hard drive replacement questions

[Henry Yen]

Sorry I've been preoccupied, so here are some additional thoughts on this
thread.

1. Badblocks only finding 13 bad blocks is a good thing. Hopefully that
   means you might have just a single bad spot on the disk, and there
   isn't a huge amount of damage. On the other hand, those middle-age
   IDE drives probably didn't do as good a job at remapping bad sectors,
   and they had fewer of them, so there may still be damage that either
   badblocks can't see, or all of the remapping spare sectors are all
   used up, which would mean many bad sectors had already been detected,
   and the remapping might not have been completely successful on some
   of them.
2. A command everyone should know is: dmesg
   That command prints out all of your boot-up messages. The identity
   of drives will either be in lines beginning with "hda:" "hdb:" (etc.)
   for IDE drives, and lines containing "Vendor:" and/or "Model:" for
   scsi/sata drives.
3. The "fsck -vcck" command (as well as use of "badblocks", by the way),
   unfortunately puts further stress on the disk. "fsck" is not 100%
   reliable to try and fix things in the face of hardware errors.
   It was designed primarily to try and recover from logical/filesystem
   inconsistencies if the system crashed or power-failed.
4. eBay and MicroCenter are two sources I've used with good success to buy
   almost any computer/electronics components. A 60GB IDE laptop drive
   runs about $40, and a 120GB drive runs about $60. You might consider
   going to a Solid-State drive ("SSD") instead (not recommended if your
   Linux kernel is more than about 18 months old), as modern SSD's run at
   about half the power consumption of spinning disks while providing
   about twice the performance. SSD's in the 40GB-ish range are at a sweet spot
   now-a-days with pricing for the SSD in an IDE laptop kit to be around $50.
5. Make sure you get a replacement drive that is at least as large as your
   old drive. You should be able to clone your old drive completely to your
   new drive using ddrescue or dd-rescue; these cousin commands are based on
   "dd", but continue on in the face of errors, and use strategies like
   copying from the end towards the beginning, and in chunks, to absolutely
   maximize the amount of data copied. After doing that, your new drive
   will be exactly the same as your old drive (including the logical damage,
   which will at least no be afflicted with physical error problems), and
   you should be able to even boot both systems just as before.
   You can also fsck the new drive, as it won't get confused by
   hardware failures any more.
6. Actually copying the drive could be tricky for you, depending on
   what hardware you have available. A USB drive converter is the least
   common denominator, but it's deathly slow. A separate "work space"
   system on your local network would be better.
7. For future reference, there are various partition cloning/backup
   systems for Linux that support Linux and Windows partitions, and are
   very fast (mostly because only blocks that are actually being used are
   copies, and blocks in the free list are not).

-- 
Henry Yen                                       Aegis Information Systems, Inc.
Senior Systems Programmer                       Hicksville, New York