XFS hanging under Ubuntu?

[PorkCharSui]

Hi I just finshed building a software raid 5 setup using Ubuntu 5.10 and after configuring the array in EVMS, I issued the mkfs.xfs format command. The command completed and brought me back to the terminal, but I was left with a process that is seeminly eating CPU time / crunching the drives indefinately.

If I issue the command again I get this error message:

derek@rsjabber:/dev/evms$ sudo mkfs.xfs raid5
dm_task_set_name: Device /dev/evms/raid5 not found
Command failed
mkfs.xfs: raid5 appears to contain an existing filesystem (xfs).
mkfs.xfs: Use the -f option to force overwrite.
derek@rsjabber:/dev/evms$

Is this process just a result of formatting, I let it run yesterday for about 1.5 hours but it never ended so I'm worried that something is possibly amiss. I deleted the raid config and set it up again today after restarting, but it still is occuring. Maybe I'm just not patient enough

top - 22:26:39 up 50 min,  2 users,  load average: 2.13, 1.90, 1.49
Tasks:  97 total,   1 running,  96 sleeping,   0 stopped,   0 zombie
Cpu(s):  6.2% us, 22.0% sy,  0.5% ni, 69.6% id,  1.2% wa,  0.2% hi,  0.3% si
Mem:    516492k total,   375144k used,   141348k free,    13724k buffers
Swap:   755012k total,        0k used,   755012k free,   160212k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 7938 root      15   0     0    0    0 S 31.5  0.0   9:32.98 md0_raid5
 7939 root      15   0     0    0    0 D  3.9  0.0   1:04.66 md0_resync

FYI, the array is composed of 4x 200gb drives running on two IDE channels of the mobo.

Thanks

 

[unhappy_mage]

It's still rebuilding the raid 5 array. Do "cat /proc/mdstat" for a progress meter. When that finishes cpu usage should drop to nearly zero.

 

[PorkCharSui]

Awesome hint. The drive activity settled down by this morning and now I have 600 gb to fill.

 

[unhappy_mage]

Just out of curiosity, how much CPU does "md0_raid5" take while you write at full speed to the disk? Change directory to the place where the array's mounted, and run this: "dd if=/dev/zero of=delete.me &"
Then run "top" and look for md*_raid5. Here's my answer:
1083 root 15 0 0 0 0 S 7.0 0.0 261:24.49 md1_raid5
so 7% cpu usage. Then be sure to kill dd and remove "delete.me", or it'll just fill the disk: "killall dd; rm delete.me".

 

[PorkCharSui]

It was averaging around 3.3 %

I'm running a P4 2.4C