I have two identical 1TB disks in a RAID 1 format, so one is supposed to mirror the other.
Last August, there was a glitch on one of them and whenever I booted the computer after that, my Configuration utility (pre-BIOS screen) told me that the RAID was "degraded" and that one of the disks had an "error". I stripped the PC down last week to replace the PSU and thought I might as well address the RAID issue at the same time. Booting the PC from one of the disks on its own, everything appeared fine. When I tried it on the other disk I got an error saying that disk was not bootable and a closer look showed that it had not mirrored anything since the date in mid-August 2010 when the problem occurred. So I guess I was lucky nothing happened to the working disk.
So the first conclusion is to deal with the problem immediately it happens, not 9 months later.
However, what are you supposed to do then? My Configuration utility told me that to take a disk out of RAID loses all the data on the disk, so I did that reasonably happily on the corrupted disk as I felt it needed a complete re-format anyway. Then what to do about the working disk? In the end, I backed it up to a spare solo 1TB drive, took the main disk out of RAID using my configuration utility, then re-formatted the main disk, loaded the back-up back onto it and re-set up RAID on the two disks. That seems to me a long way round. What is the recommended routine for dealing with a RAID error like this?
If one of the disks had not had an error, would the PC boot from either of them separately or only one of the two?
|>> Re: What to do when RAID error occurs
I havent had the opportunity to run a raid-1 in my own system, only raid-0. As you already know that with raid-0... when a drive kaaputs.... the whole array is toast... replace fail hdd, reconfigure the raid and reload...
With raid-1, if you loose a partner hdd... you should still be able to boot from it if this raid is your primary boot array and in the case if the array was just another drive, still have access to the data...
Typically the raid subsystem on the motherboard are software driven.... in your case/situation, I believe you should have had a raid utility in order to rebuild the partner hdd once it had been replaced. The way I believe it should work is that once the raid is built, both drive contains info about each other... if one of the hdd goes bad... you should be able to rebuild it by booting up to desktop on the good drive, after you have replace the suspected bad drive...... start the raid utility and have it rebuild the array group.
The raid controller on the motherboard usually only contains configuration to build or delete the array group... and it takes a software utility to do the rebuild.
Now there are separate raid controllers that are considered "hardware raid" types where the function of rebuilding the raid is built within the raid controller itself, but these types off controllers are more costly. In your situation with a raid-1, one of the hdd goes bad.... you replace it with a new hdd, go into the controller's bios (pre-post), and then here you will execute a rebuild. And many times, with the more expensive controllers..... once you replace the bad hdd with a new one, as soon as the card is powered up and all the hdd are spinning, the controller can go into an auto rebuild without any intervention.
In your case, you shouldnt have to go through the pain of a reload/image due to a failed hdd in a raid-1 group. You should've been able to boot from the remaining drive.
I havent mess'd around with pc onboard motherboard raid in a good while so I am not sure if they have gone about adding a rebuild function into the little raid controller these days or not so as to not mess with soft raid rebuilds...