Extra hard drive on hand?

[j5428]

I am curious, who keeps an extra hard drive on hand in case of a failure?  I had an issue the other night when one of my disks was disabled (Red Dot).  I was a little nervous and realized I had nothing to replace it with.  A reboot cured the problem for whatever reason.  I ordered another 750 gig just to have in case a drive fails.  Now I am tempted to install it. 

[JRS]

I've been contemplating this very same idea.  I don't have an extra hard drive on hand that matches my largest unRaid disk right now, but I am considering ordering one soon.  I think I'll mount it in the case, but not plug it in, then I am ready to do a cold swap whenever.

[WeeboTech]

I always have spares on hand.

When I buy a new set of drives, I usually buy them in pairs.

Sometimes I mirror them, sometimes I tuck one away.

When I take a drive out of service, I put it in a special box or in some external enclosure and save it for a rainy day (or eBay).

I'm about to eBay a boxload of my smaller ones as they have been idle for years and I need the room for larger ones.

 

 

[NLS]

Ain't this bad for your pocket?

 

Since unRAID eliminates the need for having SAME-SIZE-EXACTLY disks (even to replace a failed one), what is the point of doing this?

 

I'd consider it only if I was somewhere (if there is any such place in the civilized parts of the Planet) that I couldn't have a hard disk next day.

 

If a disk fails, I can go out to a shop same-day (or order on-line and have it next morning) and buy the biggest I can afford (and is in my GB/$ ratio of choice) and replace it. 99.9% of times, it will be a bigger disk or if I choose the same size (and still the size exists) it will be cheaper than if I already had it when I bought the failed one (as every couple of months the prices go down for the same GB) and I will also have more GB when I install the new disk.

 

After all it is a matter of answering two (more) questions:

 

1- If you have 2-3 different sizes of HD, you will have spare for all? For the biggest?

 

2- ...and then why don't you just add this to the array too?

 

(Hot) spare makes sense when you use some special disks that are not readily available (like some old USCSI3 on ID-marked hot-plug trays, that you had to order and wait for weeks) AND you have to have same technology/model disks to keep your array working.

 

None of the two applies on unRAID.

 

 

[Joe L.]

My philosophy is to have more space on the drives than data, so that when a drive fails I can offload any critical data to other drives in the array while I order a new replacement.   

 

Additionally, I have in one of my MG-35 media players a local hard-disk that is exactly the same size as my parity drive... if the parity drive were to fail, the media player would just go diskless until a new drive could be purchased (The disk is for when we travel and we use the MG-35 in the car and in hotels, otherwise, the MG-35 usually plays movies from the unRaid server on the LAN, as it has a far bigger selection)

 

I can usually get a new hard drive from outpost in a few days to replace any that fail.  I will replace any failed drive as soon as I can, but not at any expense... I'll still get as good a price as I can find at the time.  As you said, it makes no sense to keep exact duplicate drives on hand unless you can put them to use in the array now.  They will only get bigger and cheaper as time goes on...

 

Joe L.

[NLS]

from all this I keep this

 

I'm so paranoid about some of my data

 

 

I guess your time must be VERY expensive or you have much more money than I do (in the second case: nice going man ).

 

 

[WeeboTech]

Time is the most expensive asset.

You can never get it back. (I'm cost effective but not cheap either) LOL.

 

I have years of information on my hard drives. The cost of a spare hard drive is minimal vs the lost of some of my information.

(Or the time it would take to replace what can be replaced).

 

 

 

 

[NLS]

fair enough

 

 

[captain_video]

It all boils down to what you use your server for.  I use mine for storing rips of DVDs, HD-DVDs, and Blu-Ray discs.  I don't watch movies every day so if a drive dies it's highly probable that I can get it up and running again with a replacement drive before I actually need it.  Of course, if I was using the server for storing important data then that would be another issue entirely.  Having a spare drive on hand would just be a smart thing to do.

[SSD]

The incidence of true drive failure is incredibly low.  You are going to spend a pretty penny on an an extra hard drive equal in size to your parity drive ($225 - $250 for 1T) to sit on a shelf for 1 year, 2 years, 3 years, ...?  If history serves, the price for these drives will steadily decline.  By the time you need to use it, chances are your "great price" at purchase is now higher than you can buy it at CC or BB.  And you've gotten zero benefit.

 

If I have a true drive failure, I'd just turn my unRAID box off for a couple days and order a replacement drive.  Others might pay overnight shipping or even buy locally.  Chance are hugely in your favor that you'll come out $$$s ahead.

 

Just my $0.02.

[grandan]

Joe, mathematically you are exactly accurate.  BUT, last year I had a single drive failure in two different PCs i owned--so, of 3 PCs I had at the time with a total of 6 internal hard drives, I had 2 to fail in a period of less than 6 months.  One was WD (out of warrany) and one was Seagate (in warranty, so SG sent me a refurb in exchange).  On the WD, fortunately the only thing I lost was my e-mail files (I don't yet know why) because I did monthly backups to an external HDD.  On the SG drive, I lost about 60 GB of data out of about 360GB on the drive as I was able to transfer the rest off when the drive started acting up; all of which was ripped movies.  Since I own all my movies, this only cost me about 6-8 hours of time to repeat the backup of the 10 movies that were unrecoverable.  So, even if the math says that on average the MTBF of a HDD is 100,000 hours or even every 1,000,000 hours, some poor guy will improve everybody else's odds at his expense.  So, now I use unRAID to store all data and for file backups of my three PCs and use RAID1 drives for operating system, etc. in all but my wife's laptop--I don't like most of those recipes anyway.

[Billped]

If I have a true drive failure, I'd just turn my unRAID box off for a couple days and order a replacement drive.  Others might pay overnight shipping or even buy locally.  Chance are hugely in your favor that you'll come out $$$s ahead.

 

Similar for me, though I would probably just get the best price locally and not wait for shipping.  One has to consider, however, the cost/benefit of the two scenarios vs. the likely failure rates:

 

1. Buy a drive at a lower price but have it sit around for 1-1000 days

2. Buy a drive at a higher price but get it the day of the failure

 

It is an interesting calculation, since a sale drive can be 30-40% lower than the "crap, just gotta get one" price.  How long does it take to have drives overall go down by that much?  Probably a year or more.  Thus if you expect a drive failure in a year or less, it actually makes better fiscal sense to buy the drive now.  As a bonus, you suffer less downtime.

 

 

Bill

[SSD]

Similar for me, though I would probably just get the best price locally and not wait for shipping.

 

Although I'd certainlly look locally, if I had to spend $350 to buy locally what I could get for $225 mail order, the bargain hunter in me would make me wait at least overnight! 

 

One of the great features of unRAID is that it will continue to operate WITHOUT the failed drive.  If you have a need for some of the data and are chicken to just leave it running through the failure, you can power up the array, copy the data you need to a local hard disk, and then power off the array until you get my new HD back from the RMA process or wait for shipping of a mail order drive.

 

Another point, people complain about the need to buy 2 higher capacity drives to begin to use the next higher capacity drive in their array.  If you have to have a spare on top of that, you'd need 3.  Ouch!

 

[WeeboTech]

The incidence of true drive failure is incredibly low.  You are going to spend a pretty penny on an an extra hard drive equal in size to your parity drive ($225 - $250 for 1T) to sit on a shelf for 1 year, 2 years, 3 years, ...?

 

With unRaid I do not plan to spare out my parity drive unless it is cost effective or a common size in my network.

Yet I'll have 1 or more spares for the most common data drive size. (currently around 500gb).

 

I still do (and will continue to) setup raid1 for the operating system drives.

I just can't have some of my machines stop running.

So for these I always have a spare on hand.

 

In my case, the spare is not always idle sometimes it's a scratch/work drive.

With unRaid, I now mount them there and if need be.. shift them as needed.