Hacking a SNAP Server 2200 - How to replace a failed drive
I had a Snap Server 2000 with 2x250GB hard drives in it. The
drives were set to mirror for data redundancy. Luckily, I did as
after about 18 months, one of the drives failed! Had this been
setup as a stripe, all the data would have been lost. (I know
all about drive failure so the SNAP Server was also backed up
from another workstation as another layer of redundancy.)
The drive that failed was the boot drive, but the SNAP Server
switched right over to the mirror drive and kept on running.
Fixing the SNAP Server 2200 was a lot more trouble than I
thought it would be! I assumed I would be able to simply swap
the bad drive with another good drive, and it would rebuild all
by itself. I was wrong! I did finally get it to working but it
took some tome and research. The steps below were what I had to
do to get the SNAP Server back up and running.
The first thing I had to do was to determine which drive was
bad. I simply took the SNAP Server apart and disconnected each
of the drives until the server wouldn't boot. It would have been
much easier to go into the web interface to see the disk status
to determine the failed drive. But when I did this, it said that
the boot disk, disk 0 had failed. Looking at the hardware gave
no indication which was the boot drive and which was the
secondary drive! As it turns out, disk 0, or the boot drive is
the top drive. The secondary drive is on the bottom.
I removed the bad drive and ordered an exact replica of the
failed drive. It is important for a mirror to have an exact
replica of the good drive. I ordered two of the drives right
away in case the good drive fails in the near future.
I put the new drive in the SNAP Server (both drives need to be
configured as "Master" IDE drives) and I powered it on. The SNAP
Server would not boot! It would enter reset mode (the power and
disk lights would blink in tandem) and go no further. I cold not
even communicate with the server. So I called Adaptec tech
support (Adaptec purchased SNAP.) and they told me there was no
way to do what I wanted to do and that the server needed to be
returned for repair! How untrue! What I discovered from research
is that the SNAP Server will boot off the mirrored drive (bottom
drive) but it won't rebuild a new drive from there! The SNAP
Server must boot and run the operating system from the master
drive (disk 0 or the top drive) in order to rebuild a mirror.
I put the new drive in the bottom slot and moved the good,
original bottom drive to the top slot. I powered on the SNAP
Server. Only then did I eliminate the blinking light, reset
mode. I accessed the SNAP Server via the web interface and
looked at the Disk Status page. It showed that I had two drives!
My boot drive as logical disk 10000 and the new blank disk was
logical disk 10010.
Once I had the backup mirror drive (the one with all my data)
on top (it was now drive 10000), the Snap booted up. Now it was
time to rebuild the mirror. However, disk status showed a
cracked mirror. I could not rebuild the mirror because it saw
the new drive as a separate drive and was still looking for the
rest of the mirror. So I had to break the mirror by making both
drives individual drives, then re-establish the mirror. This
started the rebuild with no loss of data.
The exact steps are below.
1. Put good drive in top slot of SNAP.
2. Put new drive in bottom drive of SNAP. Turn on SNAP.
3. Reformat the new drive by executing the following command
from the ipaddress/config/debug web page. (For example
http://192.168.1.1/config/debug) Note: DD is the number of the
problem hard disk. co de format 100DD /reinit In my case it was
co de format 10010 /reinit (After reboot, this configures 100DD
as a SINGLE drive and FORMATS it.)
4. Reboot the Snap server.
5. Now break the mirror with this command co de config
individual 10000 10010
6. Reboot.
7. After the orphaned disk is formatted and the drives are
converted to individual drives, execute this command from the
ipaddress/config/debug page. (For example
http://192.168.1.1/config/debug) co de config mirror 100SS 100DD
SS is the source disk, DD is the destination disk. In my case
the command was co de config mirror 10000 10010 (This step
mirrors the drive with data to the drive without data.)
5. Reboot the Snap. If everything works according to plan, the
View Disk Status page should report "Rebuilding backup disk"
when it comes back online.
Once I went through this entire process, the mirror was
re-established and all the data was rebuilt. The whole 250GB
drive took around 5-6 hours to rebuild. But now I have a SNAP
Server 2200 with a new drive. The server is mirrored so I can be
rest assured that my data is safe.