Image Backup under Windoze vs Stand-Alone

Hi, have purchased TI9 as a replacement for a combination of Bootit NG and BackupMyPC. Main reason for trying TI9 was the integration of image backup with differential/incremental. Despite poking around here and reading the manual, I'm confused over whether it is better to do an image backup from within WinXP or via the stand-alone CD boot.
I've always used stand-alone DOS based image backups with Bootit, not being comfortable with the add-on lock that they use within XP to freeze the image.
Any advice around on the pros and cons of WinXP or stand-alone originated images please ?
Also, can I combine an image produced via the stand-alone boot with incrementals produced under WinXP ?

I realise this is fairly basic stuff, so please bear with me...

 

There should NOT be any difference if you believe the software developer. However, there is NO WAY these folks could test for stability with your PARTICULAR RIG. Therefore, for best performance, I only image the partition when it is NOT in use.

Many non-windows imaging proggies are under 3MB in size (Bootit). There are a few Windows-based applications (Drive Snapshot/IFW) with installation footprint under a few MBs. I personally would avoid those windows-based imaging programs over 25MB. The logic being that there are many patches within the application to address a buggy core program.

I like to KISS. My OS partition is only 900MB. It takes 40 seconds to image this partition to another partition on the same HDD. All personal data including MY DOCUMENTS are moved to an extended logical partition which I image once a month. I've never had to restore this partition except when I change HDD or replace the MB/CPU.

 

You will find some users who think it is better to use the boot CD version rather than trusting the Windows snapshot technology. Sure it is probably safer but there are very few problems posted where the Windows backup didn't get it right. I always backup from within Windows.

When doing an incremental/differential the file format is the same so there shouldn't be any concern about which version created or modified the archive. One thing you should ensure you do is make a new Rescue (boot) CD with each new build you install so the fixes and features that were released are now in your CD version.

FYI, the boot CD Full version runs in a memory-resident Linux environment and contains USB support. The True Image (TI) safe mode is a DOS environment using BIOS routines and on many motherboard's this means no USB support.

TI does not do real "updates". Each new build is a complete full version of the program. Although it shouldn't be necessary there has been trouble with various builds being installed on top of the previous build so it seems it is a good idea to remove the old before installing the new.

Be sure you register with TI because unless you do you will not be able to get the latest builds and it is another way of finding your serial number if you lose it!

 

Files are almost constantly changing when you do a backup from within Windows.

You can compare the result of a backup with the actual drives using the following:

http://www.standards.com/index.html?CompareDrives

http://www.standards.com/index.html?GetFileTypeDistribution

 

Never done this before.
I'm gonna do a full backup from the Recovery CD.
Then do a full backup as soon as the computer reboots.
May prove interesting to see what differs between the two backups.

Hmmm, the devil made me do it!

 

I'm sure there will be changes to the 2nd image file.

 

I expected that.

 

OK, I rebooted and ran the backup from the recovery CD.
I was very disappointed.

The backup, including verify, took aabout 5 hours and 9 mintes.
Way to long.

When I rebooted and ran the backup, and verify, from windows, took about 2:53:36, more than two hours less than backing up from the recovery CD, and still more than an hour longer than a backup using Ghost 10,

In addition, the creation/modification times seen to be off by about 4 hours.
Creation time is listed as 14:31, but I started the backup at about 18:31.
The Modification time is listed as 17:30, but it was actually about 21:30, and the verify finished at about 23:39.

I expect others have noticed this time bug.
I'm using Win 2000 with build 3633.
I have no reason to install a later build without knowing what bugs were fixed in the later builds, or hard evidence of improved performance.

 

Wow Howard, those timings look pretty scary...
I'll install TI9 in the next couple of days and post results of my own experiences.
It would be helpful to understand something about the backup scenario you timed - size, destination etc.

I timed the backup of my WinXP partition (15.6 GB with 2.77 GB free) using Bootit NG a few days ago. Backup to a second IDE disk together with a bit-by-bit verify took 29 minutes.

 

I am comparing for a given software/hardware configuration.
It's the relative performace that matters for a given system.

Only difference, but I do not believe that it affects results by more than a few minutes, was the destination drive. When backing up from the recovery CD, destination drive has a smaller buffer than the drive used when backing up in Windows. Both drives are the same speed.

Using build 2337, I recall that TI took only about 10 minutes more than GHost 10 for a full bacup.Full backup times have deteriorated in newer builds.
But, incremental backups have significantly improved. So TI has made some changes that affected speed both positively and negatively.

Cannot compare differential backup as, until a week or so ago, I only used incremental backup. I now use incremental with two destinations and differential for two others.

 

Actually the program is working as designed. However, as you can see from this previous thread titled <Bug in True Image Rescue Mode?>, it does lead to confusion.

 

It 's nice to see several tests with various proggies using the same setup. However, it would be better if you provided the AVERAGE transfer speed from location A to location B, and the original size of the SOURCE partition (with/without pagefile, etc).

A modern PC A64/Sempron (>1.8GH) or P4 (>3GHz) should be able to process the image file at a minimum rate of 1100MB/min. The average image restoration speed is approximately 1300MB/min (partition A to partition B transfer on the same 7200 rpm HDD). The AVERAGE maximum imaging speed for a PATA or SATA 7200 rpm drive is 1500MB/min.

Your particular build of TI is optimized to run in windows (the latest Home version has a download footprint of 75MB). Not much attention to detail is given to the program that run from the recovery disc. Theoretically, an application that's properly coded to run outside of windows should perform better because it can fully utilize all of the PC's resources.

TI has the capability to run outside of windows since version 6 (could be earlier). It's too bad the software developers are still strugging with the speed issue. <snip>

edited to remove off-topic bash - please try to remember that this is the Acronis support forum.

 

Off-topic post removed - other software may (as always) be discussed in the software & services forum section.

 

Yes, I had guessed that it was UTC, as in boot mode, there is no knowledge of the time zone settings. However, there needs to be a way to reset the time.

I can do so with a simple program, but most users are not programmers.
Perhaps, I'll post such a program.

 

I am not interested in, nor was I, comparing speed amongst systems.
I am just comparing the time needed for a full backup:

1. TI in boot mode.
2. TI in windoze mode.
3. Ghost
4. Relative to earlier builds.

If all TI is doing is checking the file structures and copying sectors after compression. then I have to wonder why there would be such a speed difference between boot mode and windoze mode.

 

Here are the results of some rather unscientific tests on various TI9 backups. All with build 3,677 and validated.
15.6 GB XP partition with 2.77 GB free from one IDE drive to another. 2.53 GHz P4 with 1024 MB memory.

TI9 Standalone, normal compression:
15 min 8.15 GB
TI9 Standalone high compression:
25 min 7.2 GB
TI9 via windows normal compression:
26 min 8.15 GB
TI9 via windows high compression:
28 min 7.20 GB
TI9 via windows maximum compression:
50 min 7.02 GB
Bootit NG standalone default compression:
29 min 7.84 GB

I realise there are various things affecting these results, not least on the validation side where (apparently) TI9 only reads checksums, whereas Bootit NG does a bit by bit verify of the whole image. Anyhow, TI9 seems to work for me - speed is good in standalone and XP mode and the user interface is good.

 

I got the program working, just need to decide on the interface.

In the process of writing the program, I realized that TI need not use UTC.
The time zone info is available just by examining the file dates and times internally. Can Linux vet those dates?