I'm aiming at a very similar setup as what you're talking about. Whether or
not you put your root on the raid is up to you. Personally, I don't, and I
have my raid mounted as /home. Anyways, your mobo should support two drives,
and each additional controller should support two drives. Thus, with one
addon controller, you can have 4 drives. With 200GB drives, that's 600GB of
space doing RAID 5. In my experience, adding more drives to an array is
extremely easy, but it is dangerous. It is possible to lose data even if you
do follow the steps verbatim, although I've done it 3-4 times myself with no
trouble. First, you must use raidreconf to grow the array onto the new disk,
then you must use resize2fs to expand the filesystem to fill the array.
Backing up during this process is highly recommended!
I personally don't use a hot spare, and I haven't had any trouble with drive
failure since I got rid of all my IBM drives. I did have a run in with a WD
drive that needed a firmware update else it dropped out of the array
periodically, though. A hot spare is something to take into consideration.
Also, you might look into some sort of monitoring script. If you don't
monitor the array, you could lose one disk and go months without ever
knowing. Something that would email you would be ideal.
you can setup mdadm in the monitor mode and configure
it to send email alerts.
--- TJ Harrell <systemloc@earthlink.net> wrote:
I'm aiming at a very similar setup as what you're
talking about. Whether or
not you put your root on the raid is up to you.
Personally, I don't, and I
have my raid mounted as /home. Anyways, your mobo
should support two drives,
and each additional controller should support two
drives. Thus, with one
addon controller, you can have 4 drives. With 200GB
drives, that's 600GB of
space doing RAID 5. In my experience, adding more
drives to an array is
extremely easy, but it is dangerous. It is possible
to lose data even if you
do follow the steps verbatim, although I've done it
3-4 times myself with no
trouble. First, you must use raidreconf to grow the
array onto the new disk,
then you must use resize2fs to expand the filesystem
to fill the array.
Backing up during this process is highly
recommended!
I personally don't use a hot spare, and I haven't
had any trouble with drive
failure since I got rid of all my IBM drives. I did
have a run in with a WD
drive that needed a firmware update else it dropped
out of the array
periodically, though. A hot spare is something to
take into consideration.
Also, you might look into some sort of monitoring
script. If you don't
monitor the array, you could lose one disk and go
months without ever
knowing. Something that would email you would be
ideal.
-
To unsubscribe from this list: send the line
"unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at
From: Lucas Albers <hidden> Date: 2004-06-02 01:34:37
I have a system that uses drives on all channels, eg master/slave.
and I have encountered any slowdown, by using primary and secondary disks
in the same channel.
Yes it is slower, but for me performance is good enough.
It serves upwards of 100gigs of data over the network per day.
I use mdadm on all my raid systems, as it was easier to setup and has
automatic email notification.
I wrote a document on using mdadm. http://rootraiddoc.alioth.debian.org
I'm trying to put hot spares in my systems because I'm getting perhaps 1
disk failed every 2 weeks to a month.
It does seem a waste of space, but hard drives are cheap.
--
Luke Computer Science System Administrator
Security Administrator,College of Engineering
Montana State University-Bozeman,Montana
I have a system that uses drives on all channels, eg master/slave.
and I have encountered any slowdown, by using primary and secondary disks
in the same channel.
Yes it is slower, but for me performance is good enough.
With two drives per channel, doesn't a disk failure pull down the other
drive, causing the entire array to fail, and require a reboot to bring the
degraded array back online?
From: dean gaudet <hidden> Date: 2004-06-02 02:18:02
On Tue, 1 Jun 2004, TJ Harrell wrote:
quoted
I have a system that uses drives on all channels, eg master/slave.
and I have encountered any slowdown, by using primary and secondary disks
in the same channel.
Yes it is slower, but for me performance is good enough.
With two drives per channel, doesn't a disk failure pull down the other
drive, causing the entire array to fail, and require a reboot to bring the
degraded array back online?
yes that is an additional failure mode possible when you use two ata
drives on a channel. this isn't really a recommended configuration.
you might have to dork with raid superblocks to recover from such a
failure. in fact chances are you *will* have to dork with raid
superblocks to recover.
-dean