DSM-G600, DNS-3xx and NSA-220 Hack Forum

fourthirtysix · 2008-06-04 00:47:44

I have two working DNS-323s but I am having problems with one of them. They are both mounted the same way and configured the same way. On the second DNS-323 I have a problem where I can only copy a few files to it before I get "Input/Output Error".

I can browse the system fine, I can telnet to it and the output of /proc/mdstat is good.

I don't know what the problem is, so I think the next step is to try e2fsck? When I follow these instructions at http://www.inreto.de/dns323/fsck/ I get a problem because the RAID is active and so the script cannot turn of the RAID in order to reboot.

I also get the following problem on the computer that is trying to access the DNS-323 in syslog:

CIFS VFS: Write2 ret -5, wrote 0

and this repeats many times.

Can someone suggest what the problem might be, or tell me how to run the e2fsck? When I follow the instructions at the link, it fails because it finds /tmp/raidup and aborts too early.

Thank you

jh · 2008-06-04 01:20:09

# e2fsck -n /dev/md0

e2fsck 1.38 (30-Jun-2005)
Warning! /dev/md0 is mounted.
/dev/md0 contains a file system with errors, check forced.
Pass 1: Checking inodes, blocks, and sizes

fourthirtysix · 2008-06-04 01:58:51

Thank you. I've found at least one problem by running e2fsck in -n READ ONLY mode.

How can I temporarily turn off the RAID to run e2fsck in regular fix-it mode?

fourthirtysix · 2008-06-04 09:14:17

I have tried several more things by rebooting my computer, the DNS unit, resetting connections, etc. The problem continues with just one of the DNS-323 and not the other.

I have not gotten any warnings so far, so I do think I need to run e2fsck, but I cannot figure out how. I still have the same problem I mentioned before about not being able to unmount the drives because i am logged into telnet which uses the mount.

There are one or two threads that discuss this, but unfortunately I cannot figure out how to make the /tmp/raidup to go away so that fonz' fsck trick will work.

Does anyone have suggestion on how to resolve this so that I can run telnet from ramdisk and run fsck??

bq041 · 2008-06-04 15:19:48

Code:

hd_verify

This is a utility run on the DNS at startup. When it runs, it unbounts the drives. You can run this (ignore its errors) and then run e2fsck on the array. Once it is done, run:

Code:

do_reboot

fourthirtysix · 2008-06-07 23:04:18

ok... i think I need to send out a big plea for HELP!

the drives are still giving me "Input/Output error" when I try to copy files to it. The drive seems to read files fine and I can browse them in Linux File Manager Nautilus.

I ran hd_verify and also "smb stop" and next e2fsck.. here is the output:

Code:

/ # e2fsck /dev/md0
e2fsck 1.32 (09-Nov-2002)
The filesystem size (according to the superblock) is 183010448 blocks
The physical size of the device is 182715248 blocks
Either the superblock or the partition table is likely to be corrupt!
Abort<y>? no

/dev/md0 contains a file system with errors, check forced.
Pass 1: Checking inodes, blocks, and sizes
Inode 15679491, i_blocks is 64, should be 24.  Fix<y>? yes

Error reading block 182747138 (Invalid argument) while doing inode scan.  Ignore error<y>? yes
Force rewrite<y>? yes
Error writing block 182747138 (Invalid argument) while doing inode scan.  Ignore error<y>? yes
Error reading block 182747139 (Invalid argument) while doing inode scan.  Ignore error<y>? yes
Force rewrite<y>? yes
Error writing block 182747139 (Invalid argument) while doing inode scan.  Ignore error<y>? yes
Error reading block 182747140 (Invalid argument) while doing inode scan.  Ignore error<y>? yes
Force rewrite<y>? yes
Error writing block 182747140 (Invalid argument) while doing inode scan.  Ignore error<y>? yes
Error reading block 182747141 (Invalid argument) while doing inode scan.  Ignore error<y>? yes
Force rewrite<y>? yes
Error writing block 182747141 (Invalid argument) while doing inode scan.  Ignore error<y>? yes
Error reading block 182747142 (Invalid argument) while doing inode scan.  Ignore error<y>? yes
Force rewrite<y>? yes
Error writing block 182747142 (Invalid argument) while doing inode scan.  Ignore error<y>? yes
Error reading block 182747143 (Invalid argument) while doing inode scan.  Ignore error<y>? yes
Force rewrite<y>? yes
Error writing block 182747143 (Invalid argument) while doing inode scan.  Ignore error<y>?

Can anyone tell me what might be wrong or how to fix it?? I can't write anything to the drive but the system does not say that the drives have failed!

bq041 · 2008-06-07 23:10:48

From my experience, I think your best bet now is to break the array, fix the disks individually, then re-assemble the array. It is not too hard to do, but it will take a little time. Let me know if you want to go that route. Somebody else may know a better way, but I do not.

I have scripts that will break and reassemble the array without data loss, if you are using F/W 1.04 or 1.05. I have not made the ones for 1.03, yet.

fourthirtysix · 2008-06-12 05:47:32

I'm trying to figure out what I'm going to do about this. It stinks that I can't write to the disks even though neither disk has failed.

It seems strange to me that there is no easier way to fix this problem? I would think that people would get bad spots on their hard drives every now and then and this would be a much more common issue!

If you have time to port the script to 1.03, that would be awesome. I haven't upgraded due to concerns with breaking a (previously) stable working environment.

I love the DNS-323 but there has got to be a better way to deal with a disk that doesn't fail entirely, but where this kind of thing happens!

AAAARGH! :-o

bq041 · 2008-06-12 06:01:04

The issue is not dealing with a disk that has not failed entirely, that is not much of a problem. The big deal is that you have an array. That makes both disk dependent upon each other and unfortunately, in order to access the disks on the DNS, at least one of them will be in use all the time and cannot be properly fdisked. I will see what I can do, but I am away from home until Friday.

fourthirtysix · 2008-06-12 06:44:40

Thanks! Maybe D-Link can add telnet or ssh as an option to one of the firmware updates?

bq041 · 2008-06-13 03:24:51

Somehow, I doubt that will happen.

bq041 · 2008-06-14 02:55:34

Why were you ignoring the errors in your above e2fsck?

fourthirtysix · 2008-06-14 22:35:57

Mostly because I'm not too familiar with running error checking on hard drives with linux tools Even though I have used linux for years, I've been fortunate to not have many HD problems until now....

Also, if I remember correctly, when I clicked "n" to ignore the error, the program was quitting on me!

Code:

/ # e2fsck /dev/md0
e2fsck 1.32 (09-Nov-2002)
The filesystem size (according to the superblock) is 183010448 blocks
The physical size of the device is 182715248 blocks
Either the superblock or the partition table is likely to be corrupt!
Abort<y>? no

/dev/md0 contains a file system with errors, check forced.
Pass 1: Checking inodes, blocks, and sizes

Inode 15679490, i_blocks is 64, should be 24.  Fix<y>? yes

Error reading block 182747138 (Invalid argument) while doing inode scan.  Ignore error<y>? no

Error while scanning inodes (91373568): Can't read next inode
e2fsck: aborted
Segmentation fault

bq041 · 2008-06-14 22:54:11

Okay, are you ready to check the individual disks? This will break the array. You can reform it, as long as you do not erase any of the configuration files. If you are ready,

Code:

/ # mdadm /dev/md0 -f /dev/sdb2 -r /dev/sdb2

Code:

/ # mdadm --zero-superblock /dev/sdb2

This will set drive 1 (left bay) to failed and remove it from the array. The second command will remove the superblock (RAID information).
Now, without rebooting run:

Code:

e2fsck /dev/sdb2

If this still gives you errors that cannot be fixed, then the next step will be to format this drive and recheck. You may have to run e2fsck more than once. If you can get everything good, we can add this disk back to the array and let it sync. Then we will remove the other disk and check it.

Keep in mind, at this point, both disks will still have your data (before refomat of course) and will have 1 disk with a broken RAID 1 array, and one individual. What may have to be done is to break the array fully (remove the other drive, but it requires some special instructions) and then re-set-up from scratch. Since i do not have scripts for this in F/W 1.03, we will have to go through it manually.

DSM-G600, DNS-3xx and NSA-220 Hack Forum

Announcement

#1 2008-06-04 00:47:44

Input/Output Error

#2 2008-06-04 01:20:09

Re: Input/Output Error

#3 2008-06-04 01:58:51

Re: Input/Output Error

#4 2008-06-04 09:14:17

Re: Input/Output Error

#5 2008-06-04 15:19:48

Re: Input/Output Error

Code:

Code:

#6 2008-06-07 23:04:18

Re: Input/Output Error

Code:

#7 2008-06-07 23:10:48

Re: Input/Output Error

#8 2008-06-12 05:47:32

Re: Input/Output Error

#9 2008-06-12 06:01:04

Re: Input/Output Error

#10 2008-06-12 06:44:40

Re: Input/Output Error

#11 2008-06-13 03:24:51

Re: Input/Output Error

#12 2008-06-14 02:55:34

Re: Input/Output Error

#13 2008-06-14 22:35:57

Re: Input/Output Error

Code:

#14 2008-06-14 22:54:11

Re: Input/Output Error

Code:

Code:

Code:

Board footer