Sponsor for PC Pals Forum

Author Topic: Data recovery: Survival Guide (Part 2)  (Read 7812 times)

Offline Clive

  • Administrator
  • *****
  • Posts: 68783
  • Winner BBC Quiz of the Year 2015,2016,2017, 2020
Data recovery: Survival Guide (Part 2)
« on: March 11, 2004, 13:39 »
PC Resurrection

A dead PC doesn't necessarily mean a dead hard disk, nor does it automatically mean your data will be damaged. I've seen hard disks brought back to life by nothing more magical than an open case and a big fan to cool them down, yet a dead PC or one that crashes every few minutes does present something of a barrier to getting at your data.

Knowing how to breathe life into the machine is, therefore, an essential skill of the DIY data survivalist. When you turn your PC on, the BIOS will perform a POST (power-on self test), which is a diagnostic to check that hardware is running okay before the actual boot procedure can occur. A POST consists of the following stages: power supply on and releasing a reset signal, CPU exiting reset status mode, BIOS readable, BIOS checksum valid, CMOS access, CMOS checksum valid, CPU able to read memory, first 64KB of memory operational, I/O bus/controller accessible and I/O bus functioning with regard to the video subsystem.  

If any of these tests fail, a series of beep codes will identify the fault via the PC speaker. The beep codes will vary from BIOS to BIOS, but for a comprehensive listing of codes and their meanings visit www.computerhope.com/beep.htm or check your motherboard manual for specifics. Either way, it's useful to have a printout of them available for emergency diagnostic use. More often than not, a dead computer hasn't 'joined the choir invisible', but instead is just resting.

There are several degrees of death as far as computers are concerned, from not a flicker through to dying after boot up. The trick is in being able to approach the problem methodically and discounting the obvious. It could take a couple of minutes to fix yourself before calling on that next-business-day, on-site engineer. If you've only got a return-to-base warranty, you could be without your PC for a week or more, and if it turns out to have simply been a loose DIMM, then you'll have wasted time and possibly money.

One of the simplest ways to resurrect a dead PC can be to take the RAM out and put it back in again. Don't be tempted to adopt the DIY approach on component replacements, though, unless you're prepared to void whatever warranty remains. Indeed, you should check the terms of your warranty, as some of the more Draconian service contracts still forbid the opening of the system case - this can usually be spotted by the presence of service seals on one or more edges. Also, before you start fiddling around inside the box, ensure you're properly grounded by using one of those static-busting wrist straps, or at the very least by discharging yourself by touching a grounded object such as a radiator first.

Once you've done that, look around and make sure all the cables, plugs, chips and cards are seated correctly and haven't become loose. This can be a particular problem with memory and expansion cards - both are easy to fix with a little firm attention, but left unchecked will wreak havoc.

Warranty issues aside, the first thing you should do if the computer refuses to boot is check the power supply. By this, I mean everything from the wall socket (check the fuse hasn't blown in the plug) through to cabling (check they haven't become loose) and any surge-protection devices or UPS units.

Surge protectors may need to be reset or replaced, depending on spec or whether they've already had to save your PC from being fried. A dead power supply is a surprisingly frequent culprit, and for all kinds of reasons. The most basic reason is that it has burnt out, and this is fairly obvious from the electrical burning smell that will hit you as soon as you open the case. Bolting in a new PSU is quite easy enough, although this isn't always necessary to do.

Check that the fuse inside the PSU hasn't blown due to overloading or a shorting caused by a loose plug, for example. If you've recently added a new peripheral/component that draws power from the PC, check if this has presented too high a load and whether any overload protection within the PSU has kicked in, preventing it from switching on and doing damage. The easiest way to check for overloading is to unplug devices one after the other and see if the PC powers up again. If you've managed to fry a PSU, instead of replacing it with the same unit you might want to uprate it from a 300W unit to a more powerful 400W one that can deal with all those hard disks and the exotic cooling system you've fitted.

If the PC shuts down without warning - usually in the middle of some processor-intensive operation or on a very hot day (or a combination of both) - you could be looking at another common failure: CPU overheating. Open the back of the system case and let things cool down for a while, then try rebooting and check if the CPU fan is working. If it has failed, a replacement can be a simple or complex task, depending on the PC in question. I recently had a cooling fan fail on a Dell box and the replacement was fitted within five minutes, thanks to it being a tool-free design.

If you urgently need to get at some data, but can't have a replacement fan delivered for a day or two, don't panic. I've kept a fanless system running for plenty long enough to copy data to removable media or across the network by opening the case and pointing a large desktop fan running at full pelt in the general direction of the CPU. I wouldn't advise running a system like this all day, but for half an hour or so it can be a lifesaver.

If the fan is okay, check the heatsink for dust. I've seen quite a few CPUs get hot under the collar due to a clogged-up heatsink. Another cause of system shutdown like this is a dodgy driver, especially if you've recently installed new hardware. Problems with drivers can also be the cause of unexpected crashes, so make sure you keep up to date with the latest drivers by regularly checking manufacturer websites. If you do get a blue screen of death and suspect a dodgy driver, boot Windows using Safe Mode (press F8 during the initial boot process and select Safe Mode from the Options menu), as this uses a minimal driver set, won't access autoexec.bat or config.sys and happily ignores much of the Registry. Using Safe Mode, you can back up important data before continuing to hunt down the cause of the instability, so helping to prevent a data disaster.

No excuses

Of course, the best defence against losing data forever is to implement a proper backup strategy, be that part of a business continuance plan for a larger organisation or just a matter of dragging your important files onto a CD-R each night for the home user. Sure, there's an element of hassle involved, but you can automate the process using fairly inexpensive tools.

A common excuse for not backing up is that 'I didn't have the time', which has to be the lamest thing a serious computer user can utter. My backup routine here is ongoing throughout the day, with 12Ghosts Backup storing changes to my documents on the fly. It only adds 15 minutes or so to the shutdown routine at the end of the working day, as an incremental backup copies only those data files that have changed since the day before onto a Maxtor 5000XT 250GB external USB 2 hard disk and then again to another identical drive that gets removed from the premises each night. Once a week, I put Drive Image to work and make an exact copy of my system drives just in case.

So yes, it takes time and discipline to have safe hex as it were, but those extra minutes each day pale into insignificance compared with the hours or days wasted as well as the business and hair lost if your data heads south without you.

The methods outlined in this survival guide aren't meant to replace common-sense computing practices; they're strictly to be filed under last-chance saloon.

Extreme measures

If your computer suffers an extreme disaster, there are some important steps to take to give the data-recovery specialists the best chance of saving the data and the day.  Fire can cause many problems, but system melt isn't the biggest of them - water is. Fire hoses can waterlog a PC, but never attempt to dry the hard disk. Remove it if you can and put it in an airtight plastic bag as soon as possible; if the drive has melted and so sealed itself, all the better. The water that gets sucked in the drive when the firemen are doing their stuff will almost certainly be contaminated by dust, ash and assorted debris that will be left behind on the platters as the water evaporates.   Try running the drive and it will be toasted a second time. Similarly, an electrical surge such as a lightning strike or power outage that's serious enough to crash the computer can kill your drive, in which case remove it and send it away for surgery as is. If the drive is still spinning, but you can't get at any data, the rule of thumb is to switch off all power and unplug all cables and leave them like this for at least an hour after the power has returned. This gives the electrical supply time to stabilise before returning power to your PC. Of course, a decent UPS with a surge suppressor would have prevented this particular extreme form happening.

Sudden noises are among the worse indicators of damage, as they're usually a result of the drive heads crashing into the spinning platters. Turn the PC off and don't attempt to reboot.  


Offline Tony

  • Loyal Member
  • *****
  • Posts: 3367
    • http://www.sugrue.ndo.co.uk
Re: Data recovery: Survival Guide (Part 2)
« Reply #1 on: July 03, 2008, 17:53 »
Yo all the Para's are over here  ;D
Athiesm is a non-prophet organization.

Show unread posts since last visit.
Sponsor for PC Pals Forum