Determining which drive failed in a RAID 1 configuration

[trents]

I built a box for a customer several years when I first started up my business. He was not open to online backup solutions like Carbonite and I wanted to come up with a solution that would backup the whole system, not just the data. If I had to do it over again I would do it differently but at the time the best solution seemed to be RAID 1 using t he motherboard bios RAID. He called me today and from what he described it sounds like one member of the RAID volume has failed as he gets an error message during boot but then it finishes and boots into the Windows desktop.

My question is, how do I determine which drive needs replacing so I can rebuild the RAID? I think they are identical WD drives if I recall. Can I just disconnect one and then the other and see which one boots into Windows? Will that break the RAID volume? Is there an easier way?

 

[HankB]

The Linux tool for examining the SMART data will also report the serial number for the drive. If the corresponding Windows tool ... You can use Smartmontools on Windows too: https://www.smartmontools.org/wiki/Download Then just determine which drive failed, note its serial number, open the case and replace it. At worst you remove one drive unnecessarily and if you're lucky, you pull the bad drive on the first try.

I know nothing about Windows RAID so I can't say whether disconnecting/reconnecting drives is a bad idea. I think it would be OK as that's how you would replace a failed drive. If you pull the bad drive and power up, any changes to the good drive will be migrated to the replacement by design (when it is installed.) If you pull the good drive by mistake, you don't really care what gets written to the bad drive.

I would do the diagnostics first just confirm that it is a bad drive and not something else.

Good luck!

 

[trents]

Thanks for the input, Hank. Some good thoughts. I'm thinking since the system still boots (or I believe it does since I don't have my hands on it yet), I'm thinking I'll just pull the SATA cable on one and then the other to see which one is still functioning as the boot drive.

 

[trents]

Hank, as you thought, the RAID interface at bootup ID'd the failed drive by serial number. So that was a no brainer. As it turns out, in Windows 10 at least you can convert a RAID drive to AHCI with Powershell commands without losing data. It's actually pretty simple. So that's what I did with the healthy drive. Now I will be able to use Macrium Reflect free to create images for the customer on a schedule to the backup drive. I will also invoke File History to automate just data backup. That's what I should have done 4.5 years ago when I built this system for the customer.

 

[habbajabba]

Are you doing backups now or both backups and the raid? Kind funny to setup raid for someone then not know how to fix it lol. I tried macrium but veeam has a cool little tray icon which is ever so handy as veeam does the scheduling for me without having to work with windows scheduling tasks, or macriums funky gui.

 

[trents]

Back when I built the box for him I had been playing with RAID and was more up on it. But haven't messed with it in several years and was kind of rusty. It all worked out well, though. I'll have to look into VEEAM. Is there a free vesion?

 

[habbajabba]

Veeam is great IF it is used as a replacement for windows' built-in system restore feature. I have discovered however that after reinstalling windows clean (erasing the original hd), veeam's bootable flash image complained about a missing dll and therefore was essentially worthless in restoring my original working environment. And this was with it's latest version. Actually it was simply a reinstall of the original windows os disk. It can in session, restore files and folders, and if you bork your running config somehow and want to fix it, it works excellent. I would recommend it for running windows installs ONLY, even if, or more precisely, they can't boot. But sadly when I simply wanted to abort the new windows 7 install and go back to my fully configured one, it couldn't simply rewrite the image of the previous install and let me move on with my life. From now on clonezilla will be the only way I backup a working OS disk. It is boot image only so it absolutely works with bare drives and doesn't care what the OS is or was beforehand. Veeam is totally free but for above mentioned reasons I will no longer be using it. Also I was able to update to the latest version of Veeam on my desktop, but no matter what I tried on my win7 pro tablet it would not let me update and I am stuck using an older version. Clonezilla will never have these problems whatsoever but you must reboot to backup or restore which is why I was running veeam in the first place as it lets you simply create a task that runs automatic and you never have to shutdown.

With my nas I learned that it will only let you create a raid if you initialize at least 2 (empty) drives as a pair in at once. Once you have a raid1 configured then you can add to it easily. I have to buy another 6tb wd red in order to create a raid1 with my first 2 just so I don't lose all the data I bought the nas for to begin with. Right now they are basically just copies of one another.

The other thing that's weird about the nas is that once you install something it somehow is not uninstallable only disableable. Worse yet, after disabling chrome I found it runs at boot and is listed in the shutdown processes as if I had been using it. In fact, anything you install, even if you haven't opened it, is listed in the process shutting down list when you shutdown. I do not like that at all. Probably all a part of the nas's vm OS environment, but it takes a lot of time for it to shutdown 37 processes. Once I get the two 960evo's I will pull all the drives and start over with just the 960's as a raid1, then add the pci-e 970evo's as cache, then reintroduce one 6tb red, leaving the other in my desktop because it will take three of them just to create a raid1 on the nas without losing all my stuff. In other words if you are not actually going to use it don't install it. Chrome will never be installed again.
Plus I need more memory if I want to run any containers or vm's. So the nas is cute but not very functional yet. Once I get everything raided and have more memory, then I can really start using it with qtier and over provisioning of the ssd's. The nas is very powerful but requires more than just a single hd to be functional. I should really add that the nas will give you alerts if for any reason a drive is failing and makes it easy to replace it. Now if I could just get those 960's.