woreilly Posted February 11, 2014 Posted February 11, 2014 Firstly, apologies if I've got the wrong section. I look after the ICT needs of a middle school and recently, 1 of the hard drives on their DC failed and when I was reporting it to Dell Tech Support they noticed 3 more drives failing - but not dead (replacement drives have been ordered and are on their way). I am told that all I need to do is simply plug the new HDDs (in a certain order) into the slots where the faulty ones come out and the disks/drives will automatically rebuild. What I want to know, is how does the server know what to write to those disks? I am a relative newbie to the world of ICT Technicians so I am still learning the theory behind RAID configurations and different types of volumes i.e. striped, spanned and mirrored etc. If you need the info to help explain it to me, they are rocking 8 x 500GB HDDS. The first 2 are running the OS and are RAID-1 (these 2 are part of the 4 'failed drives') and the other 6 are on RAID-6 with two of those 'failing'. Also the server is running 2008R2. Help me understand so I don't stress as much as I did when the techie told me I had 4 failed drives on my server **Needless to say I have made a full backup of their system before I replace any HDDs on the server**
sted Posted February 12, 2014 Posted February 12, 2014 how dosent really matter but the raid controller knows thats its a new drive inserted and copies the data to it the the drive in that slot should contain. Id guess that you need to first replace the dead one let it FINISH rebuilding then one by one replace the iffy ones
Oaktech Posted February 12, 2014 Posted February 12, 2014 Depends how the raid is set up, and the brand of raid card as to exactly how it happens... RAID - Wikipedia, the free encyclopedia
PotNoodleTech Posted February 12, 2014 Posted February 12, 2014 (edited) Basically the RAID card in the server can detect a replacement drive and automatically start rebuulding the array (rebuilding the array means, copying data on to the new drive and ensuring that there is no data corruption). Make absoloutly sure that you only replace one disk at a time and verify the array is consistant before you move on to the next disk. Be careful, take it slow. It will probaly take you several days to replace multiple drives so don't rush it. If you replaced a second drive before the array is finished building onto the first drive, you will [probably] lose the entire array and all data. My advice would be - Tackle the RAID1 array first. Replace the failed drive first, let the array rebuild and perfrom a consistancy check. Then tackle the RAID 6 array. Replace one at a time and make sure the array is consistant before you move on. I would reccomend replacing all drives. If 4 of the 8 have already failed the other 4 are not to be trusted either. What make and model of drives are they? Edited February 12, 2014 by PotNoodleTech
cogrady84 Posted February 28, 2014 Posted February 28, 2014 (edited) The first 2 are running the OS and are RAID-1 (these 2 are part of the 4 'failed drives') Sorry i'm a little late to the party, but if both of your mirrored drives are failed, that container is surely destroyed? RAID-1 can only sustain a single drive failure. Edited February 28, 2014 by cogrady84
cpjitservices Posted February 28, 2014 Posted February 28, 2014 (edited) Rebuild meaning, the RAID Array will rebuild... not the actual running system of the server. RAID has its own memory and will rebuild itself when new disks are installed as it's knows what configuration to use. Correct me if I'm wrong but won't @woreilly need to reinstall his OS if he's replacing the main OS drive if it's failed ?? Edited February 28, 2014 by cpjitservices
sparkeh Posted February 28, 2014 Posted February 28, 2014 (edited) Ok I'll attempt a basic explanation of how your RAID 6 array can rebuild itself. Lets start with RAID 5 as its slightly less complex. RAID 5 uses parity data to achieve redundancy. So using the diagram below you can see that the disks in the RAID 5 array all contain 3 blocks of data plus an extra one with the parity data (AP, BP, CP,DP). So, if Disk 0 fails we need to recover the data labelled A1, B2 and C1. If you replace Disk 0 then the RAID controller can use the parity data stored in AP, BP, CP to reconstruct the missing data. With RAID 5 you can only lose one disk as if you lose more then you can't reconstruct the data. However you are using RAID 6 So you can see that each disk holds 2 sets of parity data meaning that you can afford to lose 2 disk Hope that helps? Edited February 28, 2014 by sparkeh
cogrady84 Posted February 28, 2014 Posted February 28, 2014 Rebuild meaning, the RAID Array will rebuild... not the actual running system of the server. RAID has its own memory and will rebuild itself when new disks are installed as it's knows what configuration to use. Correct me if I'm wrong but won't @woreilly need to reinstall his OS if he's replacing the main OS drive if it's failed ?? This is what i'm getting at, if both of his RAID-1 drives have failed, he has lost all data in that container, so yes OS reinstall will be necessary, or restore from backup. 1
zag Posted February 28, 2014 Posted February 28, 2014 8 drives!!! no wonder 1 has failed! Crazy server design, especially on a Domain Controller.
sparkeh Posted February 28, 2014 Posted February 28, 2014 (edited) Also, I always recommend having a 'hot spare' drive so that if a drive fails the spare steps in and the array is rebuilt using this disk until the failed disk is replaced. Was a Godsend when we were having an issue with regular drive failures. Edited February 28, 2014 by sparkeh
cpjitservices Posted February 28, 2014 Posted February 28, 2014 I thought so too. This is what i'm getting at, if both of his RAID-1 drives have failed, he has lost all data in that container, so yes OS reinstall will be necessary, or restore from backup.
tmcd35 Posted February 28, 2014 Posted February 28, 2014 8 drives!!! no wonder 1 has failed! Crazy server design, especially on a Domain Controller. Not really. Looks like a pretty usual/sensible design to me for a larger primary/smaller secondary that perhaps only has 1 or 2 servers (I'm guessing). Certainly not out of the ordinary and the combination 2 drive RAID-1 for OS and 6 drive RIAD-6 for data seems like a good choice. I'm guessing from the date that the OP has resolved this problem by now (or his HT is pacing outside the OP's office?). I think he said 1 drive had failed and 3 where "looking dodgy". That being the case he hopefully got the RAID-1 drives swapped out before both failed completely thus not needing to reinstall his OS? The RAID-6 should survive both it's failed/failing drives going offline, thankfully.
sparkeh Posted February 28, 2014 Posted February 28, 2014 @sparkeh - what about raid 10? Well, RAID 10 (or RAID 1+0 or RAID 1&0) is basically an array of mirrors: Never sure if you can really call it a standard RAID level as it kind two RAID levels cobbled together.
woreilly Posted March 3, 2014 Author Posted March 3, 2014 Sorry guys, only just seeing your responses. In regards to the notion of the OS needing to be re-installed, I didn't have to. As far as I'm concerned with the RAID-1, there was one failed disc and the other was simply "poorly" or "ill" - it hadn't actually failed. The man from Dell Tech Support diagnosed it as "failed" even though it was still working and thus recommended its replacement. I think his diagnosis of the other two (HDD 6 & 7) proved accurate when HDD7 failed during the rebuild of HDD6. LUCKILY, I have suffered no data loss nor have I had to rebuild the OS on the server
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now