Shellshocked:pri & bu both crashed

[jimhend]

W2K sp4, ASUS MB 1GB RAM 2 maxtors, 80GB primary C:, 120GB for images D:. TI8 used for incremental imaging. I had >30 incremental files collected over time. System was working fine, I decided to delete multi-images and create 1 new image as not enough capacity on D:. Deleted old tib files, created sucessfully a new backup image. Shutdown for the night. On very next reboot I got a BSOD on C:, some kind of software checksum error, I recall (5 days ago). I think, "WOW, am I glad I use trueimage!". I boot with the boot cd and start the recovery process. About 3/4 through the recovery, I get "IMAGE IS CORRUPT", recovery halts. I bought a new drive and the corrupted image initially mounted and I tried to drag out some files, but it ultimately failed on bad sectors. I am now in the process of finding file recover software and will try to get back some data from the 2 failed drives. Has anyone experienced a failure like this? Takeaway for me is ALWAYS test an image on a new drive/partition, maybe even keep a full clone HDD on the shelf. How effective is the TI validation utility?

[DwnNdrty]

There have been many reports by users here about false positives with the validation. I never do a validation, but I do test my backups by doing a restore to a spare hard drive. I also never do incrementals or differentals - only full backups.

A good file recovery program that is not too expensive is R-Studio. The interface is easy to understand.

[seekforever]

TI expects your hardware to be working properly. You had a BSOD mentioning "some kind of checksum error" which indicates that there is a problem. You then had a TI validation error which is really a checksum error. It could be you have some problems with your system that aren't totally in the HD subsystem. To do a checksum calculation the data must be read correctly into RAM and then the algorithm must calculate the checksum. There are 2 main points of failure, reading the disk and the RAM. Regular PCs assume RAM is always correctly written and read. You can have RAM errors which are never obvious in many operations - who would notice a bad bit in a jpg file? A single bad bit out of many GB to TI's checksum calculation means "Archvie Corrupted".

The TI validation is done by writing a checksum for every 256Kbytes of data in the archive. If you can validate the archive it means that the data being read agrees with the data that was written into the archive.

The biggest problem with validations giving false positives, which I have never experienced BTW, is that people validate the archive using Windows. Why is that so bad? It is because recovering the active partition, typically C, requires that Windows not be running and to do this TI starts up a Linux recovery environment with a Linux version of the program. This is using a different set of drivers and everything else for that matter than Windows. So if you want to add confidence to your validation do it with the Linux recovery CD by booting it up.

As a rule, if you find you can validate a few times with the Linux CD version then you should have no problems restoring an archive that validates under Windows since you have demonstrated that your system is quite happy using the Linux environment. Nontheless, the best way is to do a test restore. BTW, you should always do a test restore on a spare drive after installing TI to make sure it can be done with your hardware setup.

[jonyjoe81]

You might want to try your hard drives on another motherboard. The odds of two hard drives failing at the same time are "0". When I had a problem where I had input/output read errors with a hard drive and couldn't get the computer to start even after I installed a new hard drive, I traced it to a bad motherboard, but to be safe I also replaced the power supply. It's been working fine since, and the original hard drives which I thought were bad,are still working without any problems. It's just my suggestion to check your hard drives in another computer, if your motherboard is bad, repair utilitys won't work correctly in the current system you have.
From my expierence everytime my images have verified good, they have worked when I used them to restore a new hard drive. I have 100% success rate in that area.

[Proactive Services]

Originally Posted by jonyjoe81

You might want to try your hard drives on another motherboard. The odds of two hard drives failing at the same time are "0".

This is untrue. My ISP had a double-disk failure today, as it happens. If a power surge ocurs it can quite easily fry both hard disks.

[DwnNdrty]

A few rules that any computer owner should adhere to are:
Always have the computer connected to a battery backup supply.
Never keep your backup drive connected in the running system. If you do automatic backups and don't want to take the little time it does to reconnect the drive then that's a risk you take.

[GroverH]

The more you go down that road called "experience", the more you will maintain multiple sources or methods or copies of backup data.

An incident (electrical storm/flood/theft/virus) can take out your system and you need backup data stored elsewhere--even off-site.

A good battery backup or power conditioning is better than a surge protector but even it won't stop a lighting strike. Have multiple backup copies in different places.