Introducing AX64 Time Machine - hybrid imaging/snapshot software

Discussion in 'backup, imaging & disk mgmt' started by Isso, Jan 18, 2013.

  1. Peter2150

    Peter2150 Global Moderator

    Isso, i've wanted to stay silent when you mention comparisons with Shadow Protect, but I can't. What you are stating is very inaccurate and misleading.

    From my look at AXTM comparing to when I played with Rollback Rx, I think you have a real winner.

    But on the imaging front, ie recovery from a total disk disaster, I am afraid you are yet to be close at all to SP.

    First there is reliability. SP has been around for quite a few years. I've probably got close to 1000 restores under my belt on my machines(I've tested for SP) It's reliablity is excellent. I run Continuous Incrementals and the chain survives anything and everything. The incrementals take 4-5 seconds, and even in the evening if I've done work and several hundred megabytes have change I can manually execute the job and take an incremental in 10-15 seconds.

    As far as recovering files yes your browser works great, but I can mount any image in the chain in a few seconds and then explore it as another drive.

    I hope you do get there but compared to SP in imaging you have a way to go. I do wish you the best.

  2. Isso

    Isso Developer

    Hi Peter,

    Nice to see you in this thread :) First of all I'd like to thank you for taking time to review AXTM, and for calling me to explain your ideas. I really appreciate your feedback and have all your suggestions on our feature list.

    With all due respect let me disagree with your statement that my post is "very inaccurate and misleading". Here is why:

    1. Reliability. Of course, I'm not trying to compare an alpha version of a new program with a mature imaging software as SP in terms of reliability. I mentioned both in my original post, and on the download page that this is an alpha version that may have bugs. I believe that's suffucient to not mislead the users, and I shouldn't mention that anytime I compare AXTM to other programs.

    2. Speed. I just made some tests on a fresh Win 7 x86 install in a VMWare running on Core 2 Duo 3 GHz on a Crucial SSD drive. I compared SP to AXTM when backing up and restoring the system drive that has about 7 GB of data on it. The disk size is 30 GB. Backup is made to an internal disk. Here are the results:

    SP full backup - 4 min 6 sec
    SP incremental (no changes since full) - 1 min 50 sec
    SP restore - 40 sec to boot into recovery, 2 min 46 sec to restore - total 3 min 26 sec

    AX full backup - 3 min 5 sec (30% faster than SP)
    AX incremental - 9 sec (12 times faster than SP)
    AX restore - 6 sec (34 times faster than SP!)

    Isn't this enough to claim AX faster than SP?

    3. Backup browser vs ordinary mounting. Let's say you have a file in your My Documents folder, that you've accidentally deleted and you don't remember when. Let's assume you have 10 backups made. Your steps with SP or any other imaging software:

    - Mount the last image (5 sec). Use windows explorer to browse to X:\Users\User\My documents (5 sec). See if the file is there. If not, dismount this image (3 sec), mount the next one (5 sec). Repeat until you find the correct version of the file. This in worst case may take 10 attempts, totaling 180 sec, i.e. 3 min.

    Now the same operation with Backup Browser:

    - Click on the My documents folder in Backup browser (1 sec). Use mouse or keyboard arrows to switch between the backups and see if the file appread in the right view (1 sec). Keep clicking on the backups until the file is found. In worst case you will click 10 times, which will take 10 sec plus 1 sec - total 11 sec. Which is 16 times faster than SP or any other software that uses ordinary mounting!

    Isn't this enough to say that Backup browser is a more efficient method? And AXTM can mount the images too - just like SP.

    4. Ease of use - AXTM has only two buttons. With SP even I being a long time PC user, need some time to figure out how to do backup and restore.

    Of course SP has lot more features, but I mentioned that in my post.
    Please let me know if I'm missing something.

  3. WSFfan

    WSFfan Registered Member

    Very nice.
  4. carfal

    carfal Registered Member

    I have some more good (if not exciting:D :D ) news for everyone, especially for those of us who dont really want to part with Rollback Rx and would like the best of both.

    Dont faint or go running out the house screaming....but.... I've just spent the last 2 hours testing AXTM with Rollback RX v10 and the verdict is in.

    IT WORKS.......conditions apply

    Sorry conditions apply you say. What do i mean?

    Here it is.

    Firstly, let me be clear. I tested this in VMWARE Workstation virtual machine running W7x64. I wasnt game to test this out on my only working computer at home. However there's no reason for this setup not to work on a real system (famous last words!:p )

    Also I've just installed a brand new Samsung 840 Pro SSD. Sweet!

    I already had AXTM v1.0.800 installed in a VM and i decided that I'd be the first to really test Rollback and AX together. So i went ahead and also installed Rollback v10 only protecting drive C:. Since Rollback can only protect one physical drive, i dont have to mention that it's not a good idea to protect the AX backup drive for obvious reasons. (i think)

    It passed the first test. Rollback installed successfully. I havent tested this in reverse ie. installing AX then Rollback but i would think their wouldnt be an issue. Also before i had installed Rollback i had already been using AX for a couple of days and had about 3 snapshots before Rollback was installed. I know that its not exactly a very controlled way to start testing but I think that my starting condition will make no difference to the results. (and if it does then at least we all know how to set it up so that it does work..right?)

    Moving on, to my delight, Rollback was rock solid bar one problem. I'll tell you soon.

    AXTM also had a problem. Making a backup and restore with AXTM worked perfectly with Rollback RX except when rolling back(or forth) to a snapshot using Rollback RX. The first backup or restore with AX took a while (between 4 and 5 min). Whats worse the AXTM restore restored the WHOLE drive. So now comes the Rollback problem. What does Rollback do so efficiently....track every single sector change. I took a snapshot from Rollback. Crap! 15.5Gb snapshot (roughly the size of my W7 install). Running the Rollback defrag changed little. I was very disillusioned as you might expect. I thought then that these 2 are not going to get along.

    But i persevered. I did some more rollback and restores with both programs and after a while i noticed something wonderful. When i rolled back(or forth) to a snapshot with Rollback and did a restore with AX, it was a full drive restore. The same was true for a backup.....or so i thought. For some unknown reason i hadnt checked the file size of the backup of AX that appeared to be imaging the whole drive, allbeit in 5 min?? So i looked. File size 23mb. Whats going on?

    Who cares. I suddenly had a brain storm. I ran the following test. After I rolled back(or forth) using Rollback RX, I'll do a backup first with AX (which would take a little longer but crucially use less space...15.477gb less in fact) then take a Rollback snapshot (by the way, taking a snapshot with Rollback has no effect on the performance of AX, Hooray!) and see what size the Rollback snapshot is. Unbelievable. Rollback snapshot size = 80mb. Basically the file changes of the boot up process i"m guessing. This was a cause for celebration but i thought i'd put it on hold so i could relay the good news to everybody. Of coarse i tested many Rollback and AX snapshots and both programs performed flawlessly bar the first slow backup of AX.

    For the Rollback users out there, i found it strangely amusing that throughout all of the AX restores, Rollback's snapshot marker (the green arrow) always stayed pointed at the same snapshot no matter which restore point in AX I chose. Further to Rollbacks credit taking a snapshot would dutifully branch from this point and all Rollback restore points restored faithfully what AX restored or removed and vice versa.

    Even though this is about AX it needs to be said that Rollback V10 actually performed perfectly in all of this, even the 15.5gb snapshot is not a fault of Rollback but rather the result of what restoring a full drive does. The fact that Rollback survived and even recorded the change is amazing IMHO.

    This has really become long winded and i apologize but i think that it was important to tell my experience so that the match made in heaven can be apprecaited. :D


    1. After rolling back (or forth) to a Rollback Snapshot, immediately perform a backup with AX so that subsequent backups and restores using AX occur at normal speed. DO NOT DO A RESTORE FIRST WITH AX UNLESS YOU WANT ROLLBACK RX TO FILL YOUR HD REALLY FAST. This is the conditions apply bit.

    2. If you install Rollback after you've taken some snapshots with AX, DO NOT RESTORE WITH AX TO A SNAPSHOT BEFORE ROLLBACK RX WAS INSTALLED.This is something i just thought of and i think it would be the one thing to undo Rollback RX. So perhaps its a good idea to install Rollback Rx first and then AX second so that the first AX image will have Rollback Rx already there.

    3. Just in case some clever nutter finds a way around this, DO NOT ALLOW ROLLBACK RX TO PROTECT THE DRIVE THAT CONTAINS THE AX BACKUPS. Doing so will render AX pointless because everytime you roll back in Rollback RX, you also Rollback the AX snapshots. (it had to be said)

    Thats it. Now you can all cry with joy like i have been. :eek:

    Isso, i hope you dont miss this after all of my rambling. I would like to make a request and ask if you would be willing to look into the issue of after rolling back to a snapshot with Rollback RX, AX suffers from a long FIRST backup time or a long/large restore time (which ever one one chooses to do first).

    Thankyou everyone for reading.
    Last edited: Feb 5, 2013
  5. bgoodman4

    bgoodman4 Registered Member

    That is simply awesome carfal, thank you for doing and reporting these tests. Its just about everything I was hoping for. As to the SP AX comparison, it is well known that SP is the standard by which all others are judged and IMO if AX is or is not better, the same, or not quite as good I do not care. It will work with Rx thus making my system even safer. What more could I ask for than that.

    Now I have great respect for Peter as well as Isso and will leave the issues of back-up speed to a time when I am running the programs myself. As for the file recovery it sure sounds like AX beats SP on this point.

    I probably should re-read your post carfal but I am a bit short on time. It seems to me that you said that AX captured all of Rxs snaps but I may have misread or misunderstood this. Could I trouble you to clarify a bit on this?

    Isso, AX looks/sounds awesome, even if its not quite up to SP its still appears to be a darn fine product, I look forward to having it on my PC. Heck, if its as good as it looks I may even jettison Rx eventually.
  6. Peter2150

    Peter2150 Global Moderator

    Hi Isso

    You raise good points, and dialogue is always good.

    1. On reliability. Absolutely what you say is true. Heck even as a mature program SP has had some beauts of bugs. As a beta tester for them I've found them. Maybe a better term would be Robustness. How well does it handle what it isn't supposed to handle. For example on Restore you caution not to interrupt, well happens if you do. I tested that and it seemed random based on what was being restored at the moment. In one case it must been restoring a program folder and that was corrupted, but another restore and all was well. In the 2nd case it probably was working on the system, and it was game over. Could I have restored with the recovery disk? Don't know.

    2. Speed. On full backups, I tested with the a version of SP most folks wouldn't be using and both took enough time, that there was no statistical difference.

    On Incremental, you were using all of AXTM's capability but not all of SP's. I run the Continuous Incremental option and take an incremental every 30 minutes stopping at 5pm. Average incremental time is 4-7 seconds. BUT sometimes in the evening I do work and want it backed up, so I just manually invoke the sleeping job. Amount of changed data is usually in the 300mb range. Takes 15 seconds. One could easily setup CI job to just take one incremental a day, and then do them manually as needed and let this run literally for months.

    But where your comparison is really off is the restore. Your are comparing a time machine type restore to an image restore. Apples to oranges. Only time I ever restore with SP is when my system isn't bootable. So the AX restore should be measured that way.

    3. Backup Browser. If indeed I had to search thru that many files you are indeed correct. AXM is much quicker and easier. But a more experience user, the type who would be using either software shouldn't have to do that.
    Heck last night as I did a sync I realized I wiped out 3 files. I knew they were in the last back incremental of the day, so I just double clicked on it and had it mounted. Would have had to open both AXP and the browser to get started.

    And then there is the cheat:) If I really didn't know where the file was, I'd use neither program. I'd just check my cloud back up, because there they are never deleted.

    Isso, as I said for time machine concepts you are the clear winner.

    I do have one biggie concern about the imaging/restoring. Can AXP do a bare metal restore, where you don't have the System Volume Folder and the file you put there? Also I tested the new version this morning. When I tried making a FDISR snapshot this morning, I tested a)right after install and it was perfect, and b) right after a AXM snapshot, and FDISR failed in that it couldn't copy your file in the System Volume Folder. Apparently VSS couldn't isolate it. That's when I tested Shadowprotect. But I wasn't using the Desktop version, and saw no vss or substitute show up in the log so I am not sure how they do it. I was imaging from the desktop but with a special version, not from the recovery disk.

    So what happens with the various imaging programs that people might want to run from their desktops? I don't know.

    Last edited: Feb 5, 2013
  7. Isso

    Isso Developer

    carfal, that's unbelievable! Fantastic job my friend, thank you! :thumb:

    The reason why AXTM makes full backup or restore after you use RollbackRX is apparently the following: AXTM saves its settings, including the last backup name on the drive that it protects. When you use Rollback to restore its snapshot, that settings gets overwritten. So AXTM figures out that the current drive state doesn't correspond to the last backup that it has, and uses full backup or restore (otherwise the drive could get corrupted)
    Hope this makes sense.

    Thank you very much again for your hard work and for sharing your findings - this will be very useful!

  8. Isso

    Isso Developer

    bgoodman4, thank you! Looking forward to your review ;)
  9. Isso

    Isso Developer

    Peter, thank you! Sure dialog is good. For me it's always a way to learn something new.

    For your points:

    1. Interrupting the restore. IMO no imaging program will leave the disk in a consistent state if you interrupt the restore process. That's just the limitation of imaging process, not AXTM. Should something like that happen (like power goes off during restore) you can always use recovery media to do a full restore.

    2. Speed - I didn't know about Continuous Incremental feature of SP, thanks for letting me know! Will check it out. For restore - why not compare AX fast restore with SP? It's doing the same job, the end results are exactly the same, so I fail to see why you think it's "Apples to oranges", sorry. If SP doesn't have such feature, it's their problem :)

    3. I agree, some users may prefer mounting. Anyway, when we add mounting with a double-click, SP clearly will have no advantage, thank you for the idea :)

    For restore - you can format the drive, and AX will restore it with no problem, the setting file doesn't prevent the restore. It will be a full restore though. The settings file is locked exclusively so nobody is able to copy it, apparently that's why FDISR complains. That shouldn't in any way affect VSS operation, and imaging programs should work fine.

  10. starfish_001

    starfish_001 Registered Member

    Power failure during restore is interesting I guess you have to go to the last full backup and lose the incrementals it would be nice to be able to recovery those rather than orphan them.

    I'd quite like a boot menu like FDISR - to flip between snapshots rather than adding a usb. but it nice to have the option of usb in case of real disaster Windows 8 has forced me to restore images rather than FD snapshots

    In terms of speed I use SP and FDISR - speed of AX is ok but not quite as good as SP for me. But is close enough for me to not complain.

    For me I have 4 lics for rollback RX - don't use any beyond VM because it is not good enough.

    AX looks like a great alternative to FD .. I'm hoping that this is a success
  11. Peter2150

    Peter2150 Global Moderator

    Hi again

    It's really not a problem for them, as there big dollar customers don't care. One thing they have done that I've not played with is the OS can start booting while the restore is proceeding.

    As to why not compare the AX fast restore with SP, one question, how exactly do I do the fast restore if the OS is trashed?

  12. pandlouk

    pandlouk Registered Member

  13. Cruise

    Cruise Registered Member


    Which OS is used for AX64's recovery boot disk? - I'm coming in late here, so maybe this has already been answered. :oops:

    Is it necessary to perform a restart (after installation) before creating recovery media? I ask because I have not yet restarted and I'm getting an error trying to create a recovery disk (on a 2GB USB flash)! :doubt:

    Last edited: Feb 5, 2013
  14. Isso

    Isso Developer


    Power failure during restore, and subsequent full restore will not affect existing snapshots - everything will be preserved, and you can continue taking new snapshots after restore just as usual.

    Boot menu - thank you, I'll think about it, may be a nice option. But you don't need to use USB for restore - just open Backup browser, select the snapshot and restore. USB recovery is used only when the system is unbootable.

    Thank you!
  15. Isso

    Isso Developer

    Peter, sorry there is no way to do a fast restore if your OS is not booting. Recovery media with full restore is the only option if OS is trashed.

  16. Isso

    Isso Developer


    OS for recovery media is the same that the OS that you're running AXTM on. I.e. if you create Recovery media on Win 7 - you'll have WinPE 3, if you run it on 8, you'll have Win PE 4, if on XP - the original PE.
    What AXTM does is building the recovery media using the files from installed OS.

    No restart is needed to create the recovery disk. Could you let me know the error and your OS?

  17. pandlouk

    pandlouk Registered Member

    Isso if you do implement it make it optional.

    By the way, it can be very handy with the native app for users/admins that want to have their systems in static mode. e.g. during the first start of the day or during a reboot to always restore to a given snapshot.

  18. Isso

    Isso Developer


    About boot menu - sure, it will be optional if we add it.

    Regarding native app and restore on reboot - very interesting idea, will think about it, thank you!
  19. Cruise

    Cruise Registered Member


    I'm using W7x64. The only thing AX64 reports is 'ERROR! Please try other media'. I'm quite sure the USB flash drive is good, but is 2GB large enough? :doubt:

    Last edited: Feb 5, 2013
  20. Isso

    Isso Developer

    Got it, thank you! This error happened to others as well. I'll release a new version with logging soon - that should enable me to find the problem. 2GB is more than enough.
  21. Cruise

    Cruise Registered Member

    Isso, here's some more info of interest (and concern)...

    1. By creating a Recovery ISO I was able to create the USB Recovery Drive (using PowerISO).

    2. May I suggest that you provide a 'No Network' Option in your Recovery Media Builder for ignoring network devices. It takes much too long to search for them during bootup and many of us are just working with our local PC and an externally connected backup drive.

    3. A serious issue... Upon booting up my USB Recovery, I get a functioning AX64 GUI but it does not see my internal HDD !?! :doubt:
    (fwiw, I've experienced that with other PE boot disks).

    Last edited: Feb 5, 2013
  22. kda

    kda Registered Member

    I've spent the last two weeks searching for a solution to safe snapshots and roll backs on an SSD. By accident I stumbled onto AX64 and I must say, this looks like the solution I will go with. I am really liking what I'm seeing.

    An honest, open developer with some solid ideas on how to solve an issue that affects almost every computer user that installs new software, takes updates or likes to experiment with new software. I think if you can get the word out, your market will be HUGE and I will be one of those near the front of the line.

    When you manage to take the pain and risks out of using one's computer and you do so in a simple and totally safe manner ... you are going to make a LOT of friends ISSO. You instill a lot of trust with your frank and open communication. Very impressive indeed. Put me on the list as an early adopter. :)

  23. Isso

    Isso Developer


    Sure, we'll add an option to skip network init in PE.
    For PE not seeing your HDD - that's very strange. What is displayed when you hit "Open Backup Folder"? No drives at all?
    What type of HDD do you have (SATA? RAID maybe?) and what's the chipset of the motherboard? Thank you

  24. Isso

    Isso Developer

    Hi kda,

    Thank you very much for your kind words, it's very encouraging!
    We aren't quite done with SSD support yet, but I'm planning to complete it in about 2 weeks, so you'll have a chance to test it by then. Thank you!

  25. carfal

    carfal Registered Member

    Hi Panagiotis. Your conclusions are very interesting but I'm not convinced that the live restoration method that only AXtm uses destroys the RBRX structure.

    I ran some further tests.

    1. Used RBRX to roll back to a Snapshot where i removed a program called PowerIso

    2. Ran a backup using AX to bypass the first restore issue i spoke of in my earlier post

    3. Used a AX snapshot to restore PI

    4. Updated the Baseline with Rollback - Success

    PowerIso was still there in the subsequent new baseline of Rollback.

    Continuing from there

    5. Took a backup snapshot using AX for reasons already outlined. (strangely this first backup went at normal speed??)

    6. Restored to a previous snapshot using AX that was taken from the previous RBRX structure before baseline update where PI was uninstalled. (I know.....redical right) - Success. I can't see PI. Its gone.

    7. Did another RBRX baseline update

    8. PI was now permanently gone! - success

    My Conclusions

    I think that the live restore method that only AX uses keeps the RBRX structure intact. I believe that if IFW was used on a system setup like this that it would restore fully intact. The unique live restore method must allow RBRX to successfully keep track of everychange, thus keeping its structure intact.

    Even if an image restore using other Imaging tools will in fact prove that the Rollback structure is not intact, this is a non issue for me but understand that it could be an important consideration for others.

    In my case, I would restore my image using AXTM recovery cd (because i would be able to restore right up to the latest snapshot) and then reinstall RBRX. Problem solved. :D Further more, if use a snapshot taking regime of taking a snapshot with RBRX and a matching one with AXTM i would in fact still have my snapshots intact because they would exist on the AXTM side (this is truely amazing).

    I'll consider testing the restoring an image with IFW and reseting baseline thingy at a later stage before AXTM is ready for prime time. ;)