I have four software raid volumes consisting of four 400 GB sata disks. They're working fine except for one thing - the booting time. The booting procedure takes over two minutes! (from when I'm leaving LILO to when I'm totally logged in, I'm using automatic login). Out of these two minutes, half the time (63 seconds) are spent to do something with the biggest raid volume.

These are my raid volumes:
device     size    raid type  file system  mounting point
======     ====    =========  ===========  ==============
/dev/md0    94 MB  raid1      reiserfs     /boot
/dev/md1    22 GB  raid5      reiserfs     /
/dev/md2   3.9 GB  raid5      swap         swap
/dev/md3   1.1 TB  raid5      reiserfs     /files
During boot these lines are printed to screen:
Filesystem is clean
Replaying journal..
Reiserfs journal '/dev/md3' in blocks [18..8211]: 0 transactions replayed
Then the computer performs the check (or whatever it is) of /dev/md3 which takes 63 seconds. Then the following line is printed and the booting continues:
Checking internal tree..finished
This procedure is performed for all raid volumes except for the swap volume, but the time spent on the other volumes are negligible. So md3 is the only one taking time, but I guess that is because it is slightly bigger than the other ones

It seems stupid to require 2 minutes for a raid system to boot... So:

What's my computer doing during these 63 seconds?
Is it necessary?

EDIT: Never mind, the problem is solved. The reason was that I used reiserfs, so I switched to ext3 on /dev/md3. Now md3 is mounted instantly, so the booting time is reduced by a minute.