Introducing AX64 Time Machine - hybrid imaging/snapshot software

Thanks Carfal.
I don't have the time to test thoroughly AX64 and its fast imaging and your result would help me understand how it performs/achieves the fast restore.

I wasn't sure if it used the OS apis to perform the writes or if it used direct access. Your result excludes the direct access.

Panagiotis

 

As I said to my previous post, I wanted to see how AX64 performs the fast restore.
(Related to the question through pm I send you last month about how it performs the fast restore and if in the final step it restores the mft table.)
Until carfal's test I was considering the fast restore to be more close to imaging apps but now it seams more close to ISRs.

Panagiotis

 

No... please leave it optional.

ps. unless you used direct disk access, restoring the mbr would have no effect to the RBRX structure.

Panagiotis

 

Congrats on this great initiative Isso!
Have been using RBRX 8 - 9.1 for day-to-day rescue actions. It normally worked OK but due to a few serious troubles on some restores and the concerns of only being able to save the snapshots on the system disk plus the MBR vulnerability, felt it was just not robust enough. There was always an uncomfortable feeling and I did not like the procedures needed for a defrag. Did not test anymore/yet version 10, due to the continued need of IFW/Phylock as separate imaging tool that would catch all snapshots because their separately sold Drive Cloner is still not ready for that.
So I went from RBRX to Roxio/Sonic BOT, which has served well with manual restore actions of individual files from the System Rollback Data folder (unfortunately also located on the system drive only).
Normal imaging (and restores) I have done a for a long time now with the rock-solid SP.
How beautiful it would be to have a dependable all-in-one solution as offered by AX64! I have been very impressed by the most recent alpha build 1.0.800
Here are some non-critical ideas that would make it even better without adding bloat:
- adjustable automatic backup time
- option to force an auto-backup when running an installer package (grab priority to finalise the backup before the actual installing process starts, or by any other more practical way)
- option to exclude user-selectable folders and/or files before backup (if technically posible)
- right-click option on files & folders in backup browser to open (go with explorer.exe to) their location and delete from in backup browser
- a selectable view in backup browser to show only files changed since:
* last backup
* in a selectable time-frame (between a start- and end time)
Good luck with the further development, it already looks like a serious piece of system security software to me!
By the way, could also not make the direct USB recovery image on a SD-card in a USB Multi Card Reader, possibly due to a driver problem. Via the ISO it was still possible, have not restored yet from it though.

With best regards,
defconnect

 

Mohamed, starfish, thank you for the information, I've put this to our feature list

Regards,
Isso

 

aigle,

Very good question! Yes, any operation that you do from within OS that has malware installed can be intercepted by that malware (if malware is running in kernel mode and is specifically crafted to overcome the protection). No software (not AXTM, not RollbackRX, not even antivirus) can totally protect the OS if it's running along with the malware on the same OS.
That's why if you want to be 100% sure that your machine is cleanly restored - use AXTM offline restore.
In vast majority of the cases a normal online restore would be fine though.

Isso

 

Panagiotis,

Fast restore is exactly the same as for other imaging apps, i.e. the partition is opened for direct access and written. The difference is that actual write operation is done while OS is running. That's why Rollback's driver also receives it.

EDIT: it's AXTM driver that writes to partition from kernel mode, i.e. no API is involved. And Rollback receives this just because it's driver is situated lower that AXTM. If I put AXTM driver lower, then the situation will be vice-versa, i.e. AXTM will intercept whatever Rollback writes. But it'll be a total mess then (or maybe not??)

 

Yes, it is, and for me, extra confusing. But it does not matter, all that really matters is that it works. Perhaps at some point you (and perhaps others) could create a "how to use the 2 programs together" guide. This way folks like me can make the best use of the 2 programs without having to wonder about the why it works part. I do agree with the idea that some caution is called for though. As P said just because it works on your system does not guarantee it will work on other systems. I suspect it will but a single test system is not sufficient to declare it does. Still its very exciting.

Isso, I understand you intended AX64 to replace programs like Rx but you need to keep in mind that in many cases its not possible to have continuous access to external media. I use a desktop replacement Tablet PC (convertible) and much of the time its just the 1 internal drive. The fact that AX64s backups create a compressed full image for its first snap and then incrementals means its consuming drive space. Rx snaps consume minimal space since they are only recording the datas hooks. Yes, drive space is lost due to Rx protecting the original datas location to prevent it being overwritten, but if a snap is deleted and a snapshot defrag runs the space is made free again. Given this, for me, a combination of the 2 programs is ideal. When I am able I can connect an external drive and make AX64 snaps, when I am not, I can rely on Rx to offer protection.

 

Morning Isso!

Build 800 seems to have eliminated that -99969 error... BUT my first attempt to use the snapshots has me in a quandry...

I have a baseline BACKUP and (3) snapshots. I am in the process of trying to revert to Snap#1 and the restore seems to have hung at 8.2% (about 15-min. so far). Based on this my only recovery seems to be to re-image the system back from your recovery media which I will do.

The Baseline and the (3) snaps are located on an external USB drive which has worked fine. I'll try and duplicate this after I get this thing back up.

 

bgoodman, sure, that makes perfect sense to me!

 

Hi Forggie,

Sorry to hear this Did you by any chance notice if the GUI was responsive when the program hung? I.e. were you able to move the window?
Also what OS did you use?

 

System was completely frozen (W7x86sp1)...

"I'll be bach!!"

 

Interesting. How much RAM do you have on that system?

 

It's a li'l short at 1gB ... am awaiting an upg in the mail (just brought up W7Prox86sp1 recently to "re-invigorate" the laptop <2006> vintage).

PS- the recovery went fine BUT ... I just tried to revert to snap#2 instead (closer to reference point) and the recovery hung the same way at 18.5%.

Do you think the shrunken RAM is involved? If so, what would be the minimum requirement for this package?

 

We met this problem before on a machines with small amount of free RAM, but I was sure we fixed it. I'll double-check.

 

Normally my RAM base for any machine is at least 1.5gB which W7Pro works very well in. This is an old broken down XP lappy from 2006 (it runs well, just blew apart its hinges due to hinge freezing). I have recovered it thanks to the eBay supply house and am "bringing the PHOENIX out of the ashes."

Am doing another off-lin recovery to snap#3... if there's anything you'd like me to try, especially in the smaller RAM configuration, feel free to ask.

 

Thank you Froggie, I'll most likely contact you if we are unable to reproduce the problem. I tested on a VM with 768MB RAM and Win 7 and it was working fine. Maybe you have other processes that eat up RAM - care to check how much free RAM you have right before restore?

 

Isso, the machine is pretty bare... just W7 basically.

It has from 420-460mB available RAM prior to AXTM's restore.

 

Thank you. Maybe something was broken in the latest versions. I'll recheck and get back.

 

Dear Isso,

I hope that you still remember in all that excitement of Rollback Rx, AXTM doesn't work on any of my machines due to conflict with Acronis. I cares less about Rollback Rx.

Have you fixed the problem about AXTM coexisting with Acronis?

Best regards,

 

Dear Mohamed,

Sorry, I haven't yet looked into it - we are busy with SSD support at the moment, and only after we are done with it, I can proceed to compatibility tests.
I expect it to take another 1-2 weeks, and right after that I'll look into Acronis issue.

Regards,
Isso

 

I agree. I figured out how you achieve it, a little later after I posted.
The results that Carfal got must be related with the installation order. First installed AX64 and then RBRX. RBRX Driver (upperfilters) is loaded before the driver of AX64. If he changes the order of the upperfilters he will probably encounter the problems I described earlier, with corruption of RBRX structure. (Same situation with RBRX and IFW drivers).

Congrats about your idea and implementation of achieving the fast restores.
But I have some questions.
a) What will happen if there is intense disk activity (writes) on the disk and there is a power outage?
b) What overhead does it have in the write operations of the disk.
c) How will behave a machine with limited ram, that runs 24/7 without rebooting?

Panagiotis

 

The above is akin to tail wagging the dog. In this case Rollback Rx being the tail and wagging the development of AXTM and derailing the development of AXTM.

According to Isso, his original plan was AXTM to replace Rollback Rx, and it seems he shouldn't be derailed doing so.

Thank you Panagiotis for keeping Isso on the right track.

Many thanks for keeping your promise and making it compatible with other software out there.


Best regards,

 

Well said. Just a follow up question. An offline restore will be as rapid as online one or not?

Thanks

 

Yes, I think MBR restore might be kept there as deafult( for ordinary users) with an option to turn it off for( advanced users). This way ordinary user will not be left in trouble if he is doing a restore on a machine where MBR is messed up along with the OS.