A newyear reminder for all Orbiter-Forum members.

[Keatah]

Well folks, it is the beginning of a new year and it is time to take time to make a data backup plan! All it takes is one blast of lightning, a robbery, a hardware malfunction, a flood or fire, a stupid mistake by you, a hard drive crash, a nasty mal-ware virus.. Anything like that, and you're precious data is gone, forever.

I suggest to you all to keep a backup copy of all your important stuff, whatever it may be. Stuff that you cannot easily replace or stuff that CANNOT be replaced.

That could range from --
A personal journal
Thousands of mp3's you've ripped and saved over the years
Purchased iTunes
Photos from special events
Business records & financial records
Custom Orbiter installations with add-ons
Classic gaming emulation files you've saved from the last century
Daily records from the home business
Your school papers and work
Files from your development work
Your pr0n collection
Your FBI, CIA, NSA, KGB, Scotland Yard, and TSA records.
Local police department, Boarder Patrol, and civil court records.
High-score records from your favorite video games!
Anything!

This elaborates on the topic a bit more --
http://www.carlosconsulting.com/downloads/Backup_and_Recovery.pdf
http://www.en.wikipedia.org/wiki/Backup
[ame="http://en.wikipedia.org/wiki/Data_Recovery"]Data recovery - Wikipedia, the free encyclopedia[/ame]



You can make a backup plan as simple as this:
1-Open windows Explorer
2-Insert USB keydrive
3-Copy your important stuff
FINISHED!

For bonus points test your newly backed-up files by copying them back to a bogus temporary directory to see if the keydrive can be read. I had a keydrive that failed that test! Boy was I steaming pissed off! And the police came to see what the commotion was all about.

Or you can get a little more complex and follow this plan -- http://forum.acronis.com/sites/default/files/forum/2009/08/3084/Posted_to_Forum.pdf
-- Good god!!! How in world can that be followed?!?!? But hey whatever floats your boat or gets you deltaV.



My own personal plan consists of this:
1- Weekly copies of new material and stuff - takes a few minutes!
2- Every now and then I image the entire computer and make a copy of my "Library of Stuff" and take it off-site. Takes about 2 hours.
As you can see, it is a simple, effective, and easily followed plan.

And these are the preventative maintenance tools of choice for *me*. I have tested them and they work correctly for what I am doing. What tools you use and how you use them can and will be different. Be sure to test your tools and utilities to your satisfaction!
http://softology.com.au/dirsync.htm
http://sourceforge.net/projects/freefilesync/
http://www.acronis.com/homecomputing/
http://www.wdc.com/en/products/external/
http://sandisk.com/consumer-product...um=ProductUsage&utm_campaign=HomePageTracking

Enjoy!

 

[T.Neo]

Yes, that is a great idea, but what if I have gigabytes of data and relatively little storage space and/or it takes hours to transfer?

 

[Artlav]

You can make a backup plan as simple as this:
1-Open windows Explorer
2-Insert USB keydrive
3-Copy your important stuff
FINISHED!

0-Seep thru 2Tb of data (no porn...?...!), and find all the stuff that you made yourself.

 

[Keatah]

@t.neo - this is where you might use a filesync program. Once you make the initial backup of everything you want to keep you *then* use dirsync or filesync, or something like ms synctoy.. This style of utility will backup only the changed files. And that can only take a few minutes in most cases. You can run this everyday. The initial copy might take 3 hours depending how much you have. But subsequent runs of just copying the new or changed stuff goes fast!

I have a 1tb library of all sorts of stuff. When I run a file synchronizer it looks at all the differences and copies maybe 5 or 10 megabytes on a typical backup run. And it goes by in a few minutes. Much of the time is spent examining the differences based on date/time/filesize/CRC methods. It is a one-click affair.

1-connect usb drive
2-click on the icon
3-visit orbiterforum.com
4-remove usb drive and take offsite next time I'm out.

As for storage space, I would still recommend getting a removable hard disk of a sort. Or a used drive, you can "certify" it yourself with HD-Tune or the manufacturer's low-level diagnostic utility. That should be a cheap option. You *can* use CD's but that is a real pain, and they might not last as long as a hard disk in some situations.

How many GB (or TB) of data are we talking about here? And what is your file structure like? What hard disks do you have available or can buy? Maybe a $9.00 usb interface to a re-purposed and re-certified disk from fleabay is an option?

@artlav Ah yes, 4 steps. I kinda thought step 3 would cover it, copy the stuff you want. Decide as you go along.

@ALL In order for backups to work they must be simple to do or you will never do them! And remember, the point of backups is to simply have a 2nd copy of your information in case of hardware failure. This isn't rocket science!!

---------- Post added at 05:26 AM ---------- Previous post was at 05:23 AM ----------

I am doing an experiment, I have a 10meg hard disk (Yes I said 10-Megabytes) from like 1982. It has not been accessed or powered up since then. Later this year I shall do so and see if I can read everything back from it. I expect nothing but complete success! I have more confidence in this 30-year old mechanical drive than I do the dye used in cd-roms. And CD-R/W discs are the absolute worst medium along with floppies to use in backup ops. I have "lost" so much material to CD-R/W it isn't even funny anymore. Flash based media isn't too good either. It is perfect for under 10 years, but beyond that? Who knows..

 

[Pyromaniac605]

I wish I could, the problem with doing this for me is that I have close to, if not a terrabyte of data, nearly all of which I would prefer to keep, spread out over one 500 Gb and a 1 Tb drive, with no other hardrives in the house that work, or aren't being used, I have nothing to back up to, and I'm not going to go out and spend lots of money (I have no clue how much hardrives are) to back up my data.

 

[Keatah]

I wish I could, the problem with doing this for me is that I have close to, if not a terrabyte of data, nearly all of which I would prefer to keep, spread out over one 500 Gb and a 1 Tb drive, with no other hardrives in the house that work, or aren't being used, I have nothing to back up to, and I'm not going to go out and spend lots of money (I have no clue how much hardrives are) to back up my data.

Save up $0.50usd a day, and before summer you can get a 2tb drive!
Or perhaps just save your irreplaceable stuff.

 

[Pyromaniac605]

I'm sorry but, no, I'll just take my chances with looking after my hardrive.

 

[Rtyh-12]

Yeah, except for a virus, nothing can really happen to your data.

 

[Artlav]

Yeah, except for a virus, nothing can really happen to your data.

Viruses are the least of the worries, really. Modern ones are as lame as modern programmers.

More dangerous are accidental deletion (what is Linux? Why not try it out... WTF?), or some hardware issue, like tripping over a wire, letting the system take a flight. Or a thunderstorm. Or someone using a nuke as a firework. Or kids. A curious child can take a hard drive apart, or start banging that thing that keeps his father from playing with him.

Many possibilities here.

 

[Rtyh-12]

Yeah, but many of those things happen only if you're extremely unlucky...

 

[Keatah]

Modern Virii ARE indeed very very lame. The worst that happens is they slow down your system or flood you with pop-ups. Rarely do they damage data or anything. This is because when they *DO* that, they don't replicate very well. And the idiot virus writers of today, they don't think about anything stealthy or working quietly.

I designed one years and years ago. It was small, insidious, and worked quietly. My key goal in engineering it was to be sure it did not slow the system down. I'm a white-hat and have no interest in messing you up. So I never let it out. Anyways..

It would look at the following formats..
.doc
.jpg
.txt
.xls
.rtf
And it would randomly flip a single bit from time to time in the target file.

I later spiffed it up to be sure it would work on only the user data contained within the file. I made sure it avoided the structures like headers and other formatting information. It stayed away from the operating system and other critical files necessary to the operation of your system.

What good is a virus if it destroys its host? Which is what too many "attacks" do today. They slowdown the system big-time. And you are then "alerted".

My other favorite ones were the ones that simulated bad hardware. We had developed one that would randomly place a dot here and there on the screen once in a while, making you think you had a bad videocard. Another would slow down the hard disk for a while every few hours, making you think your hard disk was going bad. We upgraded it to intercept and report bad s.m.a..r.t. data. My other simple one, one of the smallest, would make it look like you mistyped a key, it would simply replace an adjacent key with another one. You type H it tells the o/s you hit J or G, for example. And a variant on that would be a missed key, it would ignore a keystroke after say 300 keypresses or whatever number you set it to.

Fun times indeed.

 

[garyw]

My rule for backups is simple - Only backup that which you cannot recreate or redownload. So, Orbiter addons can be downloaded again. The scenario file cannot.
Various documents can't be recreated easily so get backed up but the operating system doesn't because it can be reinstalled.

By following this approach my actual backup sizes are quite small.

Finally, if you are serious about backup then you should keep a copy of your data offsite. I use Amazons S3 for my data and so far it's been very, very cheap and pretty much flawless.

 

[Rtyh-12]

Keatah, you're a genius. People really should make these and let others download them...provided they can be deleted, of course...

 

[orb]

It would look at the following formats..
.doc
.jpg
.txt
.xls
.rtf
And it would randomly flip a single bit from time to time in the target file.

I later spiffed it up to be sure it would work on only the user data contained within the file. I made sure it avoided the structures like headers and other formatting information. It stayed away from the operating system and other critical files necessary to the operation of your system.

You have mistaken the terms virus and malicious software. Virus isn't a standalone program, and its design isn't to destroy data or annoy user, but to replicate itself on the system and between systems, by attaching its code to executable part of the host file (hence not standalone).

 

[Keatah]

Yeah, except for a virus, nothing can really happen to your data.

www.ontrack.com
www.drivesavers.com
File it away for future reference!



Out of 35 or so of my personal drives, over the past 10 or so years, 5 of them failed. One just starting clicking when I powered it on. The drive was no-longer able to calibrate itself on power-up. It was a 6-month-old WesternDigital WD600AB (60GB) stuffed into a suffocating 1394 FireWire enclosure. It was one of the first of the external drives in the late 90's I think. Upon forensic examination we found it failed for two reasons.

1-It did not have an extra spindle stabilizing screw, important on a physically large platter. So the platters began to express some precession.

2-It was in an enclosure that had little or no ventilation, caused overheating of the bearings which allowed the precession to worsen over time.

I(we) spent $1795.00 to get this repaired and the data recovered. Since then I learned all about backups, and how $80.00 today will indeed save me $2 grand sometime in the future. Perhaps. Maybe.

The other drive, the same model, developed the same failure. This time 2 years after usage. It lasted longer because it was run in a pc-case and had some cooling and free air circulation. It was a sudden failure, no opportunity to get anything back. My only option was to restore the data from a backup!

And my laptop drive, well now, it simply wore out after 4 years, it was(is) a Fujitsu MHU2100AT laptop drive. It had billions of head reload cycles and trillions upon trillions of file operations. I'm sure!! I probably loaded Orbiter over 800 times on this drive. Or something.
This drive gave me plenty of warning, starting with a few bad sectors on a scan. I fixed those with HDD Regenerator or SpinRite or the mfg's low-level tool. Or something. I also immediately ordered a replacement and began saving my important work on another disk. The drive lasted a few more weeks, enough for me to get the replacement installed.
It being a laptop drive, I'm sure it got banged around and powered off incorrectly from time to time.

When the replacement drive arrived, a shiny spiffy-new Samsung Spinpoint M5 HM160HC, I swapped the drive out. That took me 15 minutes. Then I restored from last month's backup. That took 30 minutes. Then I put back my working user-dataset. That took another 30 minutes.

This "disaster" cost me about $140 and 2-hours time to FULLY and COMPLETELY recover from. $50 for the new drive and ~$90 or so for the backup drive.

I've had 2 other Western Digital drives go bad on me as well. These were the external Passport series, one was a 320GB and the other a 500GB. Both failed gracefully allowing me to get what was on them off them. Or make one last backup. It was a while ago, and I don't remember. But I did, 4sure, have backups. They were external, subject to power failure, or bounced around a little, I'm certain.
The replacement drives for those two are re-certified and seem to be working just fine, with the 500GB unit getting tossed around in lappy bag.

I use WD exclusively because I am familiar with how to repair them (nowadays) and how to milk their warranty department for all it's worth. Free upgrades and such when you complain. Works everytime.

The one most common feature of the drives that failed on me was that they were new and the first-run from the factory for their capacity type. So my recommendation is simple, get last year's capacity at a good price today and you significantly reduce the chance of failure. Today that would be 500GB 2.5 drives or 750GB/1TB 3.5 drives.

I just got a 1TB 2.5 drive, It's risque purchase but we shall see.. And I can't wait to get that old ST-506 style 10Megger going soon. I wanna make it an even 30 years since last power-up.

Also, not to fault WesternDigital or anything, I have some of their older Caviar drives in the 210MB and 340MB and 1.6GB capacity from the 90's. They still work today.

AND the best use for your old drives, is to re-purpose them into backup units. You know their history, they have been personally "tested" by you for years, you know they work!

 

[Fabri91]

I personally have an external 1TB drive from WD (I think) which gets powered up about once a month for file syncronisation and normally isn't in use.

 

[Keatah]

You have mistaken the terms virus and malicious software. Virus isn't a standalone program, and its design isn't to destroy data or annoy user, but to replicate itself on the system and between systems, by attaching its code to executable part of the host file (hence not standalone).

Well yes that is true. I used the term in the most general sense. Most folks complain to me they have a virus!!!!! I get all these popups!!! And so on.

Even more general is that "virus" means any piece of unwanted software, whether it be a key logger or packet sniffer or pop-up ad. Or intrusive application that keeps nagging you for payment and upgrade.

@Fabri91 - There you go! Sounds simple and effective.

@GaryW - Sounds like an effective setup to me. I think most folks default to that criteria as soon as you tell them to save the important irreplaceable stuff. Or stuff that would be difficult and time-consuming to recreate. And taking your stuff offsite is indeed BULLETPROOF!!

 

[Krys]

I would not say it is bulletproof. Not even close. Backups are important, and having more than one, and not all in the same place is a good idea. But bad things do happen, and its possible you lose your entire home network and the offsite backup didn't work!

Okay that would be one hell of a day, but it is possible.

 

[garyw]

I'd rather be paying a few pounds a month to Amazon for backups rather than a few thousand pounds to a data recovery firm for recovering my data and yet time and again I see people with the 'it won't happen to me' attitude. Yes, it will. One day, sooner or later your hard drive will die. It will be eaten by a power failure, power surge, fire, flood, theft or the sticky fingers of a five year old.

At that point I'd rather have the inconvienience of rebuilding the machine and a download from Amazon rather than the huge recovery bill from OnTrack. There is also no guarentee that a specialist data recovery firm can recover all the data. They can also take a couple of months to process the disk whereas with an off site backup I can have my data back as soon as I download it.

One final point to remember - no specialist data recovery firm can recover data from a stolen disk.

 

[Keatah]

@Krys -- I suppose that could happen. And that would suck beyond belief if it was a mission-critical system. I hear of folks that do 2 or 3 offsite backups AND have purchased separate computers to test the off-site restores.

@Gary -- I gotta disagree with you there. Some company someplace, somewhere, sometime is a-gonna make that claim! I bet they'll make it a spinoff of the remote-wipe some cellular carriers offer. Some experimental drives call the cellular network to see if a wipeout command is in order. If so, they execute a wipe. And only the drive mfg can restore the data. It basically kills the servo marks. That same cellular receiver/transmitter could retrieve data off the drive. This is all laboratory stuff. But for today, you are correct.

Anyways..

@ALL -- I pulled out the exam report for the 500GB drive that failed, and it may be of interest to you. It failed because one of the heads was not perfectly symmetrical. The head would not fly 100% parallel to the surface of the disk.

The outer part of the head, the part of the head closest to the rim of the disk was lower or shorter than the inner part nearer the spindle motor. So as the actuator arm moved from outer tracks to inner tracks, the head would tend to "dig down in" and move closer to the platter. That was fine, sort of. It never crashed, it always still floated. It had no problem reading data (or servo marks) as it traversed the surface.

A hard drive generally won't read user data while the heads are moving from track to track, you see, just embedded servo positioning data. And that data is often located on a dedicated platter! Yes, one whole surface is made for positioning purposes.

Now, on the movements taking the head from the spindle to the outer tracks it was a different story. The head would rise up in height then bounce back on the air cushion. The shorter outer edge of the head acted like an air-scoop and pressure built up and caused the head to fly a little higher than normal. This weakened the signal till the head settled down.

Eventually slight normal mechanical wear caused the system to compensate no further. Running the drive cold seemed to help for a while. I suspect the difference in air viscosity helped there.

It is said that hair particles hitting the head/platter interface is like driving over bowling balls in your Corvette. Smoke particles are just as bad. We're just downsizing to baseballs now!

And, did you know, that the newest hard disks have a bi-metallic strip that is connected to a hinge assembly on the head? This bi-metallic strip is powered and acts like a peltier cooler or heater, depending on current flow. And this strip, through the hinge assembly, controls the angle-of-attack of each head. This enables the head to fly higher or lower as needed. Now it's like driving your sports car over golf balls!

Not only that, the main head arm is moved by a voice-coil servo, most of us that have seen diagrams of the HD assembly know what it looks like. What you may not know is the tip of the head has a piezo element as a fine positioner. It basically wiggles the head/tip left and right. For fine positioning of the tracks. Similar to the magnetic coils in a cd laser pickup, but much faster and more precise. Now we keep your sportscar precisely positioned in its lane to about 2cm accuracy, scale.

The next generation hard disks coming out end of next year will have laser assisted recording. They use a laser to warm up the metal, I don't know how hot yet, but I would suspect close to the curie point? The disc medium will then accept a magnetic charge faster. The laser will also serve duty to image how high the head flies. This will be a direct measurement in uM, instead of relying on signal strength to estimate head altitude. AND NOW, your sportscar is riding so low to the ground that it cannot handle anything more than children's marbles without suspension failure..

If you don't know anything about sportscars, you might try the 747 analogy. It is like flying at full speed above a forest and taking pictures of each leaf and determining their orientation in 3-d space. And you do that before the leaves zip away from the camera view!

So that little circuit board on the back of your bare drive has a tall order! Of superhero responsibilities! It has to interpret commands from your o/s, it has to spin the platters precisely, it has to measure the speed of those platters. And it has to control each of perhaps 4,6,8 heads to within uM tolerances via 5 mechanisms.

1-Coarse linear speed via spindle motor rpm
2-Coarse lateral positioning via voicecoil magnet
3-Fine lateral positioning via piezo crystal
4-Signal strength via altitude, bi-metallic strip and pre-amplifier.
5-Fine linear positioning via PLL sythesizer, adjust the frequency of read/write to match rpm.

#3,4,5 are independent for each head and compensate for minor errors of 1 and 2.
#1 and 2 are common to all heads - Until they come out with a drive that has a separate motor for each platter (not gonna happen)!!

Now that it has stuff setup, it has to look at frequencies in the MHz range and count data bits (of varying strength) accurately. Usually these pulses are uV or possibly nV range. The pulses are so weak they are incapable of traversing the entire length of wire from the head to the control board. They make an intermediate stop along the way to a pre-amplifier and booster.

The ones in the binary string "1111" are much stronger than the ones in "1010". The drive has lookup patterns dyamically calculated each revolution (or so) of the platter, depicting the strengths of the data bits. It also knows what the sine-waves look like under varying temperatures and vibration.

If there is a mistake, and each and every drive has tons (millions of read errors every single day!) the 56-bit CRC is used to correct the errors. Drive mfg's have gone to dual and quad core processors (not like intel Core i5 or i7, but specific dedicated CRC checkers) just to process the errors quickly enough. There is no such thing as a perfect error-free drive! Sometimes the CRC is updated, sometimes the data is re-written, sometimes it is relocated to a different track/platter! All without the o/s knowing about it. Did you know that about 45% of a drive's capacity is related to servo positioning data AND error correction checksums, AND data 'marks' for the structure of the platters? Then the drive sometimes relocates bad tracks and remembers what areas are good and bad. So it has plenty of "hidden" spare space for future use. And this defect table is stored either on the HPA protected area on the drive platters or flash/firmware on the controller. This is why amateur data recovery operations fail, because the controller board *might* have outdated info, and DIY recovery ops often include swapping the board without consideration of the parameters stored there.

Sometimes the unused spare space hides high error rates. When that spare space is all used up, you have a drive that now develops bad sectors for no apparent reason. Insidious & ingenious, ain't it??

And to go even further, all hard disks are analog in basic nature. The data the controller circuitry and pre-amp sees is a continuous sine wave of various amplitudes. The frequency is mostly constant though, thank god!! So that controller board is basically one big-ass A/D D/A converter!

When you turn on your system, way before the BIOS completes POST, your drive is spinning up, and the computer inside your hard disk is already booted and running. It takes just a few seconds to gather all that information and go through the calibration routine. That's the head banging you hear on power up. Your entire hard disk is a whole separate computing eco-system. Very much a separate entity from your main computer. And in the early days, it was far more powerful than the system which hosted it! My old Apple ][+ had a 1MHz 6502, and the 10Megger Sider drive had a 4MHz Z-80 cpu to control it!

My point being is with all this complexity going on, you can have a failure from time to time!