What happens when a drive fails in a parity raid?
Raid3 uses a dedicated parity drive. when one drive in the raid fails, it can be a data drive or a parity drive. If a data drive fails and is replaced, the failing data need to be reconstructed from the remaing drives, most notably from the parity drive. If the parity drive fails, all that needs to be done is to reconstruct the parity drive from the original data.
Raid5 used distributed parity, so when one drive fails, both the data and the parity needs to be reconstructed on that disk. It needs to reconstruct the data from the other parity drives and then reconstruct the parity from the other data drives. This carries more overhead than in a raid3 and that is why the rebuilding takes more time than in a raid3.
Reonstructing the drive has no impact on the OS and reinstalling the OS has no impact on the raid. Just make sure that the raid functions properly and you have the McRasid software available for installation.
Harm Millaard wrote:]
Raid3 uses a dedicated parity drive. when one drive in the raid fails, it can be a data drive or a parity drive. If a data drive fails and is replaced, the failing data need to be reconstructed from the remaing ...
That clarifies things! I have a Raid3, and if it turns out that my parity hard drive fails I'll use my external backup to copy the data back
I am ready to start the "hot spare" process. My new hard drive has arrived and I can pick it up.
The Areca web based browser is still doing its Volume check. I would like to abort it so that I can get started reconstructing my raid. So far it is at 60 percent and has found 144 errors. It's taken over 8 hours just to check 60 percent!
Can I skip the Volume check for now and start the hot spare process. I'll do a new volume check on another day -- so I can finish my video editing project.
Or should I wait the several more hours it is going to take to finish that Volume check. (FYI I've had the raid for about 4 months and never did a volume check.)
If you are going to rebuild the array then finish the volume check. If you have your data backed up and you are planning on just deleting the raid volume and configuring a new one then you can cancel it.
Well since I don't want to rebuid the entire raid then I guess I will have to wait for the volume check to finish.
But what happens at the end of the volume check? Are the "Errors" automaticcally fixed?
Some controllers have a scan only option and a scan and fix. Others have just a parity check option which also attempts to fix the errors. It depends on what your controller has. Most though scan and fix.
1 person found this helpful
Your volume check is just confirming what you already knew, your array is degraded. You can let it finish or you can stop it. There is no harm in letting it continue, but I doubt there is any benefit in letting it continue.
Whichever way you want to go, you need to add the new disk first as a hot-spare to your array. Once that process has completed, you need to expand your array, so your hot-spare becomes a regular disk in the array. That will take some time. Once that process has finished too, I would do a volume check to see if the rebuild was successful. It will be faster too, because the array is no longer degraded.
WEB BROWSER-BASED CONFIGURATION
6.6.5 Check Volume Set
To check a volume set from a RAID set:
(1). Click on the “Check Volume Set” link.
(2). Click on the volume set from the list that you wish to check.
Tick on “Confirm The Operation” and click on the “Submit” button.
Use this option to verify the correctness of the redundant data in
a volume set. For example, in a system with dedicated parity, volume
set check means computing the parity of the data disk drives
and comparing the results to the contents of the dedicated parity
disk drive. The checking percentage can also be viewed by clicking
on “RAID Set Hierarchy” in the main menu.
Directly from the Areca 1600 series Manual. I would let that finish.
I see that my "Volume rate" is at 81.5%. Even though it's taking forever, I figure I'm only 2-3 hours away from the checking to finish. So I'll just wait before adding the hot spare etc.
Hopefully whatever I checked when I started it going will fix the errors automatically. So far it's found 144 errors, which I don't think is too bad for 4 tb raid having its first "Volume State Check" (This is my first raid, and I suppose in a few months I'll be a raid expert...)
The volume that it is checking says:
Volume set name: ARC 1580-VOIL #000
Raid set Name: RaId set #000
Volume capacity 3000.GB
SCSI Ch/Id/Lun 0/0/0
Raid Level Raid 3
Stripe Size N./A
Block size: 512bytes
Member disks 4
Cache Mode Write Back
Tagged Queing Enabled
Volume State Checking (in red)
Errors Found: 144
The raid testing finished and all is marked as normal.
So I assume it fixed any errors.
I replaced the failing drive with a new one.
And am ready to create the Hot Spare. I am in the DOS Areca Technology RAID Controller setup.
From the drop down I am given the choice
SELECT HOT SPARE TYPE
Dedicated to RaidSet
Dedicated to Enclosure.
Which one do I choose? See attached image
BTW Do I need to have Windows 7 Format the new Samsung drive before I go through the Hot Spare process????
P.S. While waiting for a response here (I have itchy fingers, wanting to get this done) I went to another forum and one comment was:
You can probably do any of those if you want it to replace an already failed drive. Those options are usually for when it is still "extra" and just sitting there as a spare. When it's there as a spare, you can dedicate the spare to only be able to be used if something failes in "the enclosure", "the raidset", or anywhere ("globally").
My question would be if you have hot swap capability, why don't you just swap it with the failed drive? (Check the docs on that controller) rather than adding it as a hot spare first.
Message was edited by: rowby
Hot-Spare.jpg 38.3 K
Here's the lateset update:
Raid Set Information:
Raid Set #000 : 3/4 Disks: Incomplete
Create Hot Spare
Select Drives for HotSpare, Max 3 HotSpare Supported
[X] E2SLOT 10: 1000. 2GB Samsung HD 103SJ
Create Hot Spare Yes/NO
I select Yes
Select Hot Spare Type
> Dedicated to RaidSet
> Dedicated to Enclosure
I select GLOBAL
Then I Select EXPAND RAID SET
I Select "Raid Set #000 : 3/4 Disks: Incomplete"'
And I end up with No Device Available for Expansion"
I have made a little YouTube video which is uploading showing this problem. Stay tuned for link:
Not sure what I need to do to deal with No Device Available for Expansion"
I'm getting mixed messages from my system.
When the system is booting up I'm seeing
Waiting for Raid controller F/W to become ready
NO BIOS Found. Raid controller bios not installed
If I hit TAB and go into the DOS Raid Setup
In PHysical Drive Function, I see all 4 drives including #10 listed as Hot Spare (Identifying it as a Samsung -- so it is seeing the drive, I assume)
E2SLOT 9 ---- Raid Set Member Samsung
E2SLOT 10 -- Hot Spare
E2SLOT 11 -- Raid set member
EZSLOT 12 -- Raid set member.
First of all, why is it saying No bios found when the DOS raid controller is showing what looks like a raid...
SO I reboot --and this time I will go into the WIndows based Raid Controller
WIndows 7 startup
Launch Browser based Control
My main bewilderment was why I was seeing "NO BIOS Found. Raid controller bios not installed" at bootup.
So I finally took out my new hard drive and replaced it with my old one that is apparently going to fail (although I noticed during the 20 hours of Volume check there was not one single time out in the log. So I am wondering if the hard drive is really failing after all.)
I rebooted and i no longer got the No BIOS found message.
And my raid is showing again, full drive etc.,
The Web based Raidset Hierachy is showing all of my drives as raid sets and the Volume state is Normal in RED.
So I am thinking that the bios must have been on the hard drive that I replaced with the new one.
With all of the above in mind, should I then go back to the failing drive in my raid set and make that as a hot spare -- BEFORE I remove it from the computer????
After I do that then I can swap it out with the new drive without turning off my computer.
And THEN expand the raid set?
1 person found this helpful
If you made a hotspare, the hotspare will automatically rebuild in the case of a defect disk.
If you don't have a hotspare and a disk fails, you can hotswap a new disk. Rebuild will start immediately on the new disk.
If you turn off the PC and remove the defect disk (cold swap) and insert a new disk and restart the system, the Areca will do nothing..!!
Areca has marked the location of the defect disk as bad and for the rest nothing changed. In this case you need to tell the Areca via McRAID BIOS or webbased to create the new disk as HOTSPARE, and immediately after that the hotspare will be used as a member of the INCOMPLETE ARRAY to rebuild.
The missing BIOS message is very strange and I do not have an explanation.
It's good to know that with a Raid3 I have so many options.
For now, since I am not getting any timeouts in the log I am going to keep my current setup and revisit the hot spare and the mystery of the missing bios message next week.
I should finish my project on Monday and then I can dig more into this.
I am hoping that the Volume check that found about 155 errors (mainly near the beginning of the check) fixed any bad "sectors" on the "bad" hard drive and I can continue with my work for the next couple of days.
Yes, I am making backups of my project at least twice a day just in case.
I think the mystery of the " the missing BIOS message" must have something to do why I was unable to make the new drive a hot spare.
Once my project is finished I may just do a complete rebuild of my raid configuration if that's what it takes to get the bios to return.
For now, with my old hard drive back in its slot, I am not getting that missing bios message.