Rollback versus FDISR - An Update

Hi All

As I indicated in the main thread, I've pulled Rollback from my Desktop and gone back to FDISR. The reason:

I originally tried Rollback as I had a problem with FDISR(note, it never failed me). Raxco and Leapfrog worked and fixed the problem. I was very excited about Rollback as it offered some significant advantages for the type of testing I do. The two things in particular that were of value to me. Being able to Rollback with a simple reboot, as opposed to a reboot,copy,reboot. A significant time difference. Also the ability to install two new programs and quickly separate them with a snapshot.

So what went wrong. Well first, I beta several programs that operate at the kernel level, so I am used to crashes, system hangs, etc. One of the things that I started noticing was that after many crashes, or hangs there was having chkdsk run. Never had that with FDISR. Also during these chkdsk runs, I saw some pretty scary error messages, I'd never seen before. In all fairness to Rollback things always corrected okay, but it was a bit scary. Also I am testing both KAV 6.0 and OA AV+ which uses the KAV engine. Since the install the same files, in the same place, but different versions, clearly isolation was essential. I was beginning to suspect it wasn't as good as it should be.

Last weekend I uninstalled Rollback as I had been, and then reinstalled. The installation included SSM and Avast AV. THe installation snapshot became snapshot 1. I created a 2nd snapshot as a buffer, uninstalled Avast, and installed OA AV+ and create snapshot 3. Then created snapshot 4 as a buffer, and uninstalled OA and installed KAV, and made snapshot 5. I was able to boot back and forth without problem.

Then I wanted to update versions on OA AV+ and I booted back to snapshot 3 and uninstalled the old version and installed the new version, and saved a new snapshot 3A. I got this idea from Aigle, and HDS confirmed that this is fine. Then defragged(HDS confirmed this also was no problem) I then booted back into snapshot 5 and discovered KAV had been trashed, SSM had been trashed and I had no sound. This shouldn't have happened. Backed up to Snapshot 4 and at least had sound, so I stayed there, and repaired.

I called HDS and first confirmed everything I'd done was okay. They said yes. THen I explained what happened. I was told no other people using it for beta testing had these issues, and it was suggested maybe a virus was at play. I asserted my system is clean. It was then suggested that to trouble shoot and eliminate conflicts I should install a clean system, install Rollback and go from there to try and isolate the conflict. Nuts to that. I can't install a clean system, and the whole point of using this type of program is to recover from the result of conflicts.

Since I have some more testing(read crashes and hangs) I need something totally reliable and that is FDISR. For simple uses I would be comfortable with Rollback, but with hardcore beta testing, I need reliablity, and that has returned me to FDISR. All the problems with Rollback lately totally wiped out the time saving.

I will continue to follow Rollback closely, as I've learned alot about it, and it has some real potential.

Pete

 

I pretty much had the same experience. Funny that the last thing was trying the "fork in the road" so to speak. I also wonder if it was the combo of two KAV based programs with the same name drivers. still that should be okay.

Have you guys seen it thru many crashes and system hangs, where you need to power reset??

 

Some times I've used the Home key and the intercept, and yes it always did fine. It's only after a crash or power reset and just a straight reboot, I've seen Rollback check the disk, and often it would be fine, but then occasionally no, and as I said Chkdsk always fixed the problem. Be interesting to get Chris's take.

Also my other issue may be related to two programs using the same file structure. With FDISR when you have a snapshot there are two separate copies of files. But with Rollback when I have a driver KLIF.sys and I create a new snapshot, the new snapshot doesn't have another copy of that file. So initially the new snapshot has to use the other snapshot's file. That may be where the issue is coming into play in what I was doing.

Could be a fluke, but I was dismayed by HDS's response.

Pete

 

Funny thing was I did check my disk prior to last RB install. You are right about the bare metal recovery issue. That is also a factor to me in switching back. Yes imaging takes longer, but the FD Archives are a tested way to also recover from hard drive failure, and those refresh daily in a few minutes.

 

Hi nicM

I am as puzzled as the rest of you, as that was my experience prior to what happened.

Pete

 

Just out of curiosity, what file system are you all using. I am using NTFS with no compression.

Pete

 

Lest anyone get the wrong idea, none of the crashes or hangs requiring power resets were caused by Rollback. My concern was when it happened the frequency of chkdsk running, and the error messages.

 

The link was the fact that many of the crashes and power resets were triggering a chkdsk run. Frequency of chkdsk to crash was very high as opposed to being very low when I was using FDISR. Also the messages. There were lots of index's being deleted, and replaced, files being recovered, and even saw one (didn't catch the details) that made reference to the MFT.
Now as I said after it was all repaired I didn't see any problems. But it is scary to see this when you've not seen it before, and it makes you wonder if one time...... Then add in the problem with the snapshots, and I just am leary of problems. So...

Pete

 

Hi Tobacco

Yes you are. I like Rollback as much as FDISR, but I have to have a high degree of confidence, and if there is a problem get it resolved. Up to this point the customer support from Rollback was fairly good so I was surprised. But saying no one else has a problem doesn't cut it. The bug I found in FDISR was a total surprise to them, but they worked it until solved. Thats what I expected, and didn't see. I certainly am still open to working with them, but in the meantime, I just had to switch back.

Pete

 

Hi Mike

I can't for sure say that Rollback was causing Chkdsk to run more, but by observation, I saw many more crashes and hangs cause chkdsk to run since I've been using Rollback then before.

As to your comment about a "sick" system. Define what you mean by a sick system. Then I can comment on whether I think it applies, in relation to HDS deciding it wasn't Rollback and giving up.

Pete

 

Hi Mike

Okay, I think we have miscommunicated. As it stands right now my system is very stable, but beta testing programs working at the kernel level has it moments. For example, KAV 6.0 has been fairly stable, but every now and then a new beta build BSOD's everyone who tries it. So in the process of finding out you can have several crashes. I was beta testing Process Guard in it's early days, and I'll never forget version 1.3. It was so unstable at first, that i would crash 4 or 5 times an hour. Sure take that beta off and my system was perfectly stable. But by doing that I help the programmers narrow done what was going on. I've even run prototype code for one developer. Some was stable, some was.... When I am not doing this kind of testing my system is quite stable. This is definitely not what the average user would see. So starting with a fresh system, and putting all my normal software on would result in the same stable system I have when not testing, but once testing I'd probably again see crashes and hangs.


But if HDS would walk away for the reasons you suggested(and I am not sure I believe that) then I would know I made the right decision. Let me explain.

I don't know if you read the thread where I described the FDISR problem I had. It would boot into existing snapshots, but not new ones. I was getting an error when it would attempt its roll over into a new snapshot, and would fail. I called Raxco and they had never heard of that error. Since screen capture was out of the question, they asked me to take a picture which I did. The developer came back and basically said that error could not and should not happen. But they kept sending me test programs to run so I can send them back logs. They finally nailed the problem down. I happened to have an unusually large number of files in my c:\ directory, and that caused a record in the MFT to be almost full, and FDISR could write an attribute it needed to set. They could have said that was my fault, but they treated it formally as a bug, as the program encountered a condition should be able to handle and it couldn't. The made a fix and are encorporating it into the next version.

By the same token, I had worked thru 5 snapshot, all with stable programs, then went back to snapshot 3 and updated a program with a new version that has proved even more stable. I tested all the stuff in each snapshot. So when I then boot from one to another and stuff gets messed up, how can I blame anything but Rollback. I would felt alot better if they said hey let us send you something, so you can try and duplicate the problem and get us some info.

Also consider the whole reason I run Rollback or FDISR is so I can recover if one of these beta's trashes something badly.

Hope this makes sense.

Pete(off to find out why KAV6.0 beta won't run on the laptop )