So, I use a Drobo at work and I have an unRAID unit at home. I've been very pleased with the unRAID and I haven't tripped across this issue yet at home, but I have tripped across it at the office.

 

One of the nice features the Drobo can do is if I have enough "free" space it will automatically move the files so that it'll remain protected. ie, Let's say I have 4 1.5TB drives, it has a storage capacity for 4.5TB. If I "only" have 1TB of data, if 1 of the 1.5TB drives fails, the Drobo will automatically move the files the remaining drives and since I have enough space, it'll remain protected after a rebuild.

 

I believe the unRAID can do this but it is the manual procedure on how to remove a drive found here:

 

http://lime-technology.com/wiki/index.php?title=UnRAID_Manual#Remove_one_or_more_data_disks

 

Is there some way to configure this under unRaid to do this automatically? Perhaps after a configurable timer? ie, Configure it to wait 30 minutes before rebuilding? The Drobo does this within minutes which is sometimes annoying if I'm just popping out the drive to check the serial number. I suppose this really isn't an issue on the unRAID as it doesn't officially support hotswapping at this time.

 

 

This has been discussed before. I think the consensus was that we didn't want this feature. For example, if a single drive fails due to a problem which could affect other drives too - like fan failure in the case resulting in drives overheating - I'd rather be made aware of the issue, and not have a "hot spare" drive. If other drives fail before we can shut down the server, fix the "overheating" problem, parity protection will not be protection at all.

It might be a nice optional feature though, for those who would want it.

Hmmm....I hadn't thought about that situation, but I'm not sure it would apply or if it does, it would apply to my benefit or maybe I still don't quite understand the problem.

 

For example, let's use an easy example. I have a 1 parity drive and 10 data drives. I have only 10% of the drives filled so I have lots of room.

 

If 2 or more drives fail outright at the same instant in time, unRAID can't rebuild automatically anyway right?

 

If 1 drive fails and it starts to rebuild and a SECOND drive fails during the rebuild it would be the same problem whether you manually or automatically started a rebuild.

 

I don't have a problem with it being optional as well. I was just thinking I sort of liked that feature and it didn't sound like it couldn't be done via a script or something.

I see what you're saying, but when you manually rebuild a drive, you have the option of checking what, if anything, could have gone wrong with your server. The danger lies in those situations when one doesn't even realise a disc has failed - and unRAID continues to keep the array working even harder to generate data on-the-fly.

Now that I think about it, in previous discussions, some even thought it would be better for the server to stop once a drive fails - now that's a bit too hardcore for me. Still, I like your drobo-like feature in that it offers some built-in redundancy provided one has the spare space.