Volumes mirrored - what to do in case of failure?

[B00D00]

Windows 10, nine data partitions on WD 3 TB Red mirrored to 2nd WD 3 TB Red (awkward .. in Disk Management .. one volume at a time).

In case of a drive failure ..

Question 1: Do I get notified by any means, or do I have to look for "yellow triangles"?

Question 2: Do I just need to replace the failed drive? Will the volumes (partitions) automatically be resynched or do I have to undergo the one-volume-at-a-time-make-mirror procedure?

Thanks!

 

[B00D00]

Bump .. did no one run into this problem before? I'd really appreciate your help or suggestions in this matter.

 

[jmshub]

It would be much easier and more efficient for you to use RAID for mirroring. You already have WD Red drives, which are optimized for RAID. Nearly all modern motherboards support basic RAID functionality onboard.

At this point, you wouldn't be mirroring partitions, it would mirror the entire drive automatically. However, I can't say how to do that, it would depend a lot on your PC, and what BIOS it uses.

 

[B00D00]

@jmshub:
Well, that was not really what I was asking, but anyway: Mirroring to a 2nd drive is RAID-1, even if Microsoft for whatever reason tries to avoid this term. Mirroring on a partition base still leaves me the option to move a partition without breaking the entire array. And everything beyond RAID-1 is simply indigestible in terms of performance unless based on dedicated (expensive) hardware.

 

[PepperdotNet]

Ah, yes. I used software mirroring (yes, it's RAID-1 provided by the OS instead of a dedicated controller) for a while back in the Server 2003 days. I thought I was protected but always it ended in disaster.

A common scenario with these configurations is where only one of your two drives is actually bootable, and you likely won't find that out until the boot drive is dead. When you're mirroring partitions, not drives, and the master boot record is not actually located in a partition, these things happen. The last one I wondered about for weeks, finally I just had to test it. I pulled the port zero drive and booted the machine, and it just sat there saying missing operating system. Never again. Every system I've built since then has at least had an Intel ICHxR chipset or something better to support mirrored drives.

RAID is much less complicated, and safer, when done at the bare drive level with hardware that doesn't require the OS be running to function. And many modern boards have a basic RAID controller built in. Of course you can spend hundreds on a dedicated controller card with built in memory and battery backup, but for most people it's not necessary.

To answer the original question, Server 2003 didn't support any type of proactive notifications (and I haven't bothered checking whether it was added in any newer versions); and in theory, yes you would replace the drive and do some jiggery-pokery in the Disk Management console to rebuild everything.

 

[B00D00]

Thanks for answering, PepperdotNet - but that still does not answer my question. Server 2003 is light-years ago, and I'm mirroring only data partitions. Isn't there someone out there who already checked that out with Windows 10 or ran into such problems or someone from Microsoft QA who tested that before? I would be very thankful even just for a hint where to further investigate in it or for a link to documentation concerning this matter.

 

[PepperdotNet]

OK, I didn't get from the original post that your OS is not on the RAID. So that's a different matter entirely.

I said I hadn't bothered checking whether notification existed. But with the amount of work I do on Windows servers on a daily basis, if notifications for the native disk management were in there I expect that I would have noticed by now.

There is a technique using the task scheduler and a powershell command to send an email when a specific event happens. I have this setup for the Windows Backup service to report success and failure when the jobs run overnight. If you know the event(s) Disk Manager generates when there's a RAID failure, then that could accomplish your notification.

Details:

Trigger is On an event.
Select the Log, Source and Event ID(s) you are interested in.

Action is Start a program.
Program/script: powershell
Add arguments: -Command "Send-MailMessage -To [email@domain] -Subject 'Backup of XXXXX has completed.' -Body 'The scheduled backup on XXXXX has completed successfully.' -SmtpServer [valid smtp server] -From [email@domain]"