Snap Server 1100 not accessible

[pjclifford]

I have a a fairly new (about 4 months old)Snap Server 1100 250G. It has been behaving strangely the past several weeks, so I decided to do a "Check Disk" to see if there were any problems there.

Now, the disk status seems to be stuck at "Checking. 5% complete." There is an article about this in the Adaptec Knowledge base (Article). However, all that tells me is to let the disk checking finish.

The problem is that the checking has been running for 2 days. The disk is not accessible from the network nor from the assist.exe program. I can seem to get to it via http:\\snapxxxxxx\config in a web browser. However, I can not restart it, I can not shut it down, I can't reformat it -- basically, I can't do anything with it at all.

I can't even shut it down using the power switch on the back. In desparation, I pulled the power plug (data is backed up, so it is not important to be to salvage it on this device). System started up ok, but still seems to be in "disk checking" mode. I tried to reset back to the factory defaults, and that did not help the situation any.

Should checking the disk take this long to complete? Is there any way I can stop it? Is there some way to totally reset the device back to "as-shipped" condition?

I am pulling my hair out here. Hopefully someone can help.

Thanks

PJ Clifford

[blue68f100]

The 1100 stores the OS on the harddisk. If assist can not see it, that means trouble.

Access through the web interface was proably just a cache file.

The reset procedure is posted in the wiki section of the forum.

Are any of the led blinking? The snap flash error codes to these leds.

This is from the help memu for file system check.

fsck dev [[/fix /fixfatal /altsb]] (check device's filesystem)

where dev = 10000 for your drive.

If assist can not see the hd since it was powered down and restarted, I suspect the hd has failed. If this is the case, you will need to reload the os from a image file.

Being a 150gig hd, I assume you were running either 3.4.805 or v4 OS.

[pjclifford]

I am fairly confident that the web access is not from cache (I cleared the cache, plus I am able to access it from another workstation's web browser that has never accessed it before).

Additionally, I can ping the unit successfully.

The problem seems to be that the disk checking, in spite of Adaptec's knowledge base article to the contrary, IS indeed stuck at 5%. Because the disk checking is still deemed to be active (whether it is or not), I can not shut down the unit, nor restart it.

The LED seem to show normalish activity on the unit. The SYSTEM LED is flashing twice, pause, flash twice, etc. Supposedly, this is normal for a unit that is in a disk check.

I guess what I am looking for is a way to kill the disk checking process so that I can do a graceful restart of the unit. Just pulling the plug and powering it back on seems to restart it in disk checking mode.

[blue68f100]

The disk led should be solid during a fsck.

Do you have access to the debug page, enter "co de fsck 10000 /fixfatal"

[jontz]

Any chance that you can pull the drive, copy the first 40MB to another disk using DD from the knoppix CD and install the new disk in the snap? Just as a test, I mean. Perhaps you have an old smaller hard drive laying around that you could use...

I'm just wondering if it will act normal with a different drive...

[blue68f100]

These units have OS in flashram.

You do not need to copy any thing. Just wipe the MBR with Zero's. reinstall and let the unit format and install the OS. You may want to remove all drives but one.

[jontz]

Earlier in the thread:

Originally Posted by blue68f100

The 1100 stores the OS on the harddisk

Just above:

Originally Posted by blue68f100

These units have OS in flashram.

Are you just messing with me to make me look bad or what?

[blue68f100]

No, there are 2 version of the 1100.

The way things are acting, indicates he may have a v1 (OS in flashram). If that is the case copying the boot tracks to another drive JUST MIGHT WORK.

[jontz]

I'm just giving you a rough time...because I can

[pjclifford]

Sorry I haven't gotten back to this. We have had a couple of minor fires going on here the last few days that have diverted my attention from this.

Anyway, I just tried the command line that you have me and got the following as a result:

08/01/2006 10:09:05 Command: co de fsck 10000 /fixfatal

Failed to check device 10000 - internal error <1>
------------
Command returned with console error 1.

Also, looking at the Knowledge Base Article Disk status seems to be stuck at “Checking. 5% complete”, it says:

To confirm if the disk check is still in progress, refresh the page or confirm with the LED status. The system LED should double flash while the disk activity status should be lit.

It APPEARS from all indications to be doing a disk check, but it never completes it.

I don't think I am looking to tear into the unit quite yet, as I believe it is still under warrentee and I want to preserve that in case I have to send it back. I would like to find some way of killing the disk checking process and possibly reformat the drive and start again at factory defaults.

For additional information, the version page states:

Model 	        Software        Hardware      Server # 	   BIOS
1000 series 	4.0.860 (US) 	2.0.2 	      20xxxxx 	   4.0.855

[blue68f100]

I don't know of a cmd to stop it. There are some options not on the cmd list, /force. But do not know if "/abort" is a option.

I hope you have a backup. Because it may not be recoverable, if it's a drive faliure.

[Gewehr98]

I've just experience the EXACT same problem with my 40Gb Snap Server 1100, right down to the LEDs displaying an internal disk check, the console reporting "disk checking at 5%" for a day and a half, and the error <1> code after entering "co de fsck 10000 /fixfatal" from the debug console.

Is there a server number range in the 1100 series that delineates when they switched over from the OS being stored on chip vs. the boot sector of the HD? This particular 1100 is Server #16757x, earlier than Pjclifford's example above.

I have another 40Gb 1100 that's very close in Server # range, and although it loses data, it'll still boot. I have a spare 80Gb Maxtor DiamondMax HD, and although I'm unfamiliar with the Knoppix/DD trick, I guess that's my best bet, if the OS isn't on chip, right?

Are there any folks recommended who can do the HD image swap?

Pjclifford, if you're still lurking around here since last year, how did your saga end up?

[blue68f100]

Most all 1100's were boot sector. There was a mix of 1000's that were. Since the SnapOS requires verly little HD space, its possiable that you are loaded disk cache, or the boot sectors are still in tact. The area were it is crashing is probably the sector table area. I notice a lot of HD when they fail is in the same 5% area no matter what size HD it is. At least with the snap os.

[bitor]

Hey D:

Just reading the post here. This unit is a 1100 not a 1000. I thought only(according to what you told me) the 1000's had a version 1 and 2. Just wondering.

bitor

P.S. I have figured out how to make any 1000 handle two hard drives with no problems. You can even do a mirror, but takes a long time to build.

Originally Posted by blue68f100

No, there are 2 version of the 1100.

The way things are acting, indicates he may have a v1 (OS in flashram). If that is the case copying the boot tracks to another drive JUST MIGHT WORK.