What is my 4100 doing!

[bug1124]

Hey guys, yesterday I started messing with a questionable snap server 4100 where I work. When I first got here it needed to be rebooted, after which it seemed to be functioning correctly. I went ahead and ran 'Check and Repair Disk' utility within the interface using the 'Repair most errors (severe errors will not be repaired)' option. It ended up finding an error and now it will no longer bring the drive online. Can anyone look at my log and tell me what is wrong with this thing?

Server Log

These are the specs of my 4100:

Product Name - Snap Server 4100, 120GB, Refurb
Software - 3.1.618 (US)
Hardware - 2.2.1
Serial# - 502391
BIOS - 2.4.437

[Hallis]

Very nice logs. I'll look at them in detail here in a minute.

Shane

[Hallis]

At first glance i see a few kernel panics. that is not good, It usually points to a hardware problem in the CPU or RAM or something of the like. it COULD be a corrupt OS but i wouldnt count on it. I couldnt find anything specific o nteh FSCK error 39 at first glance but i'll look deeper into it. The Kernel Panic is the more important problem and could have caused the other problems. If any of your data is available then BACK IT UP NOW!.

[Hallis]

And the SMART hdd warnings probably indicate a failing drive.

[bug1124]

Ugh. I knew I shouldn't have gotten involved with this thing. None of the data is available since the disk won't remount on the network. The data is pretty important so I think that is my first priority. What should I do to back the unit up? And once its backed up, where to go from there?

Thanks for looking at this, btw.

[Hallis]

Well, i'll defer to David for this one as far as trying to get the data back because he's got more experience with raids then i do. This is the reason im scared to do a raid5 unless i have a hot spare.

Shane

[frankb3910]

Can you see the data using the web browser? Log in using snap assist (web browser) to the admin page, hit "close" and then you should see your shares listed on the next page that comes up. At least this would confirm that the data is intact. If you don't see it here, you are probably out of luck?

What does snap assist report under disc status now?

[blue68f100]

Ok Looking over the log file, you were having a drive problem 6 mo ago. You should have a drive led amber indicating a bad drive, 10006 is drive 1.

I File System : Device 60000: file system below free threshold (10%) System 1/20/2006 11:41:25 AM W File System : Device 60000: file system above free threshold (10%) System 1/19/2006 3:19:28 PM

You have filled the unit up, Low or out of drive space. No room for the OS to work. NOT GOOD.

First Kill all access if you have not done so.

Then locate a replacement drive, and replace the one indicating on the front panel. By the log early on, Drive 1 was failing, djontz and phoenix32 can confirm.

If you can get access to debug issue "co de info" it should show which drive a long with id numbers.

If the unit is not accasable you can remove the drives (1 at a time, or lable), then use the mfg utility to access the SMART data. You will know when you find the bad one.

DO NOT MOUNT THE DRIVE IN LINUX OR WINDOWS ENVIROMENT, or ALL DATA WILL BE LOST. The snap OS is built off FreeBSD which is unix based.

If you have a copy of SpinRite v6 it is not os dependent, so it is safe to use. It will not beable to do recovery if all of the smart's spare sectors are used up.

[bug1124]

Originally Posted by frankb3910

Can you see the data using the web browser? Log in using snap assist (web browser) to the admin page, hit "close" and then you should see your shares listed on the next page that comes up. At least this would confirm that the data is intact. If you don't see it here, you are probably out of luck?

What does snap assist report under disc status now?

This is what I see when I just go to the main web page. And this is what I see under disk status.

Originally Posted by blue68f100

Ok Looking over the log file, you were having a drive problem 6 mo ago. You should have a drive led amber indicating a bad drive, 10006 is drive 1.

Quote:
I File System : Device 60000: file system below free threshold (10%) System 1/20/2006 11:41:25 AM W File System : Device 60000: file system above free threshold (10%) System 1/19/2006 3:19:28 PM
You have filled the unit up, Low or out of drive space. No room for the OS to work. NOT GOOD.

First Kill all access if you have not done so.

Then locate a replacement drive, and replace the one indicating on the front panel. By the log early on, Drive 1 was failing, djontz and phoenix32 can confirm.

If you can get access to debug issue "co de info" it should show which drive a long with id numbers.

If the unit is not accasable you can remove the drives (1 at a time, or lable), then use the mfg utility to access the SMART data. You will know when you find the bad one.

DO NOT MOUNT THE DRIVE IN LINUX OR WINDOWS ENVIROMENT, or ALL DATA WILL BE LOST. The snap OS is built off FreeBSD which is unix based.

If you have a copy of SpinRite v6 it is not os dependent, so it is safe to use. It will not beable to do recovery if all of the smart's spare sectors are used up.

There were no failed drive indicator lights lit when I first started working on it to resolve an issue regarding why it was only able to ftp files. Currently the 'System' light is slowly flashing, the 'Link' light is solid and no other lights are illuminated.

The unit has the Quantum name on the front so I imagine it has 4 quantum drives inside. What would qualify as a suitable replacement?

After talking to the other people here they have told me that space was frequently a problem and have tried to aviod filling it up. However, apparently no one has looked in the control panel for a long time so they never knew how much space there was to work with. When I checked into it yesterday (my first time) it indicated ~3GB free.

What I am wondering is how this thing was working yesterday, then after I ran check disk it now won't boot.

[bug1124]

So I guess the verdict is that I need to replace drive one? We were also considering just buying four new 120 or 160 gig drives and upgrading it as far as possible, but I guess now data recovery is the first goal.

Thanks alot for the help everyone...I am trying to do this in between my normal job here. I may just have to take the snap server home to get anything done with it.

Anyone live/work near Addison?

[frankb3910]

What you show under "Disk Status" does not look good. I have however seen the list of shares with "X" through them as you have showing the shares "off line" Here is what I did to fix that, I am not sure it would work under your conditions:

On the main web browser screen, hit Security
Then hit Network Shares
You'll see a list of all the shares, click on one of them, and then click properties.
Then click "OK".

I was able to do this and my shares became available.

As far as putting in 160's - good luck. See my other post under "Help with 4100 Craked Raid" - I have been experimenting with doing the same thing and it has not been a success.

[blue68f100]

Open the unit and record the the drive info. You will need a drive that the usable capacity is = to or > than but not more the 2%. Drive 1 is useded to calculate the RAID 5 capacity. SO try to get the same drive with the same firmware.

I would wait till djontz and phoenix32 looks at the load and confirm my findings. It is posiable that you have a mix of drives. With drive 1 being criticle to the units health.

As far as upgrading stick with 120's

bug i'm in plano if you havn't noticed.

[Phoenix32]

Originally Posted by blue68f100

I would wait till djontz and phoenix32 looks at the load and confirm my findings. It is posiable that you have a mix of drives. With drive 1 being criticle to the units health.

I looked the log over and I think Dave is right here. It is possible that sectors just got corrupted, but from what I can see, it looks like a failed/failing hard disk. Add in the fact of how full the unit is, this just magnifies the problem.

[rpmurray]

Wow, this log has it all. General Protection Faults, SMART warnings, damaged inodes, FSCK error.

Those Panics that go all the way back to January and the SMART warnings that started in August all seem to point to a system on it's way to failure. I'm kind of curious about the "Device 0x10004 SMART warning" since I can't find a physical or logical device ID that matches 10004. Anyone know which drive it's pointing to?

bug1124, could you do an "in dev" and then a "in log t" from the command line and post the results of both?. Should be interesting.

[blue68f100]

That why I'm waiting for Phonnix32 or djontz to verify. By their number scheme it should be drive 1. 10008 is were drive 2 starts.

If you don't have these units sending email alerts, it's posiable to get to a point of no return. That what I'm working on right now. Need a Mail server that does not require Auth.

[bug1124]

I opened the machine up and took several pictures of the insides and number plates. Below are those pictures. I have also managed to track down a replacement drive (LM30A011-01-A). Let me know if I should go ahead and order it.

Image1
Image2
Image3
Image4
Image5
Image6
Image7
Image8
Image9
Image10
Image11
Image12

[bug1124]

Originally Posted by rpmurray

bug1124, could you do an "in dev" and then a "in log t" from the command line and post the results of both?. Should be interesting.

How do I get a command line on the unit, or are you referring to windows? Do I need to find some utility for interacting with this unit (i.e. telnet or other) If you can you give me a step by step for what you want I will do it and post the results.

From what I can tell everyone seems to think disk 1 is dead or on its way out. Should I buy a replacement now?

You guys are a life saver...

[blue68f100]

Bug you need to clean the dust/dirt out of the cpu fan. I would advise of doing it out side, Bad as it is, it's going to create a Dust Storm.

To get to the debug page that allow cmd line add "/debug" to the address bar ( http://snapIPaddress/admin/debug).

The in dev should report a drive as bad.

[bug1124]

haha how did I know someone was going to say that about the dust.

Below are the results of the commands:

in dev
in log t

I have also updated the Server Log with this past boot since I had its cover off.

[Phoenix32]

Originally Posted by blue68f100

That why I'm waiting for Phonnix32 or djontz to verify. By their number scheme it should be drive 1. 10008 is were drive 2 starts.

I already did, look up...


Anywise, with the new logs, it looks more and more like a failed drive or interface (probably drive). And to add to it, this thing has been in degraded mode so long, I suspect corruption to the array as well. You know what that means....

Some of the numbers don't make sense, but it does look like drive 1 is the culprit.

[bug1124]

Any ideas on why this booted yesterday and all the files were accessible then after running the disk check program everything quit working?

I don't suppose anyone has a Quantum LM30A011-01-A here locally that they would like to sell me...

When I get the replacement disk what process will I need to go through to initialize the restore process?

[blue68f100]

Your unit is only registering 64 meg of ram. Do you have a 128 or 256 you can drop in?

File=/private/clients/1014_Weber/Weber info/prohelp/html/usascii/domains/manufacturing/

When you manualy mount the shares to remove data. Save this one for last. And only do it when the others are not mounted.

I think you biggest problem for not mounting may be out of disk space. As you backup data, you may want to clear non-criticle data. You need to clear a minimum of 3 gig. This is needed cache space.

The device info did not show any drives as failed, but according to the smart data it must be close. Minimise the times you start the unit up.

Do you have a pc you can move the drives over to read the SMART data? You should be able to locate the bad one. My best guess is drive 1.

[blue68f100]

Originally Posted by bug1124

Any ideas on why this booted yesterday and all the files were accessible then after running the disk check program everything quit working?

You probably have a drive that is self destructing.

[bug1124]

Originally Posted by blue68f100

Your unit is only registering 64 meg of ram. Do you have a 128 or 256 you can drop in?

I can probably track down a stick of replacement memory. What are the specs? PC133, PC100? does it matter?

Originally Posted by blue68f100

When you manualy mount the shares to remove data. Save this one for last. And only do it when the others are not mounted.

How do you manually mount shares? Will it allow that in its current state? It would be huge if I can access the data. How would I clear space on it in its current state? What if I take this to a data recovery place...or is this something I can do myself. I really just want to get the data off safely, then I can bump the RAM, upgrade the firmware, and put in four new 120GB hard drives.

[blue68f100]

Up grading the ram will help with caching, and will have no impact on your data, still safe. They use PC 100/133 ram NO eec, just std ram.

Try this by Frank

On the main web browser screen, hit Security
Then hit Network Shares
You'll see a list of all the shares, click on one of them, and then click properties.
Then click "OK".

But only mount one at a time.

If you take it to a recovery service you will be looking at upwards of $2-5k. Raid 5 systems are expensive to recover.

I do not think you are at the point yet.

I would go ahead and get the 120gig drives. We may beable to clone the drives to the 120's then reset the partition table to expand the drive. The live DD has cloning functions. But this will be our last resort. But with a drive failing you may want to go ahead and make the bin files. DD does not support writng to NTFS file structure. Since every thing is full you will need atleast 130gig of space.