Age at which to start replacing drives?

[darkryoushii]

Hey there,

So I have been running my FreeNAS build for just under 2 years now. Throughout this time only one of the 5 HDDs (Listed in my signature) I have in my Raid-Z1 configuration failed, and after a warranty replacement the array was re-silvered correctly.

I am starting to get a bit worried because of

a) the inherent problems with Raid-Z1 during a drive failure
b) the age of the drives

and am wondering at what age do you start to think about replacing your array with a bigger one or at least replacing the drives to avoid a full array failure? I have backed up all photos to crashplan but would still like to avoid losing all my media and other items I thought I didn't care about. The array is currently 40% full.

 

[9C1 Newbee]

My opinion, run them till they die. Keep a back up of the non-replaceable stuff.

But, if you could move the data off of the drives and add a drive to create a z2 pool, there is no better time than now. The pool will only get bigger and the problem harder to rectify.

Also, if you have deep pockets to buy 6 new drives, then simply create another pool and transfer the data off. Then you have the option to split the other drives into separate pools or whatever.

To trash drives based on age seems foolish to me. But it is your money and your data after all.

 

[darkryoushii]

To trash drives based on age seems foolish to me. But it is your money and your data after all.

I wouldn't trash them, but definitely would prefer to start re-purposing drives before they die on less important boxes such as firewall or HTPCs.

Who knows, maybe I have immortal drives that never die but I'd prefer not to take the risk. Also I'd need a whole secondary box to build the Z2 array as my current one is tapped out on SATA ports and drive bays.

 

[9C1 Newbee]

I wouldn't think you need HDD's for those builds. But if you have a need for 5 drives in other locations, do it ASAP. I would burn those drives in to make sure you are not adding faulty drives to the mix. Hopefully those will be the immortal ones you are looking for.

 

[marbus90]

I would run that whole system until it dies. Use it as ZFS replication destination with all the disks. Compare the checksums daily or at least weekly. As soon as checksum errors on multiple disks occur, throw the mobo+CPU+RAM away and replace it with an intel ECC system.

 

[cyberjock]

I kept my drives until I was forced to look at buying additional space. They were *all* out of warranty, and they all had a 3-year warranty.

ou are running RAIDZ1 though. I'd definitely make it a high priority to go to RAIDZ2 as soon as possible. As your drives age the will eventually fail and you definitely will wish you had that extra redundancy when that happens.

 

[jgreco]

When you start realizing that you might actually care about your data, then you might consider actually jumping up to RAIDZ3. Take it as an opportunity to back everything up, get a new mainboard, and add two drives to the system to get you up to Z3. Then relax a bit.