[NBLUG/talk] More hard drive problems

Lincoln Peters sampln at sbcglobal.net
Mon Oct 3 18:18:52 PDT 2005


On Monday 03 October 2005 01:39 pm, Kyle Rankin wrote:
> Unless I'm misunderstanding you, that's basically what I meant by saying
> you need 2 new drives--one to replace your failed drive, and a new one.

I think I get it.  And, yes, I was thinking the same thing as you; I just 
didn't realize it was the same thing.

However, I ran Maxtor's PowerMax utility on the hard disk earlier today, and 
it was able to somehow repair the problem (!).  I'm not exactly sure how, but 
I'm going to see if the disk is usable again (I'll run one more surface scan 
while running Linux, rebuild the partition table, and reattach the disk to 
the RAID array).


So, in summary, what happened was:

* The hard disk failed, taking the partition table and swap partition with it.  
My computer apparently crashed when it tried to use the damaged swap space (I 
say "apparently" because the monitor, keyboard, and mouse froze up, but there 
was still a flurry of disk activity until I finally decided to hit the 
"reset" button 15 minutes later).  Needless to say, I am now skeptical of 
this hard disk, and I will NOT use it for swap space.

* When the computer restarted, the RAID array housing my /home directory 
failed to start because a drive was missing.  At first it looked like there 
might have been filesystem errors due to the crash (which shouldn't have been 
possible), but those errors resulted from the array being offline (and 
reiserfsck wasn't smart enough to figure out that it was trying to scan an 
inactive RAID array).  Ultimately, I only needed to figure out which mdadm 
commands to issue in order to start the array using the drives that did work.

* Since Maxtor won't accept a warranty exchange without an error code from 
their PowerMax utility to tell them the precise nature of the error, I 
modified /etc/mdadm/mdadm.conf so that the array would start automatically 
without the defective hard disk (I saved a copy of the original so I can 
easily switch back), then shut down Linux and booted from the PowerMax disk.  
After doing an "Advanced scan", PowerMax said that it could attempt to repair 
the damage (I think it just re-mapped the damaged sectors so that they'd be 
invisible to the OS; thus the usable space on the disk will now be somewhat 
smaller, and I won't be surprised if more bad sectors appear).  The data on 
the bad sectors was unrecoverable, but since the entire contents of the disk 
amounted to a swap partition and part of a RAID5 array, that didn't matter.


This may sound kind of demented (and perhaps it is), but I'm now kind of 
hoping that the disk will fail completely before the end of the year.  Why?  
Because I'm sure that it will fail eventually, and I purchased this hard 
drive on January (can't remember the exact date), and if it fails within a 
year of the date of purchase, I can get it replaced for free (and my use of 
RAID5 negates any chance that its future failure will cause any actual data 
loss).  Since this particular problem was fixable, I don't think I have a 
case for a warranty replacement at this time.


-- 
Lincoln Peters
<sampln at sbcglobal.net>

Shah, shah!  Ayatollah you so!



More information about the talk mailing list