Storing Spare Drives in Server


Recommended Posts

Just wondering if anyone has insight into whether this is a good practice.

 

I ask because I recently had a pre-cleared Hitachi Green 2TB Deskstar 5K3000 with less than 50 hours on it fail as unRAID was rebuilding data on it. The drive had been mounted in the server, unconnected, for 2 1/2 years before I finally needed to use it to replace a 1TB drive that was full. It seems the controller failed. I could feel the motor spinning up but the BIOS would no longer recognize the drive. The good news is Hitachi replaced it under warranty.

 

I suppose this was just a random failure, but as I have never had any other hard drive fail  for as long as I've been using and building PCs (~25 years), I'm just wondering if there is any chance that storing it in the server drive bay might have contributed to the failure. The server is in a cool basement and the active drive temps are never more than 30°C, so I don't think it's likely this drive was ever above 24° except when it was initially pre-cleared. But maybe there was a static charge built up in the controller, since it wasn't connected?

Link to comment

I'm not a hardware engineer, but I do have a spare drive connected (warm spare). The longest time I left the drive was about 11 months. Called into service and running like a champ. So to answer your question it might have been a fluke and  awesome that it was still under warranty.

Link to comment

I had 2 drives in my 5 in 3 drive cage as spares.  Neither one was connected at all.  One of my data drives was taking read errors so i swapped it for one of my spares.  Rebuilding data died after 35 minutes on this new drive due to write errors.

 

The other disk worked just fine.  Mine too is in a cool basement and the system never gets above 25 degrees C.  So i had similar results as you.  Luckily replaced both disks under warranty.

Link to comment

I'm not sure if this still applies but in the past leaving a hard drive sit for extended periods, like a year, would cause the drive not to spin up correctly. Many drives were damaged by just sitting. Running the drive every now and then was always better and became just good practice.

 

I would be interested to see if we could back this up with some evidence. I don't disbelieve you it is just the ramifications are significant.

 

Wont be the easiest thing to google though

Link to comment

It's happened to me and a number my friends.. The trick to starting a "stuck" drive was to whack it on the side to start it and don't shut it down until you back it up. A friend of mine did that and ran the drive for a couple years never shutting it down until it finally die.

You may laugh at that low tech approach but I can't tell you how many times it worked.

Yes "stiction" is the correct term.

Link to comment

I just had this exact same thing happen to me as well. Just before the flood I bought 2 x 2TB Hitachi 5k3000 HDDs for ~$55 each (with rebates). I ran three preclear cycles on both drives and they had no errors. One sat unconnected for about a year before I added it to the array and it has worked just fine. The second drive sat unconected for 2.5 years and then I moved it to my desktop for use there. It died immediately. It had less than 100 hours of power on time. I RMA'd the drive and it was replaced.

Link to comment

It is very easy to google, stiction is the term and you will find lots of details. My second result was a video. This is well known and addressed.

 

Maybe your google foo is better than mine but I cannot find a single study to back this up.

 

I have no doubt this can happen but what I am interested in some real evidence to the scale it does happen.

Link to comment

I have no doubt this can happen but what I am interested in some real evidence to the scale it does happen.

 

+1!

 

This concerns me if this is really an issue! I have 4 spare drives sitting unplugged in my server waiting to come into play when required, and it never occurred to me that they could be damaged by not being on.... Seems counter intuitive to me... If there's no power going to the drive, and they're not being banged around, I'm not sure I understand the logic... low-level vibrations maybe?

 

Makes me want to plug them back in and keep them running as warm spares...

Link to comment

I just had this exact same thing happen to me as well. Just before the flood I bought 2 x 2TB Hitachi 5k3000 HDDs for ~$55 each (with rebates). I ran three preclear cycles on both drives and they had no errors. One sat unconnected for about a year before I added it to the array and it has worked just fine. The second drive sat unconected for 2.5 years and then I moved it to my desktop for use there. It died immediately. It had less than 100 hours of power on time. I RMA'd the drive and it was replaced.

 

Yep. Sounds like the same syndrome. I also had one that worked after about a year, but the second one that was stored longer failed.

 

But to be clear, were your drives mounted in your server? If they were just precleared and stored elsewhere that a would be evidence that the failure I experienced was not related to storing the drive the way I did.

 

Regarding the "stiction" issue raised by others. It is interesting to learn about but in my case it appears it was the controller electronics that failed, not the platter motor or heads. As far as the BIOS was concerned there was no longer any drive to communicate with, as though it wasn't plugged in at all. And of course the drive did work when it was initially powered up from storage, in fact it worked for over twenty hours before failing completely.

Link to comment

I just had this exact same thing happen to me as well. Just before the flood I bought 2 x 2TB Hitachi 5k3000 HDDs for ~$55 each (with rebates). I ran three preclear cycles on both drives and they had no errors. One sat unconnected for about a year before I added it to the array and it has worked just fine. The second drive sat unconected for 2.5 years and then I moved it to my desktop for use there. It died immediately. It had less than 100 hours of power on time. I RMA'd the drive and it was replaced.

 

Yep. Sounds like the same syndrome. I also had one that worked after about a year, but the second one that was stored longer failed.

 

But to be clear, were your drives mounted in your server? If they were just precleared and stored elsewhere that a would be evidence that the failure I experienced was not related to storing the drive the way I did.

They were stored in my server but not connected to power. My server is located in my basement which stays pretty cool (~18°C) year round. The temps inside the server case are probably in the low 20s.
Link to comment

I just had this exact same thing happen to me as well. Just before the flood I bought 2 x 2TB Hitachi 5k3000 HDDs for ~$55 each (with rebates). I ran three preclear cycles on both drives and they had no errors. One sat unconnected for about a year before I added it to the array and it has worked just fine. The second drive sat unconected for 2.5 years and then I moved it to my desktop for use there. It died immediately. It had less than 100 hours of power on time. I RMA'd the drive and it was replaced.

 

Yep. Sounds like the same syndrome. I also had one that worked after about a year, but the second one that was stored longer failed.

 

But to be clear, were your drives mounted in your server? If they were just precleared and stored elsewhere that a would be evidence that the failure I experienced was not related to storing the drive the way I did.

They were stored in my server but not connected to power. My server is located in my basement which stays pretty cool (~18°C) year round. The temps inside the server case are probably in the low 20s.

 

Thanks for the confirmation. That's two data points - with one drive model only - not enough to draw a general conclusion, but until I hear from people who have been storing spare drives mounted and unplugged without any issues, I'm going to be hesitant to do this again.

Link to comment

This is very interesting.  I wonder it moisture has anything to do with is.  It sounds like most of you are in a basement with a low amount of heating, the humidity might a bit higher.  I wonder if storing in an airtight bag with moisture absorbing desiccant similar to how the drives are shipped.  All the accounts here are antidotal , but it feels a bit higher than people talking about getting a new drive out of the box and it fails in 30 minutes.

Link to comment

Since becoming an unRAID user, I have always kept a spare drive the same size and model as my parity drive for prompt recovery from any drive failure.  I normally keep it in a 5in3 slot that I also use for preclearing.  When it is stored there, it is powered down.  I've never had any issues with these spares but I have always put them into service in less than a year.  The server lives in a finished/heated/cooled basement with ambient temps of 18-22C (65-72F).  Drive temp range from 23C (idle) to 35C (parity check).

Link to comment

I was thinking more about this over lunch..  I wonder it the right practice would be to go in and fire up the drive once a month and let it run..  then I was wondering how to wire in a switch.  I did a quick google search and came up with this.  It is a switch designed to turn drives on and off so you can boot up to another OS, etc.  You could could use this to fire up a drive once a month with your parity checks. They have a 5.25" and a 3.5" version.  How long does a drive run spin if it doesn't have a data cable attached?

 

ORICO HD-PW4101 3.5" Floppy Drive Slot 4x SATA HDD Power Switch Control (Black)

 

flips

Link to comment

Humidity can be more of a factor than ambient heat for magnetic media especially in when combined by heat. Watch out for mildew growing in those conditions, like in a basement with warm humid weather. Going through my fathers basement recently I can't tell you how many floppies and cassettes, yes remember those, and CDs I had to dump because of mold and it's cool down there. I even have it in a heated, dehumidified and well lit basement that appears to be dry as a bone.

One other issue I see with electronics when not used and warmed from use is that humidity affects the connections, like the connectors and circuit board fingers, even the gold ones, building up an oxidized layer resulting in high resistance at the connectors.

Just all food for thought.

Link to comment

I have no way to back this up, but I believe the old problem of stuck motor bearings from dried lubricant are well behind us.  I have not heard of any problems with stored HD's in a while.  Of course the abcense of evidence is not evidence, but I really don't know that this is a problem anymore.

 

Also for the average user storing a drive or two as cold spares should not be a problem.  At some point you'll need more capacity so you should press those cold spares into service and buy new drives as cold spares.  So you'll have a nice rotation going and the drives won't sit all that long.  At least that is what I've just done [shrug]

Link to comment

At some point you'll need more capacity so you should press those cold spares into service and buy new drives as cold spares.  So you'll have a nice rotation going and the drives won't sit all that long.  At least that is what I've just done [shrug]

 

This is generally how I go about it, though with the last couple of sales on drives, my cold spares have started to outpace my space requirements.

 

I suppose it could be worse!! ;):

 

I hope not I have over 200 drives on the shelf.

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.