E00070020 The archive is corrupted in ATI 11 / Vista on a new PC - HELP!!!

Hi all,
I have recently purchased Acronis True Image 11 so I can backup various partitions and data on my new PC but I can't get it to work. Now I regret not taking advantage of the 'Trial Period' before I bought ... I just assumed it would do what it advertises!

Each time I manually run the backup process the validation step comes back with an error 'E00070020 The archive is corrupt'. I have had a bit of a scout around (via Google and this forum) and there seems to be an outstanding problem with ATI. Some refer to hardware issues, some are adamant that it is a software bug.

I am attempting to backup my primary system partition so I can store a baseline Vista image. I have attempted to save it in both the Secure Area and in a regular Backup location, both of which reside on a separate hard drive to the regular 'operational' drives (which run in RAID0). Backups to both destinations fail, and to date I still have not been able to get a backup to work successfully.

The details of my configuration are:
- Windows Vista Home Premium (32bit)
- Acronis TrueImage 11 Build 8.101
- 2 x 1TB Western Digital SATA HDDs (WD1001FALS) running in RAID0
- 1 x 1TB Western Digital SATA HDD (WD1001FALS) standalone (that I want to use for backups etc)
- 4 Gb Corsair Dominator PC-8500 DDR2 RAM
- ASUS Rampage Formula Motherboard
- Intel Q9450 Quad Core CPU

I have sent Acronis support an email but thought you might be able to enlighten me on the intricacies of Acronis TrueImage 11 (as I am a first time user).

Q: Is there a process of diagnosing / resolving E000670020 errors?
Q: Has anyone had similar issues ... if so how was it resolved?
Q: Are there other related issues that I need to navigate through being a
first time user running under Vista? (I saw some messages about not being able to boot from the restored image)?

I apologise if this has been covered all before ... but lack of sleep and information overload has kicked in from trying to resolve this. Any tips from the 'pros' would be greatly appreciated, otherwise I will have to go down another path for backup software.

In the meantime I'll run 'memtest86' to see if it is memory related ...

Thanks in advance,
Tom.

 

Hello,

I assume you start the win-version of Ti.

Did you try the backup with starting from the rescue-cd?
If not try it.

I hope, the rescue-cd starts well and recognize your Raid...

Please post the result here.

BTW: Did memtest86 give a result in the meantime?

 

I tried from the boot CD and from the 'F11' boot option. Both failed with the same message at the same point (validation).

Are there any issues with backing up the system partition that you can't do it from windows (i.e. does windows lock system files that can't be backed-up??)

I haven't run memtest86 yet but will do. I am travelling away for work the next few days ... but will continue to try to resolve the issue upon my return. Just hoping for any more ideas in the meantime so I can have a whole series of tests I can run in trying to identify the cause of the E00070020 errors.

Thanks for the reply. I'll keep you posted.

Tom.

 

What kind of disk is your target?

 

The target is another internal WD1001FALS hard drive, not in a RAID configuration.

 

Try,
Set compression to NONE
hard drive write speed to about 60% on the scale.
Set validate image to on
Exclude pagefile in the image with C:

those are in make an image options.

Never fails for me. That way the drive has "more time" to write the image with less chances of a corrupt image.

 

YOu don'lt need to exclude the pagefile. ATI automatically excludes it except for a necessary placeholder since Windows has to rebuild a pagefile after you restore.

 

Aronis support requested that I install a 'SnapAPI' driver, which I did but the problem still remained.

In addition they have also requested that I run MemTest86. Which I am currently in the process of doing. It has been running for over 7 hours now and still not finished. I am not sure how long this test is supposed to take ... seems like a long time to me.

I have attached the photo of the screen which shows some errors ... I am not sure what they mean ... looks like a little more research is called for!

I'll give the backup a go with the settings you recommend ... I have nothing to lose.

Tom.

 

memtest86 is a never ending program .
Thats ok. memtest86 has run a night. When there are no errors then you can be to 99% sure that your memory is ok.

But you have RAM-errors. That is not good.

Possibly you have to increase the VDimm a little bit.

But first: Please tell us your hardware concerning motherboard (exactly included the revision, please) and your RAM.

If you enter your BIOS, can you tell me how the VDimm value currently is?

EDIT:
I've seen that memtest86 V1.70 cannot recognize your chipset. Please use the latest memtest86 (V2.01): Click

 

According to the screen you have run 10 full passes and you indeed have RAM errors which can cause archives to be declared corrupt.

While getting the latest version is a good idea, the screen indicates you have a P3 system so V1.70 would have no trouble supporting the older chipset.

You can check your motherboard manual about the possibilty of bumping up the voltage on the memory which can help but the odds are you need some replacement RAM.

 

OK ... I have downloaded the new version of memtest86 ... it now successfully identifies the CPU properly (as a Core 2 family) and the chipset (X4. I have fired that off so wil wait a few hours to see if there is any difference in the outcome.

The motherboad is an Asus Rampage Formula. not sure of the revision ... I'll find that out ... (I'll have to wait until memtest finishes as I have to unplug my PC to get it out from my desk before I can open it up to have a look!) ... unless I can find out from the BIOS ... but I couldn't see anything.

Regarding the memory it is Corsair Dominator TWIN2X4096-8500C5DF G

The only thing I have changed in my bios from the default is the DRAM Frequency from Auto to DDR2-1066MHz. Nothing has been overclocked (yet).

I could not see any direct reference to VDIMM settings ... Is this the same as the DRAM Voltage? If so it is set to 2.16V.

Sorry for being such a noob but it is all a learning experience!

 

He is the results of memtest86 v2.01 after 3 cycles ...

No errors so far.

I'll let it run for 8-10 cycles just to be sure.

 

Sorry, my mistake, I didn't realize that the P3 designation was in error.

It is best to let Memtest run overnight for good measure. This also ensures that the PC is sure to get to its normal operating temperature for most of the test.

 

Hi,
I saw the revision is not neccessary because there is only one manual for all.

Reviewing the manual I see you are right with the assumtion that VDimm is equal to DRAM VOLTAGE in this case.
If the value is "AUTO" please change it to MANUAL. Then set the voltage to 2.25 volts.
Don't change nothing more. Save the bios changes and restart the PC.

According to the data sheet of the RAMs they runs with 2.2 volts.
With 2.25 you are on the right side.

BTW:
Your suggestion to run memtest86 a little bit longer is correct.
8 to 10 hours are ok (I mentioned it before as 'a night' )
I agree with seekforever in all points.

 

Here was the final result of the memtest ... still had some errors.

Not sure what this has to do with getting ATI to work though ?!?!?!

I have upped the DRAM Voltage to 2.24. I will run memtest again overnight.

 

The hypothesis is that ATI uses so much of your ram that it uses parts most other programs do not and therefore, it can run into memory probs that other programs do not. If ATI has to rely on maprts of memory that aren't working correctly, then that's bound to mess up the file read/writes.


Also, even though it's easier for software folks to blame the hardware and vice versa, sometimes, they're actually right.

 

TI stores a checksum for every 256K bytes of data. When the archive is validated it recalculates the checksums and compares them with the values stored in the archive when the archive was created. If only 1 checksum doesn't agree the archive is declared corrupt.

Lets assume the archive was created properly with all data stored correctly. If the archive is read into a bad memory location and the checksum algorithm applied the checksum will not be correct since the data stored in some RAM location(s) is not what it should be. The different checksum causes TI to reject the archive as corrupt.

Why doesn't this happen with regular PC use? TI is comparing 4000 checksums per gigabyte of data when it is in memory - other applications don't do that or anything like it. In fact, most apps rely on the CRC checksum handled by the disk controller only and assume data placed in memory is correct. In other words, if the disk subsystem is happy all is well. TI is dealing with gigabytes of data most apps deal with a few megabytes at most so the chances of using the bad locations are smaller. Also, depending on the type of data, the effect of a bad RAM location may not even be noticed in normal use. Regular PCs do not use parity or error-correction for RAM data and the only time a PC knows something is wrong is when the bad content causes the program to run amok or there is an obvious data problem such as a TI validation error.

I believe the Memtest screen showed you have a dual-channel motherboard. You may be able to pull out 1 stick and have the machine run as single channel. This way you can see if your validation problems go away when one stick is removed. I had problems with a dual-channel board that Memtest only showed up rare errors. By running it in single channel mode the errors became much more obvious.

 

Yes, do that.
How you can see not all test-passes failed at the same address. That seems the RAM is not defect.

The 2.24V must correct this.
Retest it in the night...

When the errors appears again please follow seekforever's suggestion: Remove the RAMs with the exception of one and test again.

 

Not necessarily. To take the all the read/writes memtest gives it and only fail 6 times in 13 passes indicates the memory has a very marginal, not a hard fault. If I remember correctly test 6 is a random data test rather than a fixed pattern or bitwalk to all locations. It has been my experience that marginal errors tend to show up in random data tests rather than the fixed data tests.

 

Well ... based on all of the above. I ran the tests overnight at 2.20v. Still had errors. I then upped the voltage to 2.24 and after 3 cycles no errors were present.

I did some research on Corsair's site and the memory should operate at 1066Mhz, 5-5-5-15 at 2.1v. Mine doesn't without errors by the sounds of it ... so I decided that I would gather all my 'evidence' and go back down to the PC shop and get a replacement for my memory (which they did .... very happy).

I have just replaced the memory, reset by BIOS settings to 5-5-5-15 1066 and 2.1v and run the backup / validation process. IT WORKED!!! I nearly fell off my chair!

So to be sure I have since run another backup (which passed validation again!), and will run MemTest overnight to see what the result is.

If I get the all clear I'll test out the restore functionality of ATI to make sure it is all going to work OK. I'll keep you posted.

 

Well ... as an update...

First off ... thanks for the replies and assistance. Great to see such a supportive user group. BTW I still haven't got responses from Acronis Support ... lucky you guys came to the rescue!!!

Secondly ... I have sucessfully restored an image so I am happy with how ATI is working for me so far.

Third ... the results of the MemTest on the new RAM still produced a couple of errors ... bummer! Do you guys often get errors or does it come up clean even after several test cycles It seemed that the errors only appeared in the 6th pass (out of 10 passes). Acronis seems to be happy ... but it still makes me nervous that I still have a couple of errors appearing. Not sure if this would be considered 'normal' behaviour.


Thanks again for the help and advice.

 

All of my PCs will run Memtest overnight with zero errors and this is the only way it should be - assuming Memtest properly supports your motherboard.

My guess is that there is a problem with your new memory because of the similarity of the failing address at the 2 locations and the fact that the same data bit is in error both times. Also, the failing addresses and data bits are different than with your old bad memory which slightly supports a hardware rather than memtest support problem.

Of course, there is more to the memory system than just the sticks but the odds are that hardware problems are with the sticks.

You can set Memtest to only run test 6 which is giving the errors and see how many errors it produces. The other thing to do is to go back to the PC store and talk to them. They usually have a memory testing program that they run on new systems to check them out.

 

@tlutton
Here is the compatibility list of your MB: Click

Corsair is qualified with the RAM-Part-No "CM2X1024-8500C5".

My opinion is that you can overvolt your RAM up to 2.4 Volts...

After a little investigation I found no "CM2X1024-8500C5" at Corsair.
There are only "CM2X1024-8500C5D" used as chips.

And it is amazing: The RAM-SPD is programmed as "JEDEC standard 5-5-5-18 values at 800MHz"
Corsair tests the modules up to 1066MHz and when they are 'good' Corsair sells them as 1066MHz modules

Here a datasheet: http://www.corsair.com/_datasheets/TWIN2X2048-8500C5DF.pdf

And here the DOMINATOR overview: http://www.corsair.com/products/dominator.aspx

Very suspect...

 

Thanks guys ... It looks like another trip down to the computer shop.

Might see if I can exchange it for another brand ... any suggestions?

 

I don't know what RAMs your dealer has.

The best would be:
Meet your dealer, build in new RAM and start memtest86.
Then you have to waste time with strolling
After 4 hours you look for errors...

Have a nice day

But now serious:
Crucial, cellshock, mushkin, G.Skill