What Happens if a drive in a RAID-1 Fails?

Status
Not open for further replies.
G

Guest

Guest
Hello everyone. In my system I have my an SSD as a boot drive and I also have two 2TB HDD's for storage. The other day I was thinking and I came to the conclusion that I'm not actually using more than about 1TB and I would be better off putting the drives into a RAID-1 to improve the security of my data. I'm a bit of a noob when it comes to things like RAID and I need all the help I can get. What I would like to ask is what happens if one of the drives in the array fails? Will I just be able to boot up and access the data on the working drive as normal? Another question I have is how would I go about rebuilding the array with a replacement drive?

Thank you all in advance for your help I really appreciate it.
 
Yes, that's the purpose of RAID-1. When a drive in RAID-1 fails the raid enters "rebuild mode". When the failed drive is replaced it will automatically start cloning the data from the intact disk.

"how" you rebuild it is entirely dependent on the raid controller. Most of the time it's automatic and only involves hotswapping the failed drive out for a healthy one.
 

master_chen

Honorable
Jun 20, 2012
1,215
0
11,360
In RAID-1 all data is duplicated for both HDDs, so if you lose one HDD, all data would be safe as long as at least one HDD in the RAID-1 is alive and healthy. This also applies to RAID-6.
 
G

Guest

Guest
Thanks everyone for your responses. So, If one drive failed I would still be able to access my files on the other one, I wouldn't have to wait until I can get my hands on a replacement to rebuild the array? Also I don't have a dedicated RAID controller I was just planning on using the on board one on the motherboard. In this case will it still automatically rebuild when the failed drive is replaced or will I have to do something to make it rebuild the array with the new drive?
 
You would have to replace the failed drive :)

And just a heads-up ... some drives (like inexpensive big xTB storage drives) simply don't RAID well.

You would be better off simply backing-up to an external drive and storing it in a good safe place.

 
G

Guest

Guest
The HDD's are WD Caviar Blacks. You say these aren't very good for RAID-1?
 


The Intel raid controller is actually a firmware raid controller, it's not integrated. This means that it's a software raid controller with an OPROM that allows it to be booted from and powerful AHCI drivers which do all the work. It's not a replacement for a solid 3Ware or LSI raid controller but it gets the job done quite well for what it's designed for.
 

briantg

Honorable
Aug 20, 2012
1
0
10,510
It's not that big capacity drives will run poorly in a RAID setting... it's the "Green" style drives that will sleep the motor on inactivity that will perform poorly. The RAID (whether software or hardware) controller will have certain expected behavior for drive latency, and if a drive is operating outside the bounds of normalcy it may decide to remove it from the RAID set (thinking it's bad). When a green drive decides to take a little nap because nobody is writing/reading data from it, it does so without notifying the RAID controller. So, next time it is accessed it will take 5-10S for the motor to spin all the way up.. which may be unacceptable behavior from the perspective of the controller. The command "times out" and the drive is yanked out of the set, even though it is perfectly healthy, just a little sleepy.

Setting up a RAID-1 is a good idea, just make dang sure to set up notification. The RAID controller should either be sounding a physical alarm (if anyone is within earshot) or emailing you when a drive is taking offline. If nobody notices a drive is taken out of the set and nobody replaces it within a timely manner, it just gives you a false sense of security. Also remember RAID-1 isn't a backup of any kind... and it won't protect you from electrical events that could fry both drives, viruses, theft, human error or malicious behavior, etc.
 
Thanks everyone for your responses. So, If one drive failed I would still be able to access my files on the other one, I wouldn't have to wait until I can get my hands on a replacement to rebuild the array? Also I don't have a dedicated RAID controller I was just planning on using the on board one on the motherboard. In this case will it still automatically rebuild when the failed drive is replaced or will I have to do something to make it rebuild the array with the new drive?

Yes, you could still use the raid drive with one failed drive. It will continue to run until a replacement drive is inserted at which time the array will be rebuilt.

NOTE: RAID is not a subtitute for backups. If you accidently delete a file or get a virus, it will effect all drives in the array.
 
I don't think you need to get that complicated. Use some syncing software like All-Way Sync or Sync Toy and just set up jobs for important directories to sync across the drives every night. This way you also don't waste space making files that you don't care about redundant and don't have to worry about controllers. No matter what happens to one drive or the PC you can just take one of the drives out, pop it in any old machine or external enclosure and get your data.

Better yet, put one of the drives in another machine or external enclosure and store it away from your main machine so that if something happens to the machine there is no risk to your data.
 
G

Guest

Guest
My reasoning behind this is that i'm going to be doing a fair amount of video editing and while I will obviously make regular backups to an external HDD, with the nature of video editing these backups can become outdated very quickly. So, if I have a RAID-1 array then I am safe in the knowledge that if one of my hard drives goes down then I have a fully up to date backup on the other. If something does happen that affects both drives (e.g a power surge or a virus) then i'll still have the backup on my external HDD.

Thanks everyone for all your help and advice it's greatly appreciated :)
 
I just wonder what advantage you are going to get from a RAID1 that you can't get from a data sync type of scenario. If you are using it as a straight backup in the event of hardware failure during use then it's fine.

What happens if you edit a file one day, and then the next day do something to it that you can't undo? In the RAID Scenario you now have to roll back to the offsite while on the sync scenario you roll back to the night before. What about if you delete a file you weren't supposed to? What happens when your controller decides your array is unusable for some unknown reason and your RAID won't rebuild . . . these things happen.

Each way of doing it has some advantages and disadvantages. I just think you are going to be more likely the cause of a data recovery need over a failed hard drive, and the sync solution covers the drive failure nearly as well while also protecting from user edit fails at the same time where the RAID does not. The sync solution also has the added bonus of being able to put either of the drives in to any machine or enclosure and having immediate access to your data, with a RAID that kind of stuff gets pretty complicated (Though I've only ever had to move a RAID5, RAID1 might be a little easier in this regard). It does deserve some thought as a possible solution to your problems.
 

Kit McGuire

Reputable
Apr 20, 2015
6
0
4,510
Quick question as it relates here: when I pull a defective HDD (RAID system) is the new HDD inserted where the defective drive once resided? Or do I need to swap HDD positions? A simple out with the old and replace with the new in the exact same slot would be nice.
 

USAFRet

Titan
Moderator
Quick question as it relates here: when I pull a defective HDD (RAID system) is the new HDD inserted where the defective drive once resided? Or do I need to swap HDD positions? A simple out with the old and replace with the new in the exact same slot would be nice.
Coming from an 8 year old thread...
You out the replacement drive in the same place where the old bad one was.
 
Status
Not open for further replies.