software raid drive failed, please provide step by step, to rebuild

[David G. Miller]

Dan Carl wrote:

>I know its a raid 0 is a stripe.
>Its my swap partition.
>Why would I need fault tolerance on my swap.
>
Things can get weird if swap is on a failed device or the device its on 
fails while the system is running.  I went the opposite route:

[root at fraud ~]# df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/md5               2877756    184776   2546796   7% /
/dev/md3                248783     12014    223925   6% /boot
none                    387852         0    387852   0% /dev/shm
/dev/md2              37057024   2122108  33052508   7% /home
/dev/md0              10317752   3161284   6632356  33% /usr
/dev/md1              10317752   1177868   8615772  13% /var
/dev/hdh1              9621848        24   9133048   1% /stage
/dev/hdh2            147945308  60761372  79668732  44% /share

/dev/md4 is swap.  /stage is my amanda staging area and /share is for 
large, replaceable files (e.g., movies, ISO images, etc.)  They are 
mounted by rc.local so even if that drive fails, the system comes up. 

I tested by shutting the box down, pulling a ribbon cable off of a drive 
and then booting the system.  Lather, rinse, repeat.  Any one drive can 
die and the system still runs.  The system has also survived a drive 
failure.  Got my e-mail from mdadm and just had to figure out which 
drive in the RAID-1 pair had died.

Cheers,
Dave

-- 
Politics, n. Strife of interests masquerading as a contest of principles.
-- Ambrose Bierce