Conversation

Notices

  1. oh, and btw, good morning !tzag !fediverse! I was distracted by the sight of fsck->badblocks testing the second partition on the 'bad' disk... report to follow soon!

    Sunday, 06-Jul-14 05:42:52 UTC from oracle.skilledtests.com
    1. @mk here we are: badblocks test (R-W) of second partition of the 'bad' drive, which sits smack above the area where ddrescue reported 360+ of the 369 bad blocks to be... now has ZERO bad blocks!! looks like my strategy of 'scrubbing' (repeatedly writing over teh whole surface) was successful after all (but I'll keep a close eye on it). On to the third partition...

      Sunday, 06-Jul-14 07:46:40 UTC from oracle.skilledtests.com
      1. @mk ... and the results for the third partition: the 1 bad block found by the last pure badblock test shows up here - so still 1 bad block as opposed to 369 initially! even better, it appeared within the first 1% of testing of this partition; so, I'm going to shrink it until the bad block "disappears" and then grow the second partition so it will have that block near the end, and then re-test that partition so it will be marked in the bad blocks inode. to shrink the last partition: remove it, re-create it smaller at then end of the disk, then test no more than 2% - if bb found, rinse and repeat, otherwise we're done.

        Monday, 07-Jul-14 04:57:00 UTC from oracle.skilledtests.com
        1. @mk It is done! the original 'bad' disk (369 bad blocks) now only has one. It now has three partitions: one *before* the original area with the bulk of bad sectors, now error-free; one "over" the original area with bad blocks, now with just one bad block (recorded); and one after the bad area, also error-free. Provided I regularly check for bad blocks (non-destructively), it's safe to use this drive again. # #, too :-) \o/

          Monday, 07-Jul-14 17:02:37 UTC from oracle.skilledtests.com
          1. @mk all that's left to do is write up how I 'designed' size & position of the partitions to minimize the chances of running into bad blocks popping up again (since they were strongly clustered originally, any new ones would likely be in the same area). But first: a !beer!

            Monday, 07-Jul-14 17:05:30 UTC from oracle.skilledtests.com