To read all comments associated with this story, please click here.
Partition Magic often complains about cylinder boundries, and once lost all the partitions on an important hard drive. I managed to recover the partitions using a hex editor and a lot of trial and error. Backing up the partition table before doing anything that will alter it is proabbly a good idea.
Yes, this is one of the known reliability problems of qtparted. Rarely it may also change the head count in the partition table and also fails to refuse resizing dynamic disks. Thankfully none results real data loss (all data is accessible from Linux but Windows can't boot) and all case is recoverable without any data loss (including yours). These qtparted problems are known for 1.5 years and were communicated to three of its developers several times. Unfortunately nothing happened.
I didn't write that "most problems are due to repartitioning". I wrote that so far "all". In numbers it's about 100,000+ (partition resizing) versus 0 (ntfs resizing). The number is so high because the new disk geometry support code of the new 2.6 kernels fooled most partitioners last year and they consequently corrupted quite many partition tables: http://lwn.net/Articles/86835/ This has affected only a few percent of the users and most vendors fixed this problem very quickly.
I also wrote that the partition table manipulation is "irrelevant to NTFS" because filesystem resizers can't do anything about these bugs because they don't handle the partition table at all. Similar problems keep happenning in case of FAT32, ext2, ext3 and reiserfs too. These problems are clearly only partition table manipulation related in qtparted and they aren't filesystem level bugs what many users incorrectly think.






Member since:
2005-12-17
The last (and only) time I tried qtparted, from SystemRescueCD, I lost my NTFS partition. Why? Apparently (as far as I can tell), qtparted (which uses ntfsresize) insists that the start and end of partitions be aligned with cylinder boundaries. My 2 existing NTFS partitons (as set up by the OEM), were not aligned on cylinder boundaries. The resizing here only shrinks you partition at the end, but does not move the start, but qtparted moved the partition start without moving the data, rendering the OS unable to boot.
I know the previous poster mentions that most problem are due to repartitioning, but I disagree that that is irrelevant. A working tool must take these sorts of things into account.