How to replace a failing disk in a Raid 1 array?

A couple of years ago I set up a Raid 1 array using two identical ‘Seagate Barracuda - 1 TB internal hard drive ST1000DM010’ hard drives in an ORICO 9528RU3 enclosure.

It sounds like one of the drives is about to fail and I intend to preempt it and replace it with another one.

My questions are:

  1. How close does the new drive have to be to the current one?

  2. How can I tell which of the two drives is actually failing?

  3. How one earth do I actually do it, I thought it would be a matter of slipping the failing drive out and sliding the new one in then rebooting, but apparently not. The more I read online the more confused I get.

I have basic computer knowledge (about enough to, well, set up a RAID 1 array).

Thanks in advance! :slight_smile:

How did you set the array up in the first place? Whatever interface that was, there should be an option to mark a disk as failed (if it hasn’t died already) so the array stops using it, then, after replacement, to enable the use of the new disk, after which the array should re-build itself automatically. You should not have to re-boot, assuming you can hot-swap disks into and out of your system while it is running.

As for size, the replacement disk should be at least as big as the other disk in the mirror (might as well get the same sized disk).

I can’t remember exactly how it was set up, I hope I kept the original documentation! (I usually do)

Thanks DPRK :slight_smile:

Can you be more explicit about the symptoms? Is it literally making noise? Is it failing SMART tests? Does it have bad sectors?

I forgot to add: to determine which disk is failing (if you can’t tell based on the loud grinding noises :slight_smile: try reading the S.M.A.R.T. status for each to see which one reports a bunch of errors.

Also: if you have 1 TB mirrored, you need at least a 1 TB replacement disk like I said, but if you want to expand the array eventually with bigger disks you could get one with bigger capacity.

I recently did just that:

Replaced a failed 1TB Seagate barracuda disk in a raid 1.

Mine is controlled with an intel storage controller (some software)
I took out the failed disk, replaced it with one of the same size/size cache and waited an hour: Job done.

As to determining which disk was failing: I have a external enclosure, I put in each disk, connected that to a computer and tried to read the files. One worked, the other not.

External disk enclosures are to be had for ~15$

You normally can just take the failed disk out. No matter what you do, it’s normally no worse than “suddenly and absolutely failed”

You sometimes can put in a replacement disk without turning the power off, but I wouldn’t do that.

You almost always have to reconfigure the array to get it to use the new disk. RAID systems don’t just assume that because a disk is there, it should be added to the RAID.

It’s more fun to add a slightly larger disk, so you have rome to expand later, BUT, there are OLD controllers out their that can’t handle disks bigger than 2GB, and OLD systems out their that can’t handle arrays larger than 2GB. If it’s more than 10 years old, check before getting a big disk.

Hot swap drives normally have amber lights to indicate impending failure. I’d do that every week, walk the entire datacenter logging amber lights and swapping drives.

As DPRK describes below, its just making loud and unhealthy sounding grinding noises! And yes its giving a failure on a SMART test. Not sure about the bad sectors, I haven’t been home long enough to give it a thorough test.

Thanks for the answers everyone, very much appreciated :slight_smile:

A lot of NAS units DO actually sense that a new HDD has been added and then automatically format and copy to it. My D-Link units (older, I’ll admit) do this very neatly. Just power it down, replace the bad HDD, and then power it on. Works with both HDD-0 and HDD-1.

Part of why I love drobo.

Light turns red, replace drive, walk away. No powering down, no configuration.

I would still double-check that if you replace, say, a 1 TB drive by a 2 TB drive, the automatic system recognizes the capacity change and will expand your storage once all drives in the array have been upgraded.