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

I don't remember if there are any options needed, but I ran e2fsck on /dev/sda2.  The sata drives under linux appear as scsi drives.  Partition 1 is a swap partition and partition 2 is what makes up the 'data' volume.  When I started e2fsck it started pass 1, about 15 or so minutes later it was still on pass one.  As I was testing things out I did not have data on the drives at that time so I just reformatted.   I seems that the 323 never runs clean if you are running raid 1. 

Not sure if anyone here who knows more about linux can check the src code and see if the code to cleanly unmount the md devices before a restart is even in there, let alone getting executed.  But it's my guess that they do not.  As I'm only looking for a redundant storage, I have not checked my setup under any other raid/non-raid types.

Adding an update, figured I'd load up the scripts again and see what happens if I complete the e2fsck on both drives.  Just started on sda2 with e2fsck -y, will see how long it takes.  These are Seagate 320gb drives.

# e2fsck -y /dev/sda2
e2fsck 1.32 (09-Nov-2002)
/dev/sda2 was not cleanly unmounted, check forced.
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
Free blocks count wrong (76376011, counted=76376228).
Fix? yes


/dev/sda2: ***** FILE SYSTEM WAS MODIFIED *****
ERROR#


Now on to disk 2

Exact same error

# e2fsck -y /dev/sdb2
e2fsck 1.32 (09-Nov-2002)
/dev/sdb2 was not cleanly unmounted, check forced.
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
Free blocks count wrong (76376011, counted=76376228).
Fix? yes


/dev/sdb2: ***** FILE SYSTEM WAS MODIFIED *****
ERROR#

After the last reboot I only get the warning in one place instead of 4 or 5.  This seems to be mounting the flash FS used for root.

EXT2-fs warning: mounting unchecked fs, running e2fsck is recommended
VFS: Mounted root (ext2 filesystem).


I still see several attempts to read an ext3 journal on the md0 device.

Going to copy a few hundred gig of data and reboot it a few times and see what happens.

Last edited by LifeTap (2007-01-11 01:26:41)

I assumed since the last lines before the last e2fsck error said unbinding the drives that it wasn't actually in a mirror at that time.  Guess I should have looked.

to mig: nope, e2fsck runs fine on non root mounted FS

I have done a bit of investigation on the errors related to unclean dismounting of /dev/md0 after modifying the telnet access system (which I shall post here later)

The problem does not go away even after unmounting the /dev/md0 and stopping the RAID- before rebooting DNS323. This is definitively not right... It is still work in progress and I will share more details as I have them.

*** Update

For the filesystem to be uncleanly dismounted it is enough to umount /dev/md0, stop and start RAID. Perhaps indicates a problem with mdadm?

Last edited by skydreamer (2007-01-11 16:47:28)

is there a way to insert a check/fix disk command in one of the start-up scripts - so when the system boots it gets fixed. That way, there would be no need to telnet in ..

skydreamer, thanks for your work on this so far. 

I feel a little better about the ext3 issue, thou I think d-link should have removed everything pertaining to ext3 when they removed it as an option.  Of course they are probably working out the bugs and may plan to bring it back in the near future and didn't think people would be looking at the dmesg output.

As far as the raid not being properly set up this may or may not mean anything.  When I format the array, for the first 10% of the status bar, the drives aren't accessed.  Then the activity lights flash from about 10 to 35%.  After that the progress bar jumps to ~95% and quickly goes to 100% and asks to restart.  I thought maybe drive cache or something, but the lights stopped flashing, so I restart.

Something else that happens when I want to start over.  I'll go back in via the web interface and select to setup and format the raid array.  It goes thru and formats as above then prompts for restart.  What happens next is a bit random.  If I log into the web interface again I get one of the below.

1.  Unit is up and functioning, ready for data.  Get the standard screen to configure, view status.
2.  Page comes up that states array is synchronizing and will complete in X minutes.  After that I get normal web page.
3.  Page is displayed that my newly replaced drive needs to be formatted, formats, and asks for a restart.  After that I get normal web page.

If I pull out the drives, hook them to my linux box via usb adapter and remove the partitions it seems to work first time and I always get response #1 above.

Also these are Seagate Barracuda 7200.10 ST3320620AS drives.

Last edited by LifeTap (2007-01-11 23:52:06)

skydreamer wrote:

our friend Wilson Chen

You mean this guy (the guy on the left):



Bug me not!

;-)

/Apan

Ie do I have to unmount HD_a2 and HD_b2 before I run e2fsck? If so, what happens when I unmount HD_a2 where the telnetd and busybox is installed?

Should I use any special switches?

etc etc

Last edited by MiK (2007-01-19 00:06:40)

The message was already answered in another thread -

http://dns323.kood.org/forum/t133-Raid- … dmesg.html


The specific reply cam from skydreamer in his post on the 11th of Jan.

http://dns323.kood.org/forum/p752-20070 … .html#p752



It is not a good idea to run e2fsck on individual drives that are in RAID1. Do instead:

# e2fsck /dev/md0

skydreamer wrote:

It is not a good idea to run e2fsck on individual drives that are in RAID1. Do instead:

# e2fsck /dev/md0
e2fsck 1.32 (09-Nov-2002)
/dev/md0: clean, 16440/30425088 files, 25537481/60818048 blocks
#

hm, i'm still a bit worried.

--------------------------
disk:~# e2fsck /dev/md
e2fsck 1.40-WIP (14-Nov-2006)
/dev/md0 is mounted.

WARNING!!!  Running e2fsck on a mounted filesystem may cause
SEVERE filesystem damage.

Do you really want to continue (y/n)?
-----------------------------

Do i need to unmount now, or is it ok to run all the checks even the Raid is mounted?