Solved

how much space required to rebuild Raid 1

Posted on 2012-03-26
15
367 Views
Last Modified: 2016-12-08
Hello,
I have a server out there with degraded raid on an Intel Matrix raid controller.
One of the drives have failed and I am going out with another to replace and allow the Mirror to rebuild.
The drives are 320 GB drives, but only 13GB is free.
Will it be able to rebuild itself with this little amount of free space?
I've included a screenshot that shows the intel matrix storage console
The controller is Intel 82801GR/GH Sata Raid controller
Not sure of the other specs. It appears to be a clone machine running software.
0
Comment
Question by:bwierzbicki
  • 5
  • 4
  • 3
  • +3
15 Comments
 
LVL 9

Assisted Solution

by:OCDan
OCDan earned 72 total points
ID: 37767678
Yes it should be perfectly fine, when it rebuilds an RAID1 array it merely copies the data over to the new drive.

Although your system will go pretty slow until it copies over, space will not be an issue.

Hope that helps.
0
 
LVL 16

Assisted Solution

by:R. Andrew Koffron
R. Andrew Koffron earned 72 total points
ID: 37767691
in theory the RAID should be fine rebuilding regardless of the free space, even in a software raid I do believe, just make sure the new drive is the same or larger than the existing one. and that you have a full and complete tested backup before you try it.
0
 
LVL 47

Assisted Solution

by:dlethe
dlethe earned 285 total points
ID: 37767697
Rebuild occurs in-place. As long as the replacement drive is >=320GB, and is connected to the controller, and certified, you are good.   Obviously best practice is same  make/model, but there is no requirement that has anything to do with unused space on the drive.  A rebuild is similar (but not exactly the same) as a bit-level copy.
0
Portable, direct connect server access

The ATEN CV211 connects a laptop directly to any server allowing you instant access to perform data maintenance and local operations, for quick troubleshooting, updating, service and repair.

 
LVL 47

Assisted Solution

by:dlethe
dlethe earned 285 total points
ID: 37767713
I will mention, however, that since you are using RAID1, and if your O/S is one that has software RAID1, then you are MUCH better off letting software RAID1 do the job.

This is a fakeraid controller, and all the work is done by the O/S anyway, but if you use Win7 RAID1, as example, you will get much better performance on reads, as it load balances READS.  

The fakeraid controllers do no such load balancing.   (But converting from one to the other requires a full backup, blowing away the RAID, then restoring)
0
 
LVL 8

Assisted Solution

by:teomcam
teomcam earned 71 total points
ID: 37767732
Since you are using mirroring, the amount of left over space is not the issue. Proceed confidently.
0
 

Author Comment

by:bwierzbicki
ID: 37767738
Thanks everyone,
I will purchase a new drive and get out there asap.
After I'm done I'll post back.
0
 
LVL 28

Expert Comment

by:burrcm
ID: 37775502
1TB external drive < $100. Back it up first.

Chris B
0
 

Author Comment

by:bwierzbicki
ID: 37776969
Yes Chris, I bought a 500GB and am going to image the drive before putting the other 320 in place for rebuild. Good advice.
0
 
LVL 16

Expert Comment

by:R. Andrew Koffron
ID: 37777128
I was just about to "good advice" Chris also.  always backup I think carpenters say "measure twice cut once"  I always tell the guys that work for me it's not backed up, till there 2 copies and they've booth been tested for recoverability. :)
0
 
LVL 47

Assisted Solution

by:dlethe
dlethe earned 285 total points
ID: 37777265
Actually, it is marginally better to NOT back up first.

Reason - You are not protected until you have a mirrored array.   It will take less time on the clock to reconstruct the RAID1 then it will to backup then reconstruct.   Not only that, but USB is much slower then native SATA so it will take 4-5X longer just to backup via image.

Also if you get an unrecoverable READ error, then the image backup will likely fail and you will have gained nothing for your time.  On a reconstruction the firmware can deal with it, so it won't abort.

If the life of the surviving drive is limited, your goal should be to get the system in a redundant mode ASAP.  Let's say you did backup and have a clone and the primary disk died ...

How do you turn that clone into a degraded RAID1???   (You don't, at least not without running some specialized software and dealing with metadata and moving logical->physical blocks around ... it just isn't practical)

So no matter what, you are increasing odds of data loss by taking a backup.
0
 
LVL 16

Expert Comment

by:R. Andrew Koffron
ID: 37777342
haha everything dlethe just said makes perfect sense to me!, I'm sure he is correct


However since the primary potential loss is time, I'd still back it up I think.

@dlethe is there actually any significant potential data loss other than the remain drive could also catastrophically fail durring the backup(witch would probably also happen on the reconstruct)?  or is the risk really time on site?
0
 
LVL 47

Expert Comment

by:dlethe
ID: 37777472
Well, at this point, unless you run diagnostics on the surviving drive as well as properly burn in the drive you intend to use for the replacement disk AND the backup disk,  then no way to properly assess risk.

The one absolute is that the clock is running, and you are  unprotected.  The way to be properly protected against both a catastrophic drive failure and partial data loss due to a unrecoverable read error is to be in a redundant mode with disks in sync in a RAID1.

A backup does not provide this.  You end up with two copies, and there is no checks and balance to deal with data integrity issues, like if block X on original disk and block X + metadata size on clone disk don't match due to ECC problem, along with other scenarios..

Look, it is getting needlessly hypothetical.  I only jumped in on the tail end to set record straight that redundant, in sync RAID1 provides full protection and a backup does not, so logically you want to get the RAID1 built ASAP.  (Plus again a backup is not a true copy and can't be easily converted into a RAID1 w/o going through read and reformatting)

The clock is the enemy.  You can have it synced up and protected in 1/5th the time if you just sync the mirror then if you back up first.
0
 
LVL 16

Expert Comment

by:R. Andrew Koffron
ID: 37777609
@dlethe I was totally asking so I would know, not being a smart ass or anything. You've given fantastic advice every thread I've ever seen in. I appoligize if I came off like a doubting you. was actually curious if there was extra danger of data loss.  I agree on time being the enemy, I just usually try to take the safest route with client data.  I do see your point and will pay attention to it in the future.
0
 
LVL 47

Accepted Solution

by:
dlethe earned 285 total points
ID: 37777664
no, no, I am not on ego trip or anything like that. no apology necessary, i was just pointing out that the thread was going off on an incorrect tangent, and when you are degraded, your #1 priority is to get "undegraded".     an Image backup via  a SLLOOWW USB-attached drive in such a way that you can't turn that image into a degraded array isn't the way to go.

(If this was a RAID5 then I would have said to backup first, because a RAID5 reconstruction requires WRITES, which is much higher risk of HDD failure then reads)

That's all.   -- P.S. I am not politically correct, so I can ruffle feathers at times, and if you read my posts, you know at times I can get on a soapbox, and don't always take the side of the author of the question ... don't' ever take anything personally,  ;)
0
 

Author Closing Comment

by:bwierzbicki
ID: 37781838
Thanks everyone for contributing.
I didn't get a chance to get to the site Tuesday nite because I got stuck at another site until 9pm. It was nerve racking and very stressful to know that condition existed and I couldn't get to it. On the other hand I didn't want to go onsite with less than full faculties which was the condition I was in after a long Tuesday. I factored that against the possibility of this other drive failing and decided it was best to wait a day.I had a little comfort knowing that there were two different types of drives in this server and that probably the one that failed previously and now the second original was going. I know i was making a lot of assumptions and taking a risk but I didn't want to walk in brain dead and make my own bone head maneuver that screwed things up.

I'm glad I did because I never expected to find what I did.

When I opened the server the failed RAID member had not failed, it was unplugged. Both the power cord and the SATA cable. I thought it was odd and wondered if someone had been in the box, or has the Array been in this condition all along and I just didn't notice (not the case) but with the server on a bench I started examining the cables and to my surprise they were very loose fitting, especially the power cable. I can only assume that years of running and vibration (however slight) had caused them to come unplugged.

So, i was faced with a quandry, just plug it back in and let the rebuild occu?. I chose not to do this because, since the drive that i thought had failed was on port 0, and because it wasn't a blank disk I feared the rebuild might have been from that drive, which wasnt current, and that it would overwrite the current data. I had a similiar situation on my personal pc at home.

So what i did was hook up a new 500gb drive and the known good Array member to sata ports on another machine, boot to an Acronis True Image cd and did a clone. One that I plan to keep in the event of catastrophe, I would be able to get the server up quickly followed by a data restore from backup. It took an hour but, no problem.

I then put the good array member back in place and installed the new hard drive, booted to the Raid bios utility with a "ctrl-I" and told it to rebuild the array. After that the OS started and the rebuild began.

You were right on about this being fake raid dlethan. I never knew and assumed that hardware would do it. After seeing this process it makes sense now.

Thanks again to everyone. Although this whole process was very straightforward and went as expected the stress factor added by what's on the line makes it a very nerve racking experience. Glad you guys at EE are there to have my back by sharing your experience and making it possible to perform, under pressure, with confidence.

fyi...i connect the drive I removed via usb and it had everything on it. I could see the data and assume it was current, up to the day the cables slipped off. I used a tywrap to hold the power cable in place since there were no other available connections.
0

Featured Post

Optimizing Cloud Backup for Low Bandwidth

With cloud storage prices going down a growing number of SMBs start to use it for backup storage. Unfortunately, business data volume rarely fits the average Internet speed. This article provides an overview of main Internet speed challenges and reveals backup best practices.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

Like many organizations, your foray into cloud computing may have started with an ancillary or security service, like email spam and virus protection. For some, the first or second step into the cloud was moving email off-premise. For others, a clou…
Two types of users will appreciate AOMEI Backupper Pro: 1 - Those with PCIe drives (and haven't found cloning software that works on them). 2 - Those who want a fast clone of their boot drive (no re-boots needed) and it can clone your drive wh…
This video shows how to quickly and easily add an email signature for all users on Exchange 2016. The resulting signature is applied on a server level by Exchange Online. The email signature template has been downloaded from: www.mail-signatures…

860 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question