disParity

disParity » Bug Reports

1. Porcus
   Member

   For the hell of it, I went above the edge of full.

   Good:

   - It responds with a Unable to expand files1.dat to add message. That part works, or at least it indicates to do so.

   Not so good:

   - There are still some Object reference not set to an instance of an object.. Indeed, it seems to me that this error is chiefly on drive 0, while the “Unable” error is chiefly drive 1 (I am only protecting two drives).

   - It leaves a files0.bak file, same size as the files0.dat, modified the same second (but the .dat last accessed one second later) - and, by inspection, bit-identical. As long as they are bit-identical, it is not much of a problem, but generally I would like to know which one represents the actual state of the protection.

   No crash file created.

   Posted 2 years ago #

2. Porcus
   Member

   Oh, by the way, a feature request:

   When I first start disParity, it starts scanning. If I interrupt scanning, it goes green and up to date. Maybe turn yellow and report that scanning aborted?

   (Why do I interrupt scanning? It takes several minutes, and I might find out that “oh, I should have done this change too”, then abort (in order not to slow down read/write while I fix up), and then ... forget that I had aborted, only to see it light green next time that window is active :-o

   Posted 2 years ago #

3. Porcus
   Member

   Trying to run more times:

   Error removing z:\[pathname]1.VOB: Object reference not set to an instance of an object.
   Error removing z:\[pathname]2.VOB: Object reference not set to an instance of an object.
   Error removing z:\[pathname]3.VOB: Object reference not set to an instance of an object.
   Unexpected error while processing y:\[pathname]1.VOB: Object reference not set to an instance of an object.
   Unexpected error while processing y:\[pathname]2.VOB: Object reference not set to an instance of an object.
   Unexpected error while processing y:\[pathname]3.VOB: Object reference not set to an instance of an object.

   (Hm. Actually, the .bak files are kept even for updates which end successfully. Intentional?)

   Posted 2 years ago #

4. Roland
   Roland

   Any time you see a "Object reference not set to an instance of an object", that is definitely a bug of some sort. There's no valid reason for that message. Can you email me your log or logs from the time you ran these tests? Send them to rolandv@gmail.com.

   Yes for now I am deliberately leaving the .bak files there, as an extra layer of protection. If the original .dat cannot be read later on, in some cases the code will try to restore from the .bak. The code always creates a new .bak before any attempt to modify the .dat.

   Regarding interrupting a scan, the drive status *should* simply revert to whatever it was before. In other words, an interrupted scan is the same as no scan having been performed. Is this what you are seeing? There should also be a "Scan cancelled" message shown for "Backup status" under the drive list.

   By the way, several minutes is awfully long for a scan! An initial scan can sometimes take a while because the OS hasn't cached the NTFS file tables in memory yet, but subsequent scans should be faster. How many files are there on your drives?

   Posted 2 years ago #

5. Porcus
   Member

   You've got mail.

   Posted 2 years ago #

6. Roland
   Roland

   Received. Reply sent. Thanks.

   Posted 2 years ago #

7. rust0r
   Member

   I also received the Object reference error on a few files, I will forward the log shortly.

   Roland, one thing I wanted to request, was perhaps going back to the old logging method (I'm not sure if you switched for a reason or not), I frequently look back to see when the last parity was created, or if there were any failures on larger updates I have done. It is VERY helpful to have them all sorted by date and time so I can view them as required.

   A log window would be amazing as well, similar to how you could watch what it was doing back in the old dos/windowed version (which I just upgraded from), not crucial, but it's nice to see. I would be more than satisfied with having separate logs again so I can check them after.

   I'm VERY impressed with the new UI and features, it feels solid and very nicely done. Now I just need to get more ram to fully take advantage of it :)

   Thank you again for your work on disParity, I will continue to preach to the masses :)

   Posted 2 years ago #

8. Roland
   Roland

   Hi rust0r,

   I'm looking forward to examining your logs, if you get a chance to send them. I don't currently understand what could be causing that error so I'd love to get to the bottom of it.

   Regarding the logs, I changed the logging approach because disParity is now designed to be a long-running application rather than a single-operation command line utility. Continuous log files seem more appropriate (at least to me) for applications that remain running, although I can certainly understand why you found the old approach handy as well.

   I do plan at some point to add some more options that allow you to control various aspects of the log files.

   If you like watching the log messages as disParity is running, here is one "poor man's" approach that I sometimes use when debugging: install a simple unix command line utility like Cygwin, then "cd" to disParity's log directory, and then do a "tail -f disParity.log". You should get the same effect as you would get from a log window.

   Posted 2 years ago #

9. rust0r
   Member

   Hi Roland,

   I just sent off the log, I only had the error happen once and never again, I'm not sure whether disparity managed to correct it, or Drive Bender changed the files and disparity was able to then read them (as it was drive bender reg files that it had the problem with)

   I'll have to take a look into that poor mans log window, I definitely enjoyed seeing the progress of it. I don't have any large updates to do at the moment but will keep this in mind for the future.

   Thanks again and hope that helps

   Posted 2 years ago #

RSS feed for this topic

Reply

You must log in to post.