To make a long story short (not really), I have been copying all my data on to the array and finally finished.  Over 8tb's worth on a 13tb array!  So I decided that since I was not going to be doing anything more for at least a few days, I would shut it down.  Now then, last night I turned my UnRaid back on and I have lost one of my Seagate 1tb (model ST31000340AS) hard drives.  It will not spin up at all.  I have changed the power cables, and the sata cables.  I even tried both from the mother board, and the Sata card, just to make sure.

 

This is the 2nd of these drives (same model) that I have lost.  First one was one that I used only as a backup for my system that the UnRaid is to replace.  I thought that maybe I over heated the drive at some point using a estat dock.  But this one had never been used before in the dock, only in my Qnap 509 Nas.

 

I have a 2tb Hitachi that I RMA'd and awaiting, and the first Seagate that I RMA'd and awaiting.  The 2tb Hitachi wouldn't get through Preclear, and it would have replaced 2 of the 1tb drives and given me some backup drives for failures, but now I have none.

 

I am starting to get very worried about these drives.  I have 6 of these same drives in my array still, not counting the one that just fail.  I used the preclear on all of my drives before putting them in the system, so I kinda thought I would be okay, but one can't always tell.

 

So, I am not sure what to do.  Obviously, I need to get a hard drive as soon as I can.  But it appears that none of the local computer stores have anything in stock right now large enough (I am getting use to this).  Which means that it will be several days before I can get going, unless I next day a drive in.  Which I think I will probably do today with the Newegg sale on the WD15EARS.

 

But once I have my array back up and running, do I need to start changing out drives, or am I just worrying myself more than I need to?

 

To make a long story short (not really), I have been copying all my data on to the array and finally finished.  Over 8tb's worth on a 13tb array!  So I decided that since I was not going to be doing anything more for at least a few days, I would shut it down.  Now then, last night I turned my UnRaid back on and I have lost one of my Seagate 1tb (model ST31000340AS) hard drives.  It will not spin up at all.  I have changed the power cables, and the sata cables.  I even tried both from the mother board, and the Sata card, just to make sure.

 

This is the 2nd of these drives (same model) that I have lost.  First one was one that I used only as a backup for my system that the UnRaid is to replace.  I thought that maybe I over heated the drive at some point using a estat dock.  But this one had never been used before in the dock, only in my Qnap 509 Nas.

 

I have a 2tb Hitachi that I RMA'd and awaiting, and the first Seagate that I RMA'd and awaiting.  The 2tb Hitachi wouldn't get through Preclear, and it would have replaced 2 of the 1tb drives and given me some backup drives for failures, but now I have none.

 

I am starting to get very worried about these drives.  I have 6 of these same drives in my array still, not counting the one that just fail.  I used the preclear on all of my drives before putting them in the system, so I kinda thought I would be okay, but one can't always tell.

 

So, I am not sure what to do.  Obviously, I need to get a hard drive as soon as I can.  But it appears that none of the local computer stores have anything in stock right now large enough (I am getting use to this).  Which means that it will be several days before I can get going, unless I next day a drive in.  Which I think I will probably do today with the Newegg sale on the WD15EARS.

 

But once I have my array back up and running, do I need to start changing out drives, or am I just worrying myself more than I need to?

 

It sounds like a lot of failures... What power supply are you using?  It might be just your luck, or, it might be a power supply voltage issue.  It is not uncommon for multiple drives from a single manufacturing lot to suffer the same failures...  Perhaps that explains the two identical drives failing.

 

Joe L.

It is suspicious to have so many drives die in short succession, but not conclusive.  What are your system specs?  Specifically, what PSU?  Also, how's your cooling?  What are your drive temps like?  Excess heat and bad power are two easy way to kill drives.

 

You say the drive won't spin up?  Does it make any noises, like it is trying to spin up but is somehow stuck?  You might try the heavy-handed approach - place the drive flat on a table (label facing up), and give it a good solid smack with the palm of your hand just as it is trying to spin up.  I've seen drives with failing motors or bad bearings that just can't get past that first little bump.  Once it starts spinning (if it does), then you can transfer all the data off of it, then RMA it.

 

I would definitely grab that 1.5 TB WD EARS deal if I were you.

 

I don't know much about Qnap, is it a striped RAID system?  If so, then it likely wore all the drives that were inside it evenly, which increases the chances of them all failing in conjunction.  So if a lot of your failing drives came from a striped RAID system, then the close failures would make sense to me.

 

I hope you start having better luck  Otherwise you may have to change your name to 'Loose_pockets'

Could it be due to the seagate firmware bug?

 

Just a thought

 

Tony

Do you have your system on a ups?

sorry for delay on posting, tried to answer all question as they showed up, here goes

 

I am running the US version of the Budget build.  With the exception of the PSU.  Here I opted for the CORSAIR HX Series CMPSU-750HX, figuring that I would be re-doing my system to a 20 drive in the future.

 

Temp has been approx. 24C at idle and as high as 39 while running the system.  Preclears on 3 drives at once have seen temps as high as 43C.  Then I covered all of my unused fan ports and the temp dropped down to 39C, but most the they time stay in the low to mid 30's, depending on the room temps. 

 

As for the sound, hard to hear when 12 drives are all suppose to be starting, but I can say that it is the only drive that I can't feel when the drive start to spin up.  I could disconnect all other drives and try to start the system and know better?

 

The Qnap is a striped RAID, and the last drive came out of it, the other was one that I purchased just to backup some of my data.  This also means that they are not from the same lot.  The serial numbers are not similar, and they were purchased at different times of the year, and also from different venders.  The only reason that they are the same model is because they work well for me in the Nas, that I went with the same.

 

Their are 4 more that are the same lot as the 2nd one that failed.  And their are 2 more that are the same lot as the first failed drive.  I didn't know any better, so I would purchase several drives at once.

 

Tonight I can give the smacking of the drive a try.  This actually sounds like something I would do, but I have been trying to refrain from such behavior as I get older.

 

I have not ever updated my firmware on a hard drive.  The one that I have here says "firmware SD15".

 

I do have my entire system, theater equipment and all, on UPS and power conditioners.

As for the sound, hard to hear when 12 drives are all suppose to be starting, but I can say that it is the only drive that I can't feel when the drive start to spin up.  I could disconnect all other drives and try to start the system and know better?

I think that is a good idea.  Alternately, you can hook up the drive outside the case with the help of some long cables.  This will help you isolate the sound.

 

Tonight I can give the smacking of the drive a try.   This actually sounds like something I would do, but I have been trying to refrain from such behavior as I get older.

What can I say, sometimes mechanical components need a bit of tough love.  The key is violence with purpose.  Still, I'm sure the first person to figure out this method (it wasn't me) did so purely by rage-filled accident.

 

Also, this may sound a bit odd, but your drives may actually be too cool!  The ideal range is the 30s, or maybe high 20s to low 30s.  A widespread google experiment on consumer drives found that 20C was too cold, and that it actually decreased the lifespan of drives.  Your drives are also undergoing quite a differential, from 24C to 39C.  If possible, perhaps you could let them stay a bit warmer so that they only shift by about 5-10C when going from idle to active.

If you hadn't said Seagate, I would be concerned that it is something else in your system.  But you said Seagate, so multiple drive failures and dead drives after six months are the norm.

 

I've also heard of one other person who could not get a 2TB Hitachi to pre-clear using Joe L.'s script, but it would clear when they just stuck it in the unRAID and cleared and formatted without the pre-clear script.  After building parity, there were no errors in SMART, so I assume it was okay.

 

I'm also still iffy on EARS drives.  I know they work okay with the jumper, but I'm still concerned that there could be long term problems that haven't cropped up yet (like what happens when unRAID does start detecting 4k drives, or does the Sector +1 jumper fix jeopardize your data in some way).

 

Last Friday I bought some 1.5 TB Samsung Eco drives for $80 each.  I know that it $10 more than the EARS drives, but it may be worth it to stay off the EARS drives a little longer.

I'm also still iffy on EARS drives.  I know they work okay with the jumper, but I'm still concerned that there could be long term problems that haven't cropped up yet (like what happens when unRAID does start detecting 4k drives, or does the Sector +1 jumper fix jeopardize your data in some way).

 

As soon as unRAID can start using 4k drives natively, all you have to do is remove the jumper and let unRAID reconstruct the contents onto it.  You have to do this for each drive one at a time, so it will be time consuming, but it should work.

 

I assume you are staying away from the Samsung 4k drives for now since you are the one whose testing determined that they are problematic, right?

As soon as unRAID can start using 4k drives natively, all you have to do is remove the jumper and let unRAID reconstruct the contents onto it.

 

Nooooooooooooooo.  The jumper has NOTHING to do with the 4K sectors... it only affects the offset.

 

The 4K sectors are INVISIBLE to the OS in the 2TB EARS drives... the drive does 512b to 4K translation.

 

The FIRST drive to expose the 4K clusters to the OS is the (as yet, unreleased) WD 3TB drives.

If you hadn't said Seagate, I would be concerned that it is something else in your system.  But you said Seagate, so multiple drive failures and dead drives after six months are the norm.

 

 

Ridiculous and untrue. 

As soon as unRAID can start using 4k drives natively, all you have to do is remove the jumper and let unRAID reconstruct the contents onto it.

 

Nooooooooooooooo.  The jumper has NOTHING to do with the 4K sectors... it only affects the offset.

 

The 4K sectors are INVISIBLE to the OS in the 2TB EARS drives... the drive does 512b to 4K translation.

 

The FIRST drive to expose the 4K clusters to the OS is the (as yet, unreleased) WD 3TB drives.

 

I keep mixing those up, sorry.  Still, what I said above should be true once unRAID addresses the sector starting at 64 directly, right?

I have been very happy with my Seagates, until now.  But since I have never had a drive fail on me before, it's probably just my luck and I was due.  Still worries me.  Time will tell.

 

As for the smacking of the drive, I did it, and it felt soooo good!  But sadly, it made no difference at all.  I have a couple of the WD15EARS on their way and here tomorrow.

 

Now I have a question concerning the preclear of the new drive.  With one drive showing that it is missing, should I hook up the drive and start the preclear as I normally would?  Or can I disconnect all drives and only hookup the new drive and then preclear?

 

So my thinking was to disconnect them so as not to have them powered up for the number of hours that it will take to process.  It wouldn't change anything in my config to this, would it?  Of course, I copy my flash drive, as I have done each time I have maded any changes to the hardware, so if anything did go wrong, I can try to get it back to its previous state.

 

I just don't want to do anything that might cause another drive to fail. 

As for the smacking of the drive, I did it, and it felt soooo good!  But sadly, it made no difference at all.  I have a couple of the WD15EARS on their way and here tomorrow.

Yeah, that was more of a placebo effect I was hoping for...just kidding, it does actually work once in a while.

 

Now I have a question concerning the preclear of the new drive.  With one drive showing that it is missing, should I hook up the drive and start the preclear as I normally would?  Or can I disconnect all drives and only hookup the new drive and then preclear?

 

So my thinking was to disconnect them so as not to have them powered up for the number of hours that it will take to process.  It wouldn't change anything in my config to this, would it?  Of course, I copy my flash drive, as I have done each time I have maded any changes to the hardware, so if anything did go wrong, I can try to get it back to its previous state.

 

I just don't want to do anything that might cause another drive to fail.  

 

Preclear doesn't affect the array at all.  You can preclear up to 6 drives while running the array if you wish.  However, you concern is valid - I also wouldn't want to run my array in degraded mode for 30+ hours while waiting for new drives to preclear.  I would do as you've described: unplug all the array drives while the others preclear.  Before you do this, take a screenshot of your 'devices' page and save it on your desktop somewhere (anywhere but the server).  Then once you are ready to get the array going again, you can be sure which drives go into which drive slots.  unRAID will usually figure this out for you, but just in case it is good to have it written down somewhere.

 

The other option is to preclear the drives using a different machine and leave your main server powered off the whole time.  You could either use your primary unRAID flash drive to boot and then preclear the new drives, or you could prepare a separate unRAID flash drive.  I keep a spare flash drive around with unRAID Basic installed and ready to go for this and other purposes.

Well I think that I  am going to us my main system.  I won't be doing anything with it until the drive is ready.  And since I have 2 drives coming, I will place one on the exact same power and sata hookups that the failed drive was on (this was on sata card), and the other will be on a different power cord and the mother board sata slot.  If there is some other issue, it might show it self.  If not, then I will get to see the difference in the speeds of the Mother board and the sata card.

 

One more thing.  If anyone wants to know if the hot swap bays are worth it.  Heck yes!  Of course I don't have them myself, sure wish I did now though.  All the time spent, and all the cables to disconnect just to get one drive out really stinks.

 

They will be on my list of upgrades to do in the future.

 

My wife says that she doesn't understand my "call sign" at all.  I'm never tight about my hobbies.  She is right, I should have spent the extra.  But no, the Tight_wad in me said "No, just save a few bucks, that's just eye candy, you don't need those things."  Man, being a Tight_wad can really suck.

 

Thanks for all the help everyone! I will let you know what happens next.

Update

 

I have precleared and replaced the failed drive.  All has turned out OK.

 

Except!

 

Now one of my other drive makes brief sound like bacon cooking when it first spins up.  This is bad, right?

So should I wait and see what happens or should I replace it immediately ?  I can use my new extra drive, I precleared it already to try to be prepared for a problem.

Update

 

I have precleared and replaced the failed drive.  All has turned out OK.

 

Except!

 

Now one of my other drive makes brief sound like bacon cooking when it first spins up.  This is bad, right?

So should I wait and see what happens or should I replace it immediately ?  I can use my new extra drive, I precleared it already to try to be prepared for a problem.

Funny noises coming from a disk are usually not a good sign, regardless of when they occur.  At the very least,, get a SMART report on that drive and see if it is complaining about Spin-Up time.

 

Joe L.

...one of my other drive makes brief sound like bacon cooking when it first spins up

 

Haha, that's a new one to me....so your drive sizzles?

...one of my other drive makes brief sound like bacon cooking when it first spins up

 

Haha, that's a new one to me....so your drive sizzles?

 

Oh ya, laugh it up.  Wasn't it you that said "Smack it!"

I think he was.... 

Yes, I did recommend the 'tough love' procedure.  I'm not laughing at your misfortune, I just haven't ever heard a drive sound described that way.  Now that I think about it, I think I know what you mean though.  A drive doing a lot of random reads can sound sort of like it is sizzling.

Years back I had a couple of deskstars that got hot enough to cook bacon on them.

 

Last night, after I moved my drives in my case around, changing my slot order, I started playing with the sizzle issue.

 

I removed the spin up groups for each of my drives.  Then I spun each one up, on at a time.  No bacon sizzle from any on them.  So I spun down, reassigned the the group spin, ( spin all in one group ).  Spun them up, and I make bacon.  I think it might be the combined sound of each on that I was hearing.  So when is making bacon not really bacon? 

 

This weekend I will smart them and see what I have there.