Second Drive Failure in three months!

[FlyingPete]

Well, it happened my second drive failure in three months! This time it was a Maxtor (last time a Seagate), no SMART warnings on this one either, also this time my primary photo store!

Fortunately for me this was one drive in a mirrored set (RAID1), as it is my primary photo store, I take this stuff seriously.

It seems to be media errors, the RAID controller attempts to rebuild, but gives up around 60% after excessive block errors.

I also have two offsite removable drives, in case I did loose the lot.

There is a lesson here though, never rely on a single drive for all your photo storage, these things fail, I am just lucky I am yet to actually loose anything!

[tommykjensen]

RAID1 is NO guarantee! Always backup to secondary media, second PC, external drive, DVDs whatever.

I have had 2 drives in a RAID1 die at the same time!

[FlyingPete]

RAID1 is NO guarantee! Always backup to secondary media, second PC, external drive, DVDs whatever.

I have had 2 drives in a RAID1 die at the same time!

Eeeik :eek: Sounds like a failure in the often overlooked other part of the equation, the controller. That is why I have two offsites (I NEVER have both at home at once), mainly for house burning down scenarios, but also good for the one in a billion two disk failure (hey statistically it happens to some people, perhaps it is more common that we would like to think, as I have seen it happen twice, and heard of other instance like yours!).

You should never rely on a single storage ‘entity’ for you photos.

One thing I think people overlook is the MTBF’s quoted on drives are an engineering specification, they are designed to last that long in optimum operating conditions, in reality, my experience is more like 1 in 100000’s of hours rather than the millions often quoted.

If you take a MTBF of 1 in 100000 (typically what we see, I am a storage engineer/consultant for my day job), then that is statically a failure every 12 or so years, if you have four drives, you are going to statically loose one every three years.

Here is my setup:

Primary Workstation – 160GB SATA disk – All working and recently downloaded images stored here

Photo Server – 2x 120GB PATA drive in hardware RAID1 (mirror) – All sorted and process images stored here

2x Offsite 80GB PATA drives

I Robocopy everything to my server regularly whilst working, and is part of my camera upload procedure, so there is always two copies of the files. Each time I delete images out of my working area (every few weeks), and move them to the Photo store on my Server, I run one off my offsites. This means that I have at a minimum two copies of each photo, but at least three after a few weeks.

Currently my photo store is around 40GB (I delete anything not worth keeping to stop it getting out of hand).

[pcasciola]

You know, I used to lose a LOT of hard drives. I honestly think 90% or more of computer hardware failures are related to voltage fluctuations. A few years ago, right after yet another hard drive failure, I upgraded to high end power supplies instead of the cheapo ones that come with the $50-$100 cases, and I also installed a UPS which also serves as a line stabilizer, and I've never had a single hardware problem since. Now, when the lights flickers, my UPS kicks in and I just smile intead of sweating out another potential hardware loss. I bought the UPS less for the uniterruptable power aspect, and more for protection against voltage sags and spikes.

[gasrocks]

I just lost my hard drive without any warning....along with much data & pictures, of course. Thought I had been making back-ups often enough. Not so. Advice for all - back it up now! External hard drive and back-up software is on the way.

[FlyingPete]

You know, I used to lose a LOT of hard drives. I honestly think 90% or more of computer hardware failures are related to voltage fluctuations. A few years ago, right after yet another hard drive failure, I upgraded to high end power supplies instead of the cheapo ones that come with the $50-$100 cases, and I also installed a UPS which also serves as a line stabilizer, and I've never had a single hardware problem since. Now, when the lights flickers, my UPS kicks in and I just smile intead of sweating out another potential hardware loss. I bought the UPS less for the uniterruptable power aspect, and more for protection against voltage sags and spikes.

Yep, not 'optimal' conditions for those high rated MTBF's. My PSU smoked (literally) a drive a few years back too!

I think optimal conditions are perfect power, around 21degC around the drive, probably around 65% humidity, no vibrations at all, I doubt anyone here has a machine that can provide that, few Datacentres I know of can!

The server this drive was in has a good PSU, and is on UPS, so it shouldn;t have been a power spike.

[DocFrankenstein]

I don't think humidity matter for harddrives at all, they're supposed to be sealed.

I still have all of the pics on a single drive. Can't get around to backing it up.

That UPS is an interesting idea, how much do they go for?

[FlyingPete]

I don't think humidity matter for harddrives at all, they're supposed to be sealed.

Most harddisks have brething holes on them (for atmospeheric preasure changes), very few are completly sealed, there is a filter there though to stop dust getting in. THe main issue with humidity is if it is too low, you can get staic build up, which is very bad for electronics, and if it is too high you can get condenstaion if you are moving equipment around.

I still have all of the pics on a single drive. Can't get around to backing it up.

Pleeeassse say you are kidding! I have heard too many horror stories, and almost one divorce over this! It doesn't even have to be a failure, one guy I know was having issues with upgrading his harddisk, and accidently erased the wrong one, no backuk, and he had the first four months of baby photos on there! I hear these stories all the time, another local guy was trying to sue the computer shop he bought he PC from after he lost everything in a failure, you even hear the stories right here on this forum!

[joeseph]

I haven't ever had a drive failure, although due to brain failure I deleted the FAT table & partition info from a drive a year or so ago. Learned a lot about recovery at that time!
Now I have two machines, one running 24x7 (webserver) and one that I use for everything else. weekly copy my docs folders between the two & occasionally to a USB external drive that I keep at work. Seems to work for me at this point although could still lose a couple of weeks worth of data of the house (lord forbid) burns down. My normal process involves copying off flashcard to an external HDD (30Gb Tripper) so I've usually got around a month or so kept there as well.

[Harry Settle]

Installed a battery backup/surge supressor system a couple of years ago and quit having problems like this. It isn't the solution to everything, but it is the answer to a lot of things.

[pcasciola]

That UPS is an interesting idea, how much do they go for?A good APC brand UPS is around $100. The newer ones have USB and pretty good software, too, so set it to shut down your PC gracefully when the power fails. I've had mine for a few years now, and recently replaced the battery for about $40. Only gives me a few minutes of power, but enough to save anything I am working on at the time.

But you GOTTA backup your photos and important data, at least to a second external drive. You can catch 120Gb and 160Gb drives on sales for $40-50 these days. Throw that in an external case, run a program like Mirror Drive and you will never have to worry about it again.

I use three external drives in addition to DVD backups of my really important data. Paranoid? Maybe. But I've lost too much data in past to let it ever happen again.

[FlyingPete]

I use three external drives in addition to DVD backups of my really important data. Paranoid? Maybe. But I've lost too much data in past to let it ever happen again.

There is no such thing as paranoid in the storage business. Unfortunately it usually takes a bad experience to make people consider backup.

I've got a customer that has a synchronous copy between two sites of all data, so any writes at one site, also get written at the other, they worked out that by the time they restored off backup, re ran the logs to get them back to where they were, they would be out of business!

I was going to mirror between a friend of mine and my house over DSL, but I go away and come back with 10GB of data far too often to be practical. Eight days to replicate that due to our restricted upload speeds, too much hassle to wander of there with a HD.

[mdm]

I backup to a external 120 gig. I need to back that up with some dvds to be on the extra safe side. Then I'll feel ok when they are in the fire-proof safe.

[Citizensmith]

Of course this reminds me I need to go replace my UPS. Supposedly a 15 minute unit, and it did used to be. We had a power out a couple of weeks back, the UPS started bleeping away so I started to shut down. I hadn't even made it to the shutdown dialog before it gave up.

Or, its all very well spending the money on UPS and external HDDs and such. Just make sure you use and maintain them! Otherwise your backup will fail when you need it most.

[FlyingPete]

Of course this reminds me I need to go replace my UPS. Supposedly a 15 minute unit, and it did used to be. We had a power out a couple of weeks back, the UPS started bleeping away so I started to shut down. I hadn't even made it to the shutdown dialog before it gave up.

Or, its all very well spending the money on UPS and external HDDs and such. Just make sure you use and maintain them! Otherwise your backup will fail when you need it most.

Hmmm must do the battery in mine, good for around 30 seconds at the moment! Only good for power fluctuations, which is what I get the most of.

[robertwgross]

A buddy of mine is a design engineer at one of the large hard disk companies. He claims that the two worst things for hard disk life are:
1. Heat
2. Power fluctuations, especially spikes

---Bob Gross---

[DocFrankenstein]

Yeah, I've heard about heat too... that's why I keep my computer open. Makes me wonder when I get cancer on my right leg (the one next to the comp), cause my 2.4 gHz is pretty much the frequency of a microware oven. :lol: The power isn't the same, but still...

I don't have a copy, cause I have a few CDs which are the "best of pics" of my collection... and most of the worthy stuff I have hard copies of.

[pcasciola]

1. HeatYES!! I forgot about this. At the same time I upgraded my power supplies and installed the UPS, I also switched to two aluminum cases with like 99 fans in them. One of them sounds like a small jet, but it runs cool.

I also never use those really small external drive cases because there is not enough airflow around the drive, so I use the ones that are the size of a half height 5.25" bay with a built in fan. The kind where you can also put DVD-RW/ROM drives in them if you want. Never had a problem with one of those, but I have had hard drives overheat in the smaller cases.

[FlyingPete]

1. Heat


Yep probably the main cause of my weekend failure, I recently installed a third no RAID drive for all that unimportant stuff (mainly CD ISO images, so I don't have to hunt down the disk each time!). Case side is off so air circulation has probably bee quite poor, must load that box up with some fans!

[CyberDyneSystems]

RAID-1 = GOOD!
But should not be only solution. Pete,. glad you had it mirrored! I too will not live without it.

High Quality Power supply = GOOD!
Phil, I totally agree. Get a good power supply and you'll notice everything tends to last longer!

Back-Up Solution:
I recommend RAID-1 as your first layer of defence against a single hard drive failure.

I recommend a third hard drive in an external enclosure:
Use it for full automated backups of your image data. (and whatever other data you cherish)
Turn it on at least once a week,. or every time you alter the data on the RAID-1 array.

I recommend manual backups to DVD-R or CD-R:
Every time you copy a days shoting to your RAID-1 array.
It's best to make two copies, and check them/verify.
It's best to keep one of these copies off site.

[felix21685]

so your saying a 3.2 ghz with a nvidia 5900 ultra in a small shuttle case is too warm without a fan ?

pfff...lol (jk) i dont have the shuttle case..those things are so dumb..for my application

[tommykjensen]

Now I have two machines, one running 24x7 (webserver) and one that I use for everything else. weekly copy my docs folders between the two & occasionally to a USB external drive that I keep at work.

Oh no! I hope Your important data are only unbackedup on the webserver for a short time (minutes)! Remember the webserver is on the internet and can be hacked at any time!

[tommykjensen]

I've got a customer that has a synchronous copy between two sites of all data, so any writes at one site, also get written at the other, they worked out that by the time they restored off backup, re ran the logs to get them back to where they were, they would be out of business!


Several years ago the worlds largest containershipping company had to move its datacenter from Ney York to Copenhagen. The amount of data that had to bo moved was counted in TeraBytes and it was estimated how much bandwidth were required to copy the data via network. The results was staggering and would have been extremely expensive not to mention that it would have taken weeks to copy the data with the available datalines. Instead all the disks was first mirrored on site (they were already mirrored but a third copy was required). The 2 copies were shipped on a plane. From the site to the airport in NY the 2 sets of disks were shipped with 2 different trucks. On arrival in Copenhagen they were also shipped with 2 different trucks to the destination. The operation went well with time to spare and the story made local news.

[joeseph]

I mean the data is backed up to a drive on the webserver machine - it isn't online at the same time as the web side of things! I keep a fairly close eye on the logfiles and frequently use a sniffer to see what's happening realtime. perils of being a techie I guess.
Come to think of it, I haven't Ghosted the webserver for a few weeks - now sounds like a great time to do it...

[Wazza]

Bugger Pete.

I've just changed my setup again.

I've got a 200Gb and a 120Gb on this current setup, and had a 80Gb spare.
With the old system it is, and lack of money, I was just going to be taking out the drive, and copying it through when need be. Obviously this wasn't up to date enough. So I put the spare 80Gb into the main household PC box, which already holds a 80GB, and transferred all the pics through the network. Took a few hours to transfer 50Gb of pics from 20D.
Now I'm copying data through every time I've done some shots. Obviously, I'm going to be out of space soon on the 80Gb. Might just buy another cheap 200Gb.

"Why don't you delete all the unwanted pics" - The family asks.
Haha, have they not realised there are 25,000 shots to sort through. Even reviewing them every 2 seconds - delete? Yes/No, would take a whole day. :eek:
So the easy solution, don't delete and buy more storage. :rolleyes:

[napolar]

Ok, I have a Compaq server with RAID 5 as my main data store. I back that up to a DLT drive and keep the most recent copy in the safe deposit box at the bank. I, also, keep a second copy of my pic's on my laptop and a third copy on my workstation at work. The server and tape drive at home are on a UPS. The only thing that I do not have at this time is a "hot spare" for the RAID array.

I have seen too many drives fail without warning and I do not want this to happen while my pictures are in a single location. The one thing that I have learned is that if you have a single poit of failure you are begging for trouble.

[lost]

Hey docfrankenstein Ihave seen where opening up the computer actually worsens the heat problem. Your case is designed to pull air through from front to back. When it is open the fans dont create the air flow over important components.

For example. My case has a fan in the front that sucks air through the front cover and one fan in the back that blows air out the back. With the case open my processor runs a good bit hotter than with it closed.

Just an FYI. If you are conscerned about heat add more fans.

[FlyingPete]

Hey docfrankenstein Ihave seen where opening up the computer actually worsens the heat problem. Your case is designed to pull air through from front to back. When it is open the fans dont create the air flow over important components.

For example. My case has a fan in the front that sucks air through the front cover and one fan in the back that blows air out the back. With the case open my processor runs a good bit hotter than with it closed.

Just an FYI. If you are conscerned about heat add more fans.

Most high end server gear actually spins up the fans to full speed when the case is opened due to the loss of cooling efficiency.

Also watch out adding addional fans, make sure the air flow is correct, I use a 'smoke test' usually a blown out candle of match to observe airflow. One thing that was quite common a few years back was the PSU fan and case fan were at odds with each other, net result little of no airflow.

[Mudflap66]

Philip is right folks. I have had a line conditioner for the past 15 years. Also fans, fans, fans everywhere. My new desk top has 9 fans in it. I have never lost a motherboard, harddrive or any other equipment in thoes 15 years. I turn off the system when not in use. (boy that last statement will start a war)
Good forum I sure enjoy it.
POP

[blue_max]

Single drive, no backup, computer next to a radiator.

Simple solution – buy another computer. Backups are for wimps.:lol:

nb I bet people didn't dupe trannies as a backup in the olden days! (mine are in my sock drawer I believe).

Graham

My "next" pics will always be better anyway, so no great loss. ;)

[Avalonthas]

ive gone through anywhere from 50 to 75 different harddrives since i got into computers many years ago and have yet to have one fail. I currently have 2 500gig, 2 250gigs, and 6 100gigs always running, and no problems as of yet. I usually get a new set of harddrives every year though so there relativly new each year. All free of course, so i dont mind so many of them, but i always backup every shoot on a DVD. I got about 300 DVD's in my basement from all the major shoots i have done.

[DionM]

I am planning to move away from a RAID system. After my second controller hiccup I have lost confidence in the integrity of the data being written.

I will replace it with a USB case or maybe a removable HDD caddy.

My HDDs in my PC sit in a special HDD cage with fan blowing directly across them (Antec 1080AMG case). Keeps 'em nice and cool.

My worst HDD failures have been two Maxtor drives. That was my first and last experience with Maxtor. I have about 10 Western Digital drives scattered across my PCs, ranging in size from 20GB to 200GB and no problems. I also have only 2 Seagate drives (20GB and 80GB) and the 20GB just died a few weeks ago.

[FlyingPete]

nb I bet people didn't dupe trannies as a backup in the olden days! (mine are in my sock drawer I believe).

No, but they did often store them in 'safe' places (I think a sock drawer qualifies!); also trannies don't have a habit of self destructing, as they don’t ‘fail’ (what is the MTBF of one?).

I do know a lot of people who store their negatives and prints in two different locations. Anyway digital photography is an enabling technology, it enables us to do things we didn’t used to be able to do or was not practical to do, being able to make multiple copies of the ‘negatives’ is one of them.

Perhaps you images are not that important, but the first 18 months of my little girl growing up is all digital, for her sake and ours (my wife and me), I need to protect that data. If I only had sunsets, landscapes etc, things that I will get another chance at, I wouldn’t be that worried.

[Citizensmith]

Perhaps you images are not that important, but the first 18 months of my little girl growing up is all digital, for her sake and ours (my wife and me), I need to protect that data. If I only had sunsets, landscapes etc, things that I will get another chance at, I wouldn’t be that worried.

I agree. The content of the photos can be much more important than the actual quality. Unless you make your primary income from photography the family photos are what really needs protecting. Digital technology has given us the ability to easily copy our albums. It has also given us the ability to watch the instantly disappear.

[FlyingPete]

Oh cr*p, there goes the second drive now, bad blocks, some photos unreadable, haven't got the replacement drive yet, so I am building a new box now, a nice well cooled box with a good PSU.

Luckly I have offiste backups, so still no real data loss!

As someone around here said, RAID1 is good, but you still need more!