on a new usb drive, after lots of badblocks iterations to verify the drive, i created two luks partitions identically (default options, same size, etc.), and stuck ext3 on both. the only difference is that i set a larger inode ratio on the second. both partitions worked perfectly for several days, (the first partition still works perfectly). i rsynced lots of data to both. then i rsynced some data to the second partition, and it failed with no space left on device. however, df and common sense showed that there was more than 100GB free. this persisted after doing a umount then a mount. then i did a luksclose on it, but luksopen failed saying that it is not a luks partition. isluks and luksdump likewise. the first partition opens and closes fine. i never treated the partitions differently that i know of and did nothing fancy (not even lvm). i was pretty careful. do i have any chance of recovering the data? what are the commands to overwrite the luks header with the header from the working luks partition, or to check to see if the header is corrupted? also, what should i trust the most or least: dm, dmcrypt, luks, ext3, usb, or the hard drive? could the fact that the hd goes to standby have affected anything? kernel is 2.6.18. it has both aes and aes-586 built in because i didn't know which one i needed for athlon mp (smp). i am now compiling 2.6.21 in hopes of being able to read the sucker. thanks. p.s. i am posting via gmane, so please cc: if you can. --------------------------------------------------------------------- dm-crypt mailing list - http://www.saout.de/misc/dm-crypt/ To unsubscribe, e-mail: dm-crypt-unsubscribe@xxxxxxxx For additional commands, e-mail: dm-crypt-help@xxxxxxxx
