Help with Raid 5

DocGreen said:
When I reset the disk I'm given the option to mark the array as 'normal' at which point it attempts to rebuild, then fails. But... if I pull the bad disk and replace with a good disk, it doesn't give me the option to reset/mark as normal. It shows the bad drive as 'missing' and the new as a non-member.
Click to expand...

Are you doing this from inside the Intel BIOS or using the Intel software from within within the OS. Every time I have swapped out a drive in that environment I have always used the BIOS app to do things.
 
markverhyden said:
Are you doing this from inside the Intel BIOS or using the Intel software from within within the OS. Every time I have swapped out a drive in that environment I have always used the BIOS app to do things.
Click to expand...

I expected to need the BIOS utility, but no... everything is done in the windows GUI.
 
jbartlett323 said:
Your problem is the new drive, and your screeny's tell me so. You are trying to install a drive with a physical 512 sector size, and all the currents are 4k sectors. Basically, this drive is incompatible with your array, try another...
Click to expand...

Can anyone else confirm this? Exchanging the replacement drive would be the easiest fix by far.
 
A quick Google search shows that both the old and new drives are 512 sector size.

You should try making these changes in the BIOS instead of the GUI to see if you have any luck.
 
DocGreen said:
Can anyone else confirm this? Exchanging the replacement drive would be the easiest fix by far.
Click to expand...

If you have a 1, 2, or 3 TB laying around that is 4k/Advanced Format, you can easily confirm it for yourself. A drive in an array can bigger than all other drives, just not smaller and you will not have access to the remaining space. You can then replace it again later, but will need to rebuild again.

I'll admit, I could be wrong. Don't have an Intel array handy to try it on but my own server. But then I haven't bought a 512 sector drive in quite a few years either.... Oh and HGST has this bad habit of some drives being 512 and some 4k on the same model. Depends if it was made before or after the WD buyout, and/or some other factor I am unaware of. Can't seem to do that quick google search and confirm what this guy has, you have a definitive Taptech? My Google-fu has failed me today, cause I can't even find it on their data sheet: https://www.hgst.com/sites/default/files/resources/DS7K1000.C_DS_final.pdf

What I can say is that I'm thinking that Samsung you are trying to insert is really, really old. Great model, run em myself for over 5 years in my server. But not sure that's what I would want to put in a box today...
 
TAPtech said:
A quick Google search shows that both the old and new drives are 512 sector size.

You should try making these changes in the BIOS instead of the GUI to see if you have any luck.
Click to expand...

Unfortunately the only options available in the bios are to delete a volume or create a new one.
 
jbartlett323 said:
Can't seem to do that quick google search and confirm what this guy has, you have a definitive Taptech?.
Click to expand...

Just some tech specs from old websites/reviews. If the sector size is changing on the same model, though, then forget what I said. It does look like the Intel utility is reporting both drives as 512. I can't think of much else, except trying another drive, maybe even a larger one. Might be hard to find a drive that isn't AF and larger than 1TB laying around though. Well.. maybe that new drive, being slightly different, is slightly smaller than the others. As in the logical volume is set to the same max size as the physical drive. New drive has 1MB less space, so it won't work. Maybe.

edited to fix quote format thingy.
 
Reply to my reply... maybe you should check in Disk Management to see how the actual drive sizes compare. Check the RAID array against what's available on that replacement drive.
 
I have been watching this thread since the first post. Docgreen, I feel for you indeed. I hope you get this cleared up soon.

Whats the progress on this now? Did you replace the drive?
 
coffee said:
I have been watching this thread since the first post. Docgreen, I feel for you indeed. I hope you get this cleared up soon.

Whats the progress on this now? Did you replace the drive?
Click to expand...

Not yet, I've been trying to take care of a few customers and clear some invoices first before I dive back into that whole debacle. Honestly I'm going to need to purchase a few drives... mostly because the raid array was the biggest I had. I don't have anywhere else to store that much recovered data, LOL.

markverhyden said:
If you look at the pics he posted the failed drive is 512/512 just like the new drive.
Click to expand...

Thanks for pointing out the obvious answer we all seemed to miss, lol. Would it still be worthwhile to try a different replacement drive?

I'm also really curious as to why the array never showed as degraded. I was logged into the server when the drive crashed and never got any kind of pop-up warning from the IRST utility, and when I restarted, the BIOS utility showed the array as failed. I thought that a raid 5 wouldn't show as failed unless more than 1 drive dies... shouldn't a single drive failure trigger a degraded status?
 
lcoughey said:
My new slogan from earlier this month.

If it is too big to backup, it's too big.
Click to expand...

Most of it is customer HDD images, recovered data, etc. I try to keep that stuff on hand for 90 days in case the customer finds something I missed. I have a local backup that was supposed to be backing up all of MY stuff... business stuff, tools, sensitive personal stuff, etc. But for whatever reason the backup got screwed up. When I went to check on it, it listed \\?\Volume{blahblahblah... instead of D:\blahblahblah... Only a little annoyed that I never got any popup notification about the backup failing.

I've been meaning to start backing up with an off-site service that I partnered with... I just never took the time to set it up. Whoops.
 
DocGreen said:
Most of it is customer HDD images, recovered data, etc. I try to keep that stuff on hand for 90 days in case the customer finds something I missed. I have a local backup that was supposed to be backing up all of MY stuff... business stuff, tools, sensitive personal stuff, etc. But for whatever reason the backup got screwed up. When I went to check on it, it listed \\?\Volume{blahblahblah... instead of D:\blahblahblah... Only a little annoyed that I never got any popup notification about the backup failing.

I've been meaning to start backing up with an off-site service that I partnered with... I just never took the time to set it up. Whoops.
Click to expand...
Just to be clear, that wasn't meant to be a direct remark to your lack of backup...more an observation of the growing problem of very large storage arrays that are nearly impossible to backup and, sadly, too much reliance on RAID.
 
DocGreen said:
Thanks for pointing out the obvious answer we all seemed to miss, lol. Would it still be worthwhile to try a different replacement drive?

I'm also really curious as to why the array never showed as degraded. I was logged into the server when the drive crashed and never got any kind of pop-up warning from the IRST utility, and when I restarted, the BIOS utility showed the array as failed. I thought that a raid 5 wouldn't show as failed unless more than 1 drive dies... shouldn't a single drive failure trigger a degraded status?
Click to expand...

To be honest trying another drive will most likely not make a difference. When a RAID is marked as failed a bit or bits are set which creates this flag. Much like M$ has a dirty bit to force a check disk. Your RAID is basically a software RAID as that is what those Intel chipsets are. However I've never worked with one of those at a low level. A couple of customers had them on workstations as mirrors. When one drive failed I would just swap the drive and it would start rebuilding when powered up.

I'd venture a guess that the Intel may follow the mdadm protocol. I've seen that with drives pulled from LaCie RAID devices. So you may be able to image each drive and mount them in a linux. There are techniques to repair RAID in Linux, which includes clearing the failed state. Of course it is risky. As has been mentioned you need to make sure you have copies of everything. Preferably two which have been verified.

https://raid.wiki.kernel.org/index.php/Recovering_a_failed_software_RAID
 
Recovering the data is almost certainly possible - if you think about it what gets the drive marked as failed is errors over a very small percentage of the data on the drive(s). Just restoring the RAID at this point sounds like maybe a lost cause and better replaced with newer hardware.
 
According to what I have read on the Intel site it is compatible with mdadm and linux. You should be able to load up linux (Mint is my choice) and recover your raid. There is also an option of clearing the failed bit if I remember correctly. I would be happy to help out if you need it or would like to call me if you have a question.

I run 2 software raids here at my shop. I have my work computer that I have a RAID1 and then my server runs RAID5 for backups. They both run on linux and my RAID5 has been running about 7 years :)

coffee
 
You must log in or register to reply here.

Similar threads

britechguy
Today's BSOD, Code BAD_SYSTEM_CONFIG_INFO - What's actually going on here?
Replies
13
Views
1K
fincoder
fincoder
Appletax
[SOLVED] Intel Rapid Restore Driver Causes BSOD - DRIVER_POWER_STATE_FAILURE
Replies
3
Views
537
Appletax
Appletax
E
[SOLVED] New 1TB install issues with blank ssd
2
Replies
30
Views
3K
ell
E
HCHTech
How to screw up with class - :-)
Replies
5
Views
1K
HCHTech
HCHTech
Big Jim
Dell T320 Server and RAID question
Replies
18
Views
3K
nlinecomputers
nlinecomputers
Back
Top