Rebuild Snap 2200 with Raid 0

[swandive]

Hi there

Ive been reading though all the posts on this fourm looking for answers. Im glad i found this site as it has been very useful so far.

A few questions i have though. Hoping everyone can help me out.

I have a Snap 2200 (2x250gig) that was setup as a Raid 0. One of the HD went bad over the weekend but i want to see what i can do with the server. I realize that my data is lost. Fortunaly i have a good backup.

I would like to rebuild the server using 2x300 gb drives but not quite sure about a few things. I would like to put 2 new HD's in the unit to start fresh. Im not sure which one of the 250's went bad.

From what i have seen on this forum the OS for the 2200 is stored on the HD's themselves? If i replace both drives will i have to reinstall the OS as well?

Also how do i find out the Firmware version and other information about my current setup. I still have access to the Firmware but couldnt find what i was looking for

[blue68f100]

Good for you for having a BU. Running Raid0 (Stripped) has no redunducy. If the unit still boots, look at the error log. The top disk is 10000, with the lower 10010. The Strip is 40000. Error will show on 40000 and on one drive.

Data is not recoverable from a RAID 0 failure.

The 2200 is advertised as bootable from ether drive, but I think this only applies to RAID 1.

The OS is stored on the HD. The easiest way to move up to 300 gig drives, would to copy the boot tracks from your good drive to your new ones. This is best done with dd from the Knoppix CD.

To find out what your current OS, HW ... version is, clink on the SnapAppliance Logo from the admin web page. A popup window will apear with the info.

[swandive]

Thanks blue. I figured as much about recovering from Raid 0. I didnt set this server up to begin with and always figured it would come back to haunt me. Thank God for backups. Always worth the time and effort when something goes wrong.

I saw some instructions on the wiki on how to get the img off. Is that the best way to do it? Let me know what you think

swandive

[swandive]

Here is an update to the specs of my snap appliance
Model Software Hardware BIOS
2000 series 4.0.830 (US) 2.0.4 3.4.758

4.0.860 is the latest version?

[blue68f100]

That the best way we found, and is univeral between hardware. Since the hd format is a XFS file system, most have a problem reading it correctly. You can go from hd to hd. Its just a matter of change the in=file to in=/dev/hda? Just don't get it crossed and whip out your good drives OS.

If you were lucky enough for the 10010 hd to fail. You may be able to knock it back to JBOD. Replace the failed drive then build a RAID 1. This would put the OS on the new drive. Powerdown, Replace the old with a new. Reboot. It will will copy the OS over to it. The 2200 some times does not auto repair the raid disk (50000). With no data no problem. Since the drive capacities were different, JBOD, Then back to your final setting (raid1?).

Time wise, since formating for raid takes a long time. Copy the boot tracks to both drive and just install them. Then issue the "co de format 100X0 /reinit" to each drive.

[swandive]

Blue

Thanks again for your help

Let me see if i read your last post correctly..... It sounds like the OS needs to be installed on both drives?

So if the 10010 drive is busted then i can put in a new 300 gb drive. that will turn it into a jbod configuration. It will then copy the os to both drives. once that is done i can remove the good 250 drive and put the other new 300 in and it will do the same thing again. Finally i can set it up as a raid 1 which is where i want it to end up.

Now if the 100000 drive is pooched i have to use knoppix and dd command to get the boot files from the 10010 drive and copy them to both of the new 300s Then build the raid from there.

Am i correct for both situations here? Please let me know if im off course here.

Thanks again

[blue68f100]

It needs to be only on the 10000. TOP drive.

I was attempting to explain how to recover with out using the KnoppixCD.

If 10010, bottom drive failure.
Installing a new drive (10010, bottom), Snap will format it. Breaking RAID to JBOD. then Changing to RAID 1 will copy the OS to the new drive. Then you can swap out the top drive.

Let wait and see which drive has failed.

[swandive]

Ah that makes sence. Sorry for the confusion

So im trying to figure out what disk is fried. This is what my disk log is full of
File System Check : Cannot Read: Blk 13533888 Disk 40000 3/27/2006 8:07:24 PM

It has been trying to do a disk scan for about 3 days now. The Block number has been a few differnt blocks but this seems to be the one its stuck on. But from this i cant tell which disk it is. Where else could i look?

[blue68f100]

40000 is the raid set, not the specific drive. Your looking for a 10000 or 10010.

Unplug one of the drives and see if it boots. If it does. Replace the other and see what happens.

[swandive]

Ok now im getting confused. I unplugged both drives seperatly and booted up. Both times i was able to get into the firmware with no problems. When i checked the disk status of course i got some errors saying that the disk was broken

W File System : Set-up of logical disk failed. Device ID: 40000 (hex) Disk 40000 3/28/2006 9:04:22 AM
W File System : 1 member(s) missing in logical device (original ID: 40000) Disk 40000 3/28/2006 9:04:22 AM
E File System : Logical set member 0 not found. Original device ID: 40000 Disk 40000 3/28/2006 9:04:22 AM

These are from the error when disk 0 isnt plugged in. I get the same errors when disk 1 isnt plugged in.
I did try unplugging both drives and booting up but couldnt connect at all.

The other thing i noticed is the system led. When disk 0 was only plugged in the system led blinked with a double blink. Like it has been doing for the last 3 days as the system check has been running. When disk 1 was plugged in alone the system led blinked just once at a time. Dont know if that means anything

Im still pretty sure that one of the disks is bad because of the constant Cannot Read BLK xxxxxxxxx errors that im getting. My guess is that the boot sectors for both of the drives is ok and i could take the boot image off either one to copy to the new drives.

Anywhere else i should be looking? any thoughts? Is there anyway else to load the OS onto new drives? you're my boy blue!

[blue68f100]

You have a 2200, you just confirmed something for me. The 2200 is advertised as being able to boot from either drive. If you can get to the disk utility from the web interface, it may tell you which drive. Repair is easy. Just install a new drive into disk 1. And see what happens. Worst case it want boot. If it doesn't boot move it to the other position. Since you can still connect, you can monitor what happening from the Disk Utility section. One the drive is formated, and ready. Power down and install the other new drive. Once it boots off of the top drive (10000), you can safely change the drive configuration. I don't think it will allow you to change it to a raid 1 with one drive bad. But try, if no go change to JBOD.

Since it is bootable you will not have to copy the boot tracks manually.

[swandive]

Sweet. Thanks for your help blue. Im picking up the new HD's this afternoon. Will keep you posted.

[swandive]

Hey Blue. Just wanted to let you know that everything worked like a charm.

Stuck one new drive in and it did a bunch of formating and check etc then added the other new drive and the same thing happened. Then it was build the RAID and now im in business.

But now im thinking i should have used 500 gig drives rather than the 300s. DO you know if that would work?

Thanks again for your help

Swandive

[blue68f100]

I havn't heard of anybody installing that large. A few have thought about, but most have stopped in the 250 - 300 range, lower cost/meg . 500 gig I think is the largest Snap/Adaptec list.