"Rollback RX" luvs "Image For Windows" - REDUX

Discussion in 'backup, imaging & disk mgmt' started by TheRollbackFrog, Apr 26, 2015.

  1. TheRollbackFrog

    TheRollbackFrog Registered Member

    Joined:
    Mar 1, 2011
    Posts:
    3,042
    Location:
    The Pond - USA
    Greetings! Some of the old time Rollback RX users may have been familiar with an effort championed a long time ago by many of us (previous?) Rollback RX users to find a successful way of imaging Rollback protected partitions, including all their snapshots. We wanted to be able to do this on a LIVE Windows system so that we didn't have to shut the system down in order to run our protection mechanisms.

    The effort culminated in a document, previously available in a Horizon DataSys Forum, that detailed how to accomplish this effort, to insure that our Rollback protected systems were properly backed up and that we could reproduce them if catastrophic failure occurred within our system.

    Following the production of that document, evolution of both Rollback RX and the Image For Windows products caused a failure in that procedure. It was recommended at that time that the procedure be abandoned until such time as a new procedure. was properly vetted for success. it's been a long time since that occurred.

    Recently, a group of us has briefly gone back into the "lab" to determine if a procedure may be developed to accomplish this task once again. The result has been very encouraging. We have successfully reproduced some of the success we enjoyed years ago but need to vet this new process across both newer operating systems (W7, W8.x & W10) and newer BiOS and disk geometries (UEFI, GPT). If successful, we may have finally found a way to protect our Rollback RX systems (and their snapshots) against a series of possible failure modes that have plagued us for years.

    I will try and answer any questions you may have at this time, and will keep you up to date, via this thread, as to our progress... please wish us luck (we're gonna need it... it's Rollback RX, remember :eek: :D ).
     
    Last edited: Apr 26, 2015
  2. pandlouk

    pandlouk Registered Member

    Joined:
    Jul 15, 2007
    Posts:
    2,549
    Froggie, that method can potentially be dangerous for the live RBRX environment.
    PHYLOCK's temp file is created with normal windows APIs so it resides in the current snapshot and is not a problem.
    But we do not know how PHYLOCK releases its' buffering/updates the file system after the imaging process is finished. If RollbackRX can intercept the changes and isolate them in the current snapshot is ok, but if it also "touches" the baseline filesystem the user will have a corrupted baseline until he performs a restore of the IFW image.

    Panagiotis
     
  3. TheRollbackFrog

    TheRollbackFrog Registered Member

    Joined:
    Mar 1, 2011
    Posts:
    3,042
    Location:
    The Pond - USA
    Panagiotis, thanks for reminding me of that :thumb: We talked about that years ago and it was always a concern but never got to the bottom of it. I'm going to see if I can get an "inside" to a Phylock/IFW developer and get some additional detail without exposing any intellectual property.

    Phylock was originally developed when VSS was not available on the OSes they were marketing to... since then, VSS has been the GoTo method for freezing a FileSystem during an imaging operation by almost all imagers... even IFW offers VSS as a method if the user wishes to use it. Maybe PhyLock secrets are not as sensitive as they used to be. If they still are, I'll just get Edward Snowden to get me the strait skinny on PhyLock's design :argh:
     
    Last edited: Apr 27, 2015
  4. TheRollbackFrog

    TheRollbackFrog Registered Member

    Joined:
    Mar 1, 2011
    Posts:
    3,042
    Location:
    The Pond - USA
    Panagiotis, I've issued the appropriate questions in the TeraByte Forums but I also tried to get a head start on this issue. In doing so I did the following... lemme know if you think it sheds any additional light on the situation.

    With the UpperFilters set to allow PhyLock to be first in the chain (the apparently successfully method), I then fired off an "All Sector" IFW image of a Rollback protected partition. While the image was in progress, and hopefully PhyLock cacheing any system changes during the process, I generated some significant changes to the file structure. My expectation was at the end of the image, those changes would be merged into the Windows FileSYstem... what I didn't know was where on a Rollback protected partition... into Rollback's current snapshot or the original installation Baseline.

    Following the image and the release of PhyLock's lock, I BOOTed into a WinPE and used an Explorer to see if those cached changes were in Rollback's original baseline... they were not. Then, using Rollback's sub-console, I BOOTed into Rollback's baseline and made the same check... no changes in the baseline were found (I expected this as the WinPE found nothing). I then rolled back to the snapshot where the test "All Sector" image was done and found all my cached changes at that point in time.

    I can't account for everything PhyLock does during the LOCK release (I just don't know how it all works yet), but the above checks kinda tell me that the LOCK release and cached changes are being merged back into Rollback's snapshot rather than the baseline.

    Wha'd'ya think?
     
  5. pandlouk

    pandlouk Registered Member

    Joined:
    Jul 15, 2007
    Posts:
    2,549
    It seems that does not touch baseline's file-table. But for being on the safe side run also a chkdsk to the partition from inside a PE; if it comes clean (e.g. no fixes, invalid entries, etc.) it should be ok.

    Panagiotis
     
  6. TheRollbackFrog

    TheRollbackFrog Registered Member

    Joined:
    Mar 1, 2011
    Posts:
    3,042
    Location:
    The Pond - USA
    Clean as a whistle! Who knows... we just may be close :ninja: :doubt:
     
  7. pandlouk

    pandlouk Registered Member

    Joined:
    Jul 15, 2007
    Posts:
    2,549
    Then you are ready to go... although I cannot understand why anyone would want to capture all the snapshots; it's not like someone can continue using them for more than some months, sooner all later one will have to update RollbackRX and he'll have to start over= too much effort for nothing.

    Panagiotis
     
  8. TheRollbackFrog

    TheRollbackFrog Registered Member

    Joined:
    Mar 1, 2011
    Posts:
    3,042
    Location:
    The Pond - USA
    It was even worse for me when I was using this extensively a few years back. I had 3-4 separate timelines running off the installation Baseline... all were separate system environments. When the time came to do what you mention above... it was brutal :eek:

    Panagiotis, thanks for all your help with this!
     
    Last edited: Apr 29, 2015
  9. pandlouk

    pandlouk Registered Member

    Joined:
    Jul 15, 2007
    Posts:
    2,549
    For such scenarios the ideal program is InstantRecovery; you can update/upgrade/downgrade it's snapshots and archives at will.
    You are very welcome.

    Panagiotis
     
  10. jwcca

    jwcca Registered Member

    Joined:
    Dec 6, 2003
    Posts:
    716
    Location:
    Toronto
    That's where Raxco's IR (or FD-ISR) would help, a separate snapshot (plus archive(s) for each timeline. It would take more disk space but it's much easier to control.
    J
     
  11. pvsurfer

    pvsurfer Registered Member

    Joined:
    Sep 1, 2004
    Posts:
    1,400
    Location:
    California - USA
    Hi Panagiotis.

    I would not argue with you on that point when you speak about retaining RBrx snaps over a period of months. In my use of RBrx a snapshot is taken automatically each day at startup. Whenever there are Windows patches or updates to my apps I will manually take a snap before each such update. Sometimes I also use snaps to work in a 'test environment'. Once a week I always make a Cold-Raw backup (using DS). This isn't too much of a burden (time & storage wise) as my system partition is relatively small. Every one or two months I uninstall RBrx to the current system state, run chkdsk/f, then do a file cleanup and a file defrag. Then I reinstall RBrx and the cycle repeats.

    You (and others) may question if using RBrx is worth the time and effort, but to me it definitely is (I have the routine down to where it's not much of a chore). And who know, after all these years DC6 may wind up a blessing. If not, Froggie's IFW-PhyLock method may very well be of interest. ;)

    pv
     
  12. pandlouk

    pandlouk Registered Member

    Joined:
    Jul 15, 2007
    Posts:
    2,549
    Actually the way you use RollbackRx is probably the only way to not brake it. (e.g. not use more than 15-20 snapshots and reset or update the baseline and defrag every 1-3 months).
    And for a scenario like yours the ideal app (since February 2015) is Macrium 6. If you have a spare pc to give it a try, you'll come to realise how much time and effort you actually spend in babysitting RollbackRx.

    Panagiotis
     
  13. pvsurfer

    pvsurfer Registered Member

    Joined:
    Sep 1, 2004
    Posts:
    1,400
    Location:
    California - USA
    :argh: LOL... Well with all of the fanfare MR6 is getting on Wilders I guess it's worth 'a look-see', so I will d/l a trial and install it on my old Dell (XP) desktop. It supports XP I hope.
     
  14. Kurtis Smejkal

    Kurtis Smejkal Registered Member

    Joined:
    Mar 17, 2015
    Posts:
    253
    Location:
    Vancouver, BC
    That's a very sensible way of looking at it. Here's hoping Drive Cloner Version 6 is a good fix for that.
     
  15. Peter2150

    Peter2150 Global Moderator

    Joined:
    Sep 20, 2003
    Posts:
    17,039
    Amen on two things. First for some applications you just can't beat IR. May be "slow" and "Clunky" but still unbeatable. Not to recently I installed Rollback 10.3 (or latest version) on my play desktop. Creating snapshot was extremely fast compared to IR or normal image taking. Same was true with switching snapshots. But that all got blew away with Macrium V6. Snapshot creation as quick as RBX, restoring an image quick then swapping over in Rollback. But no issues with regular defraggers, no problems with Windows Updates, no issues with other imagers, and in fact no need for others.

    So with Macrium v6 I would have to ask why bother with Rollback?
     
  16. pvsurfer

    pvsurfer Registered Member

    Joined:
    Sep 1, 2004
    Posts:
    1,400
    Location:
    California - USA
    Pete,

    Don't take this as a RBrx is better than MR6 remark, but I don't believe your 'no issues with defraggers' remark. It's got to be one and the same issue with RBrx and MR6 (or any other incremental/differential imager). Using a file defragger between snapshots or images will result in extraordinarily large subsequent snaps/images.

    pv
     
  17. TheRollbackFrog

    TheRollbackFrog Registered Member

    Joined:
    Mar 1, 2011
    Posts:
    3,042
    Location:
    The Pond - USA
    PV, I think what Pete is talking about is not "normal" defragging that surely causes large image DELTAs, he's referring to the BOOT time defrag that used to kill AX64 and other "tracking" type of imagers.
     
  18. pandlouk

    pandlouk Registered Member

    Joined:
    Jul 15, 2007
    Posts:
    2,549
    Actually it depends on how you configure Macrium 6.
    e.g. "Synthetic Full- Forever incremental" does not suffer from such problems
    https://www.youtube.com/watch?v=6aeWSmT2TP0

    Panagiotis
     
  19. pvsurfer

    pvsurfer Registered Member

    Joined:
    Sep 1, 2004
    Posts:
    1,400
    Location:
    California - USA
    That's not the way I read it. I read it to say that using (file) defraggers brings about issues for RBrx but not for MR6. :rolleyes:
     
  20. TheRollbackFrog

    TheRollbackFrog Registered Member

    Joined:
    Mar 1, 2011
    Posts:
    3,042
    Location:
    The Pond - USA
    Yep, those issues are exactly the same with any imager whether it uses COPY ON WRITE (standard imaging) or REDIRECT ON WRITE (Rollback RX).
     
  21. pvsurfer

    pvsurfer Registered Member

    Joined:
    Sep 1, 2004
    Posts:
    1,400
    Location:
    California - USA
    Panagiotis, I just don't see that. The incremental following a file defrag will be much larger than it would otherwise. I see where the synthetic 'rollup' will eventually mask that, but it's the same issue as far as I can tell.
     
  22. pvsurfer

    pvsurfer Registered Member

    Joined:
    Sep 1, 2004
    Posts:
    1,400
    Location:
    California - USA
    Exactly (and you say it so much better)! ;)
     
  23. pandlouk

    pandlouk Registered Member

    Joined:
    Jul 15, 2007
    Posts:
    2,549
    Sure, until the full is merged with the incremental the issue is the same but after it merges is gone.

    Panagiotis
     
  24. jwcca

    jwcca Registered Member

    Joined:
    Dec 6, 2003
    Posts:
    716
    Location:
    Toronto
    or just use an SSD and skip defragging...:shifty:
     
  25. appster

    appster Registered Member

    Joined:
    Jun 19, 2007
    Posts:
    530
    Location:
    Paradise (Hawaii)
    pv, I recall you saying (in another thread) that you considered it risky to run system utilities inside a Rollback Rx installation. Based on your long and successful use of Rollback Rx I have embraced the above procedures. :thumb:
     
Loading...