Sanp 4200 FSCK fatal error

[tta013]

Hi All,

First of all sorry I have posted the this topic to post on the following one because for some reasons, I can't start a new post at that time.

I have got a Snap Sever 4200 which is used for backup purpose.

Whilst doing the back up the other day, I saw a report saying that there is a media error.

When check the disk with disk utilitie, I have got a message saying
File System Check: FSCK fatal error = 39 Disk 1
File System Check: Bad state 0 for inode I = 4065440 Disk 1

When I try to access to the server via IP address,
I can see that all three folders are shown as *off line*. Therefore, I can't access it from the short cut drive. But I can still go into the administration mode and still do the disk utilities.

Please can someone help me out on this. How can I get the disk back to normal. The way I see is if the Disk 1 has got a problem, I should be able to save my data on other disks.

I would really appreciate if someone can get back to this post.

Thanks
Tin

[re3dyb0y]

What RAID format do you have your 4200 in, as this will be key to working out how to sort you out

If it is mirrored you should be fine,but if you cannot view the disk, it seems like it is more likely to be striped

Let us know please!

[tta013]

Thanks for the reply re3dyb0y. Probably sound silly, but I don't want to give the wrong answer. How do I find out the RAID Format on 4200?
Cheers

[tta013]

Hi again,

Here are some of the info I can extract but I am not sure that's what you asked.

info panic

Current boot number is 68
FATAL ERROR During prior boot: 57
PANIC : ffs_alloccg: map corrupted


04/12/2006 12:01:37 Command: info version

Product: 2 - "Spinnaker/Laser: 4 drive IDE" - Platform Bytes "2.2.1"
Server-3.1.603, Build Date: Apr 5 2001 18:33:31

fsck dev

Current fsck mode for device DE: -1 (RAM-based fsck, 90% of RAM used for i-node cache)

BACKUP - Unavailable: disk is off-line.
DATA - Unavailable: disk is off-line.
SHARE1 - Unavailable: disk is off-line.


Regards
Tin

[tta013]

I just run Disk Repair. Still having problems and got following messages.

E File System Check : FSCK fatal error = 39 Disk 1 12/4/2006 13:17:09
E File System Check : Bad state 0 for inode I=4581596 Disk 1 12/4/2006
W File System Check : Summary information bad (Salvaged) Disk 1 12/4/2006
W File System Check : Blk(s) missing in bit maps (Salvaged) Disk 1 12/4/2006
W File System Check : Free blk count(s) wrong in superblk (Salvaged)

I am quite new to the Snap and has totally no idea how to solve this.

I was thinking about formatting the drive but then I wasn't sure that's the silly approach.

The OS is Windows by the way.

I would appreciate if anyone can help me out on this.

Regards
Tin

[blue68f100]

Sounds like disk 1 has failed in raid5, with no spares.

You will need to replace disk 1. Setup as Hot spare. Then the snap should rebuild the array. Depending on the array size it may take 24 hrs to repair if there is a lot of data on it.

Raid 5 scatters data across all drive in the array. The other drives have parity info for rebuild a failed drive. The reason you can't read the data is 1 chunk is missing. The dada is not lost.

[tta013]

Hi blue68f100,

Thanks for getting back to me.
But I am a bit lost. You asked me to replace disk 1. Do you mean physically replace with a new drive ? To be honest, I don't understand what you mean by Setup as Hot spare. As I said, I am quite new to Snap but this is something I need to get on with. :P

So please, can you give me some instructions - "dummy guide" on how to solve this problem.

Regards
Tin

[webboy]

sounds to me like incompatable drives. I was getting the same error with a 4100. No matter what i did with the drives it wouldnt hold a drive configuration.

http://forums.procooling.com/vbb/showthread.php?t=12849

how long have you been using the device? and what kind of drives are in it?

[tta013]

The company has been using Snap/devices for nearly 2 years. I just got into this environment. I am guessing that the drives are Segates.

[blue68f100]

Snap Appliance used IBM/Hitachi drives depending on the age.

tta013: From the info you gave me, Drive 1 is having a problem. It should show a problem on the front panel.

Run the cmd "info devices" from bebug. It will give you a detail report of your drives.

In a raid5 config. All data is spread across the array. Each drive contains parity data on the other drives. So if one fails. The data is not lost. But not be accessable till a replacement drive is replaced and the array rebuilt.

IF YOU REPLACE THE WRONG DRIVE ALL DATA MAY BE LOST. So you need to know which drive is having the problem. One the drive is replaced and set as a HOT Spare the OS will use it to rebuild the array.

I think the 4200 has seperate controllers for each drives. Drive ID's are like this: 10000, 10007, 10010, 10017, where raid 0 = 40000, raid 1 = 50000, raid 5 = 60000. There are other drive id that reference to system directories.

As a rule of thumb on raid systems. All drives need to be the same size (with in 2-3%) and config ( head, cyl, cache ...).

I assume that you have Snap OS and NOT Guardian.

I hope this clear some things up for you. The wiki section has a link to Adaptec if you need the admin manual.

[re3dyb0y]

Mine had maxtor's

But that was quantum, i suppose, as they were part of the same group

[tta013]

Thanks for your replies.

Yes I am suing Snap OS not Guardian.

When I run info devices, I have got the following information:

Logical Device: 10006 Position: 0 JBOD Size (KB): 32296
Free (KB): 24440 Private Mounted

Label:Private Contains system files only
Unique Id: 0x5E58CC586E6D7C95 Mount: /priv Index: 0
Order: 0
Partition: 10006 Physical: 10007 FS Size (KB): 32768
Starting Blk: 515 Private

Physical: 10007 Drive Slot: 0 IDE Size (KB): 29316608
Fixed

Logical Device: 60000 Position: 1 RAID Size (KB): 0
Free (KB): 0 Public Unmounted

Label:(Unknown) Unable to access the information for this disk.
Unique Id: 0x295BDA6965E9EBFB

Partition: 10000 Physical: 10007 R 60000 Size (KB): 29067088
Starting Blk: 31062 Public
Physical: 10007 Drive Slot: 0 IDE Size (KB): 29316608
Fixed

Partition: 10008 Physical: 1000F R 60000 Size (KB): 29067088
Starting Blk: 31061 Public
Physical: 1000F Drive Slot: 1 IDE Size (KB): 29316608
Fixed

Partition: 10010 Physical: 10017 R 60000 Size (KB): 29067088
Starting Blk: 31061 Public
Physical: 10017 Drive Slot: 2 IDE Size (KB): 29316608
Fixed

Partition: 10018 Physical: 1001F R 60000 Size (KB): 29067088
Starting Blk: 31061 Public
Physical: 1001F Drive Slot: 3 IDE Size (KB): 29316608
Fixed

------------
Command executed without error.

If I understand correctly, the error is on the Disk 1. Am i right?

Do I just buy a new drive (any drive or same brand? silly question really but want to make sure I do it right) , open the box and replace with the Disk 1? Am i right?


Regards
tta

[blue68f100]

Thats not quit what I expected. I expected one of the drives to show up as bad. The only drive that is showing bad is the RAID 5 set.

E File System Check : FSCK fatal error = 39 Disk 1 12/4/2006 13:17:09

This is reporting disk 1, which is the 2nd drive position (0,1,2,3).

You may have problems finding a 30 gig drive. They are a little scarce. RAID 5 are a little picky when it comes to drives. Open the case and get the mfg, and model number and go hunting. It best to have them all the same.

It's my under standing you need the data off, Correct?

[tta013]

Data off? Actually, I want the disk back to normal - being able to use again. If that means I need to format the disk, I would do it.

Because as soon as Snap has problems, I redirected the bkf(s) to somewhere else.

So would it be idea to format the drive instead of finding a perfect match drive?

Regards

[blue68f100]

Data OFF Good News.

If you are going to do an upgrade. Knock it back to JBOD, before you swap in the new drives. You may want to do that and re-format the drive 1 that is giving off the errors. The try to build the array and see if it clears. I don't know about you, I don't trust drives that generate errors. I replace them.

You will proably find it makes better econimics to replace all 4 drives than trying to find a match for the 30 gig. Don't know the age of the drives, you may be getting into that window. It would be nice to have a 5th drive for a spare if you have more that one of these units, or other unit tha can use them.

The last time I checked you could buy 250 gigs for under $100 each.

Do you know what size PS is installed in the 4200?

v3.1.603 is that a Guardian Version?

Don't recall if the 4200 has OS in flashram, think not. on first drive (0).
If you replace all execpt the first one. I think the OS will be copied to the other drives. Then install in a new drive 0. That way you have an good drive with the OS on it. Just incase you need to copy the boot tracks.

[re3dyb0y]

There is conflicting information

You cant have a 4200 on Snap OS!

It is a guardian unit

Is it a 4100 or 4400 or 4000?

[blue68f100]

There have been a few that tried the switch to Snap OS because they could get it. Using a image file it worked, i think. It's been over a year.

[tta013]

It is Quantum. Snap Server M4100. Model Number 70700071-001.
Does anyone know the command for fomatting the disks and building back the array? I wonder if there is a step by step tutorial for this.

I would be really grateful if someone can help me on this please.

Regards

[tta013]

I just open up the server and I can see that there are 4 disks:

Disk1
ID#
10000

Disk2
ID#
1008

Disk3
ID#
10010

Disk4
ID#
10018

All disks have got Quantum Fireball (Tm) Plus LM
3.5 Series over it.

Thought it would be helpful to know above info.

[blue68f100]

Disk 1, bad with errors. is the boot disk. OS in flashram.

tta03, typo on disk2 10008

If no data change to JBOD, then back. or Use the "co de resync 60000" cmd.

If the disk as been replaced, you just make it as a hot spare and it will join the set.

[tta013]

What would this command - co de resync 60000 do?

Will it reformat the hard drives?
Will it delete only data or do I need to install the OS as well?

I would like to reformat the dirve. But if I do will I also lose the OS as well?

At this stage, I am getting confused. Really.

Thanks

[webboy]

Can we change the name of this thread to say 4100 not 4200.

Buy 4 matching 120 gig drives. pull all the drives out and replace with new drives.

Boot up the snap server wait 2 minutes then hit it with a web browser by its ip address. Login as admin and go to Disk Utilities/View Disk Status and watch as it formats all the drives.

The OS is in FLASH MEMORY and not on the drives. When its done formatting the drives you can build a new raid array somewhere around 390 gigs a hell of alot better then the 90 gig raid array you have now. :D

[tta013]

Thanks for all the repies. It helps a lot. I really appreicate it.

As I understand correctly the error was on Disk 1. Do I really need to buy 4 drives? Can I not buy one hard disk (Quantum Fireball (Tm) Plus LM
3.5 Series) and replace for Disk 1 and keep the other three? Sorry, I am sure this sound quite wrong for you, but I want to do I really have to buy 4, 120 G drives.

[tta013]

Hi All.

Again it is me. I now understand why buying 120 G drives was suggested.

I couldn't find an exact match for Quantum FireballTM Plus LM. I called a couple of resellers and nobody seems to have that drive anymore.

So it is time for 4 new hard drives.

SO I was wondering is there recommendation for Snap Appliance 120 G Hard dives?
Which brand, which made, Where to buy in UK etc.


Regards
tta013

[blue68f100]

The older model snaps cam with IBM / Hitachi drives. I generally use Hitachi or Seagates. When you set raids, It is prefered but not required that all drives be the same models, size and firmware. The drives must be with in 2-3% of each other. If you start mis-matching drives you run in to problems. Different mfg have different, heads, cyl, platters.....

If you need to recover the data you will need to find a drive that is close. There are a few utilities that allow you to change the capacity.

Server drives are built a little heavier the most drives, also more expensive.

[re3dyb0y]

Bust most server dives are SCSI


Here in the uk i tend to use a company called CCL www.cclonline.com

Rather than people such as ebuyer, as ebuyers tech support is crap and its very hard to get stuff replaced

IDE hard drives
http://www.cclonline.com/product-categories.asp?category_id=111

Take your pic on size, but get 4 oif the same

Like david, i wouuld reccommend seagate

250GB
http://www.cclonline.com/product-info.asp?product_id=3297&category_id=111&manufacturer_id=0
£61.52 inc

300GB
http://www.cclonline.com/product-info.asp?product_id=3298&category_id=111&manufacturer_id=0
£72.87 inc


Thats as big as you would want i should've thought

And delivery is £7.50 +vat for std usually next day delivery!

[tta013]

Hi,

Thanks for your reply.
So are you suggesting that if I can find a drive similar to Quantum Fireball (Tm) Plus LM 3.5 Series, it is ok to replace with Disk 1 ?

Sorry , what does it mean by (The drives must be with in 2-3% of each other) ?

You mentioned that "when you set raids, It is prefered but not required that all drives be the same models, size and firmware". You also mentioned that "If you start mis-matching drives you run in to problems." That makes me confused.

Now, as you know I have got Snap (4100) with 4 Quantum Fireball (Tm) Plus LM 3.5 Series Drives. And that Dirve 1 has got a problem.

The way I understand is I need to either replace Drive 1 with another Quantum Fireball (Tm) Plus LM 3.5 Series or buy 4 Dirves (120 G) to replace all drives.

I tried the first option but couldnt' find the exact match for Quantum Fireball (Tm) Plus LM 3.5 Series. If you are suggesting that I don't need to buy drive with same model, firm and size, what sort of hard dirve I should buy to replace Disk 1 which is Quantum Fireball (Tm) Plus LM 3.5 Series to avoid mis-matching.

Thanks in advance.

[tta013]

Thanks re3dyb0y. I just saw your post after submitting the previous one.

Thanks.

[blue68f100]

RAID arrays are picky.

Look at the specs of the Fireball, Platters, heads, cyls.... Capacity. If the heads,cyl, platters are the same with the same capacity it may work. All drives these days have a lot of spare cyls/capacity. Speed, seek times need to be close also. Since the no longer test/format the drives any more, they relie on the SMART technology to do the correction. You are below the LBA48 which throws another twist in the works.

One brand may list there drives as 120, other may show 123 .... With the smart technology you can actually adjust the drive size/capacity. Hitachi can do this.