored on it, and I got that all hooked up no problem pointing to the path on the RAID drive...with just the one drive running.

However, at this point if I actually start trying to read/write to the drive, it hangs and eventually starts giving me messages that the drive is corrupt and can't be accessed. I'm hoping that's just because it's looking for the other drive, not finding it, and getting confused, but then when I get the replacement and pop it in there, rebuild with both, then hopefully things will be good, so I'm just not touching it until I get the new drive.

 

[angelleye]

This whole thread is a testament to needing an actual backup behind a RAID of any kind.
Or behind any data you feel is critical.

With a viable backup, this could have been fixed in an afternoon.
Recreate the RAID array with blank drives. Resurrect the data from your backup. Done.

I do have other backups. That is not the issue.

The issue is that there is apparently no use what-so-ever for RAID 1 if I'm going to have to restore from my regular backup on a blank array just like I would with a single drive (and did for many years).

The whole point was, ok great, I'll do RAID 1, I'll have 2 identical drives, and if one happens to fail, no problem. Just get a new one, work from the good one while you wait on the new one, and when the new one comes you pop it in, the array rebuilds, and now you're right back where you were. At least that's what I thought.

What I'm learning is that running from the single drive in Degraded mode while waiting on the new drive isn't working out as planned. It does load, I do see the drive, I can see all the data there on it and browse it without an issue, it just doesn't really let me work from it in this status.

So again, I'm waiting on the RMA replacement now, which of course the holiday didn't help with that, and as soon as I get it I'm hoping to pop it in and be rocking shortly thereafter. I'd have to wait on that anyway, so no, this wouldn't have been fixed in an afternoon regardless. We're also talking about nearly 3 TB of data, so it would be a lot quicker once I do have the replacement if the raid works out for me.

If I get the new drive, pop it in, it rebuilds and everything is good I'll be happy. I can deal with waiting on a new drive (or even keeping a spare on hand) to let it rebuild and be back to normal.

If I get the new drive, though, and the rebuild simply fails, or even works but then the data is all jacked up or still lags and causes problems, then I'll have to wipe them and restore from my regular backups as if I didn't have RAID at all.

So then the question will be is RAID 1 just a complete joke? No, I wouldn't think so. Did I do something wrong? All I did (well all they did when I bought the machine) is use the built in settings to create the raid 1. Very simple and straight forward. I won't really know if I'm at that point until I get the new drive and try the rebuild.

All of that said, I would love ideas on how else to keep backups of 2.5 - 3.0 TB of data. Right now I have it on Carbonite remote, but to restore that I have to pull down all that data. Not a huge deal, but not ideal. I also use Windows File History on an external drive here, but that's a little bit finicky at times and takes a while to restore that much data. I also use Google Drive for some of the most important stuff, but of course I'd rather not have to pay them $100/mo for the space there.

Anyway, I'll update again when I get the new drive and try the rebuild.

 

[USAFRet]

There IS a use for a RAID 1.
When you absolutely need 24/7 uninterrupted operation. Even if a bit degraded. For instance, running a webstore. Down time = lost sales.
A RAID 1 would allow the system to limp along until you can schedule the downtime needed to get it all back and running well.

It is NOT a backup for general use.

How to create a residential backup plan for your house? 2-3TB? No problem.

Here's mine:

2 PC's
System 1 - my main PC.
4 x SSD + 1 x 3TB HDD
4.5TB total drive space.

System 2 - HTPC/house server
1 x SSD (C drive)
5 x HDD. Movies, music, etc, etc
12TB drive space

OS and applications:
Using Macrium Reflect, the C drive (250GB SSD) is imaged every night at 2AM to a folder on the 3TB spinny.
Keep 2 weeks worth.
Every Sunday, another image to a drive and folder on the other PC. Keep 4 weeks worth.
With these, I can resurrect the OS and applications from almost any point in the last month.
A full image of the C drive takes about 20 mins. The C is currently ~150GB used, the Macrium image ends up at about 85GB each.

The other PC (HTPC/house server) does exactly the same in reverse.

Automated and happens in the background.

Data files:
Using SyncBackFree, a somewhat similar schedule to the OS drive.
Since this is 'just files', nothing fancy needed. And I can be as particular as needed as to what gets copied over.
This also includes the Steam games that live on a secondary drive in System 1.
And, SyncBackFree can be set to do only new or changed files. No need to do the whole thing every time. So after the initial setup, it is very fast.
Automated and happens in the background.

Anything that can't be easily replaced is on at least 2 drives, usually 3.

So, in the event of a C drive fail, boot from the Macrium rescue DVD and apply the chosen image to a new drive.
For fail of any other drive, slot in a new one, and just copy the data back over from the relevant copied location.

If the 3TB HDD drive used for backup fails...well, that's why it is a 'backup'. A second independent copy. Toss the drive, put in a new one. The automatic drive imaging and copying continues again at 2AM.

This also points to the benefit of different drives for different purposes.
In my main:
250GB SSD = C drive. OS and applications
250GB SSD = photo and video work
960GB SSD = games, doc, video, etc
120GB SSD = scratch for Adobe Lightroom and video applications.
3TB HDD = the main backup drive.

 

[FireWire2]

Your post is VERY confusing
1st " so now with the smart error drive unplugged, the system boots ..." , then " So at this point I should be able to pull the bad one, get it replaced" ?!!!! Does the RAID run with ONE drive?

Base on the info, my guess there is ONLY one drive in RAID1, correct? If so, now go and get the new HDD and just plug it in,
It should rebuild, the RAID engine of your MB need to initialize the HDD before it can rebuild - look at the MB manual it will tell you

Yes, this whole thing has been very confusing for me, too.

The drive that was showing the smart error is apparently the bad one. I can't get the RAID to to load at all with only that drive plugged in.

With that drive pulled, though, when I start the machine it shows a failed RAID array, but if I go into RAID BIOS and choose to rebuild, it switches to Degraded. Then when I continue and let Windows load from there, the drive is accessible, the data is there, and I was even able to then install things like Google Drive, which I use this storage drive to sync to along with all the other data actually stored on it, and I got that all hooked up no problem pointing to the path on the RAID drive...with just the one drive running.

However, at this point if I actually start trying to read/write to the drive, it hangs and eventually starts giving me messages that the drive is corrupt and can't be accessed. I'm hoping that's just because it's looking for the other drive, not finding it, and getting confused, but then when I get the replacement and pop it in there, rebuild with both, then hopefully things will be good, so I'm just not touching it until I get the new drive.

There is a misconception about RAID engine, RAID from Intel, AMD, ASMedia, Jmicron, Marvell...they have different way to handle failed drives, most of Motherboard RAID is pretty BAD of handling failed HDD, where RAID specialist - Raid card maker - not Mobo - know what to do with certain failure mode, most of the time

Let face it, RAID with Mobo were that good, there wont be Adaptec, LSI, Areca, Highpoint, Atto....

As I stated in many posts, if you are serious about using RAID, consider a true hardware RAID card. particular in RAID0/1 they are not that expensive.

Use a drivers-less RAID controllers like
http://www.datoptic.com/ec/host-controllers/driver-less-hardware-raid-controllers.html
https://www.startech.com/HDD/Adapters/sata-dual-hard-drive-raid-adapter~S322SAT3R

This will ensure you can move or upgrade OS w/o ANY headache

 

[mjslakeridge]

Man, reading through this thread confirms to me that RAID is not for the casual home setup (not that I was ever inclined to try it). I have read that setting it up is the easy part, recovering from a drive failure is where the problems can come in, and that is certainly confirmed here. The other area of concern with RAID is that if you mess up a file (say an Excel spreadsheet) and save, it is messed up on both drives, there is no backup to go to. (I know this is not the case in the OP's case as he has a backup), but I would bet a lot of people don't understand this aspect.

 

[angelleye]

Man, reading through this thread confirms to me that RAID is not for the casual home setup (not that I was ever inclined to try it). I have read that setting it up is the easy part, recovering from a drive failure is where the problems can come in, and that is certainly confirmed here. The other area of concern with RAID is that if you mess up a file (say an Excel spreadsheet) and save, it is messed up on both drives, there is no backup to go to. (I know this is not the case in the OP's case as he has a backup), but I would bet a lot of people don't understand this aspect.

Yes, it's very important to understand that this configuration is not to be considered a primary backup. The whole idea was zero down time if a drive should happen to fail. That's it. If the RAID does fail, if the system is physically stolen or damaged, etc. you would need a different backup. Always good to keep a remote backup of some kind, at least for the absolute most important stuff.

 

[angelleye]

There is a misconception about RAID engine, RAID from Intel, AMD, ASMedia, Jmicron, Marvell...they have different way to handle failed drives, most of Motherboard RAID is pretty BAD of handling failed HDD, where RAID specialist - Raid card maker - not Mobo - know what to do with certain failure mode, most of the time

Let face it, RAID with Mobo were that good, there wont be Adaptec, LSI, Areca, Highpoint, Atto....

As I stated in many posts, if you are serious about using RAID, consider a true hardware RAID card. particular in RAID0/1 they are not that expensive.

Use a drivers-less RAID controllers like
http://www.datoptic.com/ec/host-controllers/driver-less-hardware-raid-controllers.html
https://www.startech.com/HDD/Adapters/sata-dual-hard-drive-raid-adapter~S322SAT3R

This will ensure you can move or upgrade OS w/o ANY headache

If I use a RAID card will I be able to experience zero down-time when a drive fails like I had planned here?

That really does seem to be the case with the mobo config. It technically seems to be working, but it provides very little in the way of software to see what exactly happening with the drives, their status, etc. I have a Western Digital NAS drive that does RAID and provides much better management tools than what I have on this desktop at the moment.

So, if I get a nice RAID card, setup there, and in the future a drive fails, will it give me notifications that a drive is bad (I never even knew this current problem was happening until I rebooted and happened to watch the POST screen), tell me which one, and allow me to continue running without a performance hit from the good drive?

 

[angelleye]

For a general update here, I just got the new drive and popped it in. The RAID BIOS saw it and popped up a message telling me it sees a degraded array, and it also sees a new drive ready for rebuild, and asks me if I want to rebuild. I simply say Yes, and it puts it in rebuild mode, and tells me it will rebuild in the OS.

Ok, great. So now Windows is loaded, but I have no indication of where it's at in this process. I'd imagine it will take a little bit of time to rebuild a 3TB array and copy everything over..??

So now I'm worried that if I start reading / writing I'll screw something up, or if I reboot the machine when it's actually at 98% done with the rebuild and I just didn't know it I'll screw something up.

I guess I'll just let it sit for a few hours and risk the reboot later so that I can get back into RAID BIOS and confirm that it shows everything is good now..??

@FireWire2, again, is this another area where I would have software to help me see this sort of thing if I was using a RAID card instead of mobo?

 

[mjslakeridge]

I would let it go as long as your patience allows. Are you seeing the HDD activity led light up periodically or constantly?

 

[angelleye]

I would let it go as long as your patience allows. Are you seeing the HDD activity led light up periodically or constantly?

Ugh, well, it's not looking good so far.

I had walked away for a bit to let it sit and rebuild. When I saw you mention activity I figured I'd come back in and check the task manager / resource monitor real quick.

The OS now is extremely laggy. It took a long time to turn the displays back on. When I right click on the app bar and choose task manager it just never comes up. When I ctrl + alt + del it took about 30 seconds for that to pop up, but again, choosing task manager there won't actually open it.

I can open Windows Explorer and it sees the drive and shows all the top level folders, but when I try to open anything it tells me it's corrupt and inaccessible. I actually did see that message once before while running on the single drive, but then it worked fine again on a reboot after that.

So I guess at this point I'm going to let it sit for a while and hold out one last little tiny bit of hope that it's rebuilding right now which is causing all the lag and performance problems, and when it finishes maybe everything will kick back in.

If not, I guess I'll delete the raid array and probably just go back to using the 2 drives synced up manually (automated with software sync). Restore from my other backups and forget I ever tried all of this.

 

[mjslakeridge]

If it is rebuilding, and disk activity is at 100%, that would cause your system to be laggy. If the disk activity LED quits lighting up for a while, then I would suspect a problem, but give it time.

 

[angelleye]

If it is rebuilding, and disk activity is at 100%, that would cause your system to be laggy. If the disk activity LED quits lighting up for a while, then I would suspect a problem, but give it time.

The disk activity is 0%. I'm giving up on the RAID rebuild.

I have split the drives up onto separate SATA controllers now. The drive that was still good from the original RAID array has been moved to a separate SATA controller, and now that drive shows up in the system with all data accessible, so the data is not lost.

Looks like I'll just go back to the old way I always did things before.

 

[mjslakeridge]

Well, that is good that you have your data. Like I said earlier, rebuilding a RAID array seems to be the most problematic part of RAID. I use Acronis True Image Home to make periodic images of my OS drives. If I ever suspect a virus or some other instablility, it takes no longer than 15 minutes to restore from the image. The main thing is to have an established procedure for doing backups and the discipline to follow the procedure.

 

[USAFRet]

Well, that is good that you have your data. Like I said earlier, rebuilding a RAID array seems to be the most problematic part of RAID. I use Acronis True Image Home to make periodic images of my OS drives. If I ever suspect a virus or some other instablility, it takes no longer than 15 minutes to restore from the image. The main thing is to have an established procedure for doing backups and the discipline to follow the procedure.

I use Macrium Reflect for this.
Full image of the C drive, every night at 2AM.
Every Sunday at 3AM, to another PC.
All automated.

 

[angelleye]

That actually brings up another question I have about all of this. It seems like anytime I see people talking about making images they're just doing it with the system drive. I really don't care about my system drive, though. I can be up and running from clean install of Windows in a very short time. It's simply the storage data I'm worried about.

Can I make an image of a typical storage drive just like I could a system drive? And then if I restore the image to a fresh drive it would be setup exactly like my original storage drive, right?

 

[USAFRet]

That actually brings up another question I have about all of this. It seems like anytime I see people talking about making images they're just doing it with the system drive. I really don't care about my system drive, though. I can be up and running from clean install of Windows in a very short time. It's simply the storage data I'm worried about.

Can I make an image of a typical storage drive just like I could a system drive? And then if I restore the image to a fresh drive it would be setup exactly like my original storage drive, right?

Yes, it works exactly the same way for a storage drive.
But for mine, I use SyncBackFree
Folder and file copy utility.

Macrium Reflect to image the OS/application drive, SyncBackFree to copy files and folders from the storage drives.
I can be a bit more granular in exactly what to copy, rather than the entire drive.

At the scheduled time, it does a compare between source and target folders. Only that which is new or changed gets copied over. Not the whole thing every time.