Solved How involved is it in a Raid 1 array when one drive is going

Custom built pc / Ga-z97x-ud3h
February 10, 2015 at 14:34:16
Specs: Server 2003, i5-4690K
How involved is it to rebuild a Raid 1 array when one drive is going?
At work, our old system is on an older Windows 2003 Server with two drives in RAID. The system is used now only for customer order history since the newer system has been up for almost a year now. One of the drives apparently has been going and I was just told that the system has been complaining for a while about one drive. My thought is to purchase two new drives of about the same size, swap out the bad one and rebuild the array and then swap out the other one and rebuild it again. They are only 250GB SATA I drives and I found SATA II WD RE3 250GB drives that are quite reasonable that should work fine. Since I have not built or repaired a RAID set up yet, how involved is it? I found this article: http://redmondmag.com/articles/2014... --- Is this accurate? Is there anything specific I need to know or do? Do the drives need initializing or is this done automatically? Is there a better way? (The drives are SATA 150GB/s Maxtor DiamondMax10 and the motherboard is either SATA I based on age or possible SATA II which is why I looked for the SATA II 3GB/s drives for their backwards compatibility)

You have to be a little bit crazy to keep you from going insane.


See More: How involved is it in a Raid 1 array when one drive is going

Report •

✔ Best Answer
February 12, 2015 at 19:31:36
Do you know if the raid config is for a mirror, does it tell you the raid config? In any event, I would simply. Unless you must keep the machine running, I would shut down, remove the old one, put in the new one turn it on, and it should recognize the drive, and ask if you would like to rebuild. Once rebuilt and booting, make sure, then add ther other new one, rebuild again. You would still have both old drives to put back in, if there was "uh oh". The jazz about size was; the original mirror was created with as 2 x 250g ish. If say you could not find a 250g replacement hard drive, but instead used a 500g, it would still create a 250g mirror on it, effectively leaving half, the drive underutilized. Doesn't seem to matter in your case here, but that's how it works. If after putting in the drive you get a wonky prompt you are unsure of post back.

To err is human but to really screw things up, you need a computer!

message edited by HopperRox



#1
February 10, 2015 at 15:00:03
It should just be a question of pulling out the bad drive (make sure you get the right one!) and putting in the new one. If your RAID controller allows it you don't even need to take the system down to do. I used to have to swap out the disks in RAID arrays (RAID 5 in my case) 3 or 4 times a year (several arrays - not just one!) and never had to even reboot.

This is assuming that you are using a hardware RAID controller. I think the article that you linked to is talking about the software RAID built in to Windows Server.


Report •

#2
February 10, 2015 at 19:32:57
I assumed it would be hardware based RAID controller built in to the motherboard but I guess I will have to check. I only remember one expansion card which was an NIC card. I will note down the motherboard model tomorrow and look into its properties.
If it is hardware based and will rebuild itself (which is what I originally thought it would do), how would I verify that it is/has rebuilt the array? On a 250GB drive that is probably less than half full, assuming a SATA 150GB/s original drive, how long should it take and should we limit or restrict use during this period? Should I give it a day or more before swapping the remaining drive to have two newer matching drives (the drives are likely 8 or 9 years old)?
Thank You for your help.

You have to be a little bit crazy to keep you from going insane.


Report •

#3
February 10, 2015 at 23:18:09
The RAID controllers that I dealt with (I can't remember the model but this was exclusively on IBM servers) provided some sort of monitor program that allowed you to check the progress of a rebuild. In any case, I would expect messages in the Windows event logs indicating completion. On SCSI disks it would take in the order of four hours to rebuild a 150GB RAID 5 array.

I would imagine that a RAID 1 array would be faster than this as it is, essentially, a simple copy operation. I can't remember ever having to rebuild one of our RAID 1 arrays. There is no need to restrict access to the server whilst the rebuild is progressing, but it's sensible to start the rebuild at a time of light activity. I think if you leave it a day before swapping the second disk you should be safe. The disk activity lights will give some sort of idea that the rebuild has finished.

A slight confession - we always had a spare server (generous, or sensible, employers) which gave me the ability to practice all these sort of procedures before applying them to live servers. For example, I was able to test increasing the size of a RAID array by swapping the disks out one-by-one and then resizing the array without a reboot.

Take it calmly, and don't try to hurry things, and it should go like a dream. That's the whole point of RAID.


Report •

Related Solutions

#4
February 11, 2015 at 07:32:48
Yepper, as above, slap in the new hard drive, and turn on the unit, it should pick up the hard drive and ask if you would like to rebuild the array, you say yes, and it carries on with building the array and tells you the % complete.

To err is human but to really screw things up, you need a computer!

message edited by HopperRox


Report •

#5
February 11, 2015 at 13:31:06
Usually you have to designate the drive as a spare before the raid controller will use it to rebuild the array.

Fingers your link to storage arrays don't apply to rebuilding your raid array.

Understand that if you put in a 250gig drive to mirror a 150gig drive you will make that 250 a 150gig drive. There is also the issue that since the controller only reports the drive as a 150gig after the mirror even if you break the mirror you may not be able to expand the 150 to 250 since the raid array is reporting the drive as a 150. This is controller dependent. Some controllers will allow you via them to allocate, partition and use the rest of the drive [which means it recognized the available disk space]. it certainly is worth a try.

Answers are only as good as the information you provide.
How to properly post a question:
Sorry no tech support via PM's


Report •

#6
February 11, 2015 at 21:07:11
Sorry Wanderer, misunderstanding I said they were 250GB drives but they were SATA 1 (1.5GB/s speed, though I made the error of typing 150GB/s which is not correct and made the confusion). Again, Sorry. The drives are the same capacity as the original drives, just faster. Also I recognized that I needed to say with SATA II drives since SATA III drives will not be backwards compatible with the SATA I controller.

All: I looked again and there is no add on RAID controller so I looked up the motherboard numbers (Intel) and it does not have an onboard RAID controller so this must be a software RAID set up through the Windows Server 2003 OS. Does anyone have much experience with it? I am still hoping for a plug in and either automatic rebuild or maybe making a few selections and then it will begin. We will probably be ordering the drives tomorrow through Newegg and do the install Tuesday and Wednesday if they arrive by then.
Again, thank you.

You have to be a little bit crazy to keep you from going insane.

message edited by Fingers


Report •

#7
February 12, 2015 at 19:31:36
✔ Best Answer
Do you know if the raid config is for a mirror, does it tell you the raid config? In any event, I would simply. Unless you must keep the machine running, I would shut down, remove the old one, put in the new one turn it on, and it should recognize the drive, and ask if you would like to rebuild. Once rebuilt and booting, make sure, then add ther other new one, rebuild again. You would still have both old drives to put back in, if there was "uh oh". The jazz about size was; the original mirror was created with as 2 x 250g ish. If say you could not find a 250g replacement hard drive, but instead used a 500g, it would still create a 250g mirror on it, effectively leaving half, the drive underutilized. Doesn't seem to matter in your case here, but that's how it works. If after putting in the drive you get a wonky prompt you are unsure of post back.

To err is human but to really screw things up, you need a computer!

message edited by HopperRox


Report •

#8
February 13, 2015 at 07:33:50
Thank You. That is as I thought is should be, just trying to make sure. Right now (after replacing the exhaust fan and power supply) the disks are listed as healthy in Disk Manager so we are waiting. Tuesday morning they are going to restart the machine and if there is still the warning about the drive (that is when they always saw the warning), I will order the new drives, if the warning is not there, I will assume that the old power supply was below voltage on one or more of the rails and that was the issue and not the drives (or was causing one of the drive's problems). While looking around on the server I saw the drives noted somewhere as RAID and somewhere else as Mirror (this first may have been in Disk Manager, the other I do not remember) so it should be straight forward. The new working system is on a new server, this one is only now used for access to customer history which while still important, is only used occasionally. It was down for one day when the power supply died (I was off) and there was one customer that had to be called back the next day to reinstate an old credit to their account (no big deal). If it was the new server that went down, they would have had the system guy's on the phone in five minutes and probably on site within the hour, it was not important enough and much cheaper to wait for me the next day to deal with it. This way I get to learn something new (I understand RAID in theory but I have not installed or managed one yet) and I get to play with a server. I guess at the time RAID controllers and server boards with them were expensive and the software version was cheaper (being included), now all mid level boards include RAID (0,1,5 at least) so there is no need and it died out.
Thanks again.

You have to be a little bit crazy to keep you from going insane.


Report •

#9
February 17, 2015 at 21:27:57
Update: Since replacing the fan and power supply the server has been working well and Disk Manager reports both drives as healthy drives. We decided to give it a couple of days and then the person who told me about the original warning said he would restart the server and show me the warning this morning. All started correctly and the warning he told me about pops up. I read it and is does NOT say that a drive is bad, what it does say is that 'a driver or service is not working or could not start' (maybe not exact since it was from memory from this morning) which is not even close to the same thing he reported to me before. This warning has been going on for a long time and I gather dates back to about the time when the built in NIC went bad and an NIC card was added to the server. I am now assuming that the person who installed the new NIC card either did not properly disable the onboard NIC in the BIOS and/or did not uninstall/delete the driver for it. Since it has been otherwise working perfectly and this server is only for histories from our old system, I am not doing anything more and leaving it alone.
Thank you all (as always)
Fingers

You have to be a little bit crazy to keep you from going insane.


Report •

Ask Question