FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Debian > Debian User

 
 
LinkBack Thread Tools
 
Old 01-05-2009, 05:46 PM
Dov Oxenberg
 
Default Partitioning and allocating a logical disk....

Hi,

Hopefully someone can provide*guidance (enlightenment?) to this relatively newbie Debian admin.

I recently built a Debian 4.0 stable system on an HP ProLiant DL380 G3 hardware platform (x86 32 bit).

The Server has five 72GB hot swap SCSI disks and I configured two arrays (using the HP ACU) - one logical disk 72GB RAID1 to house the O/S and the remaining 200GB array as a second logical disk which I intend to use as storage for FTP upload/download.

VSFTPD is installed and running, I have generated and installed a Certificate, everything there is fine.

What I need help with is in using the available space on the second logical disk.

I suspect I fudged the partitioning - I THOUGHT I partitioned the 200GB drive as EXT3 and the space was available for use, but here is what is happening.* I created a directory under /home named "secureftp."* I also added this entry to my FSTAB, contents of which look like this:

SERVERNAME:/# cat /etc/fstab
# /etc/fstab: static file system information.
#
# <file system> <mount point>** <type>* <options>****** <dump>* <pass>
proc*********** /proc********** proc*** defaults******* 0****** 0
/dev/cciss/c0d0p1 /************** ext3*** defaults,errors=remount-ro 0****** 1
/dev/cciss/c0d0p5 none*********** swap*** sw************* 0****** 0
/dev/hda******* /media/cdrom0** udf,iso9660 user,noauto**** 0****** 0
/dev/fd0******* /media/floppy0* auto*** rw,user,noauto* 0****** 0
/dev/cciss/c0d1p1 /securedata** ext3*** defaults******* 0****** 0

*

Now, when I try to mount this directory, wishing to place it on the 200GB storage array (sorry if I am not wording this correctly, my background is in Windows and have only been working with Linux for a couple years) I run the command "mount -t ext3 /dev/cciss/c0d1p1 /securedata"*the system responds with "mount: wrong*fs type, bad option..."* Hopefully you will know what I mean without me having to type out the rest of the message.* In the message I am directed to have a look at syslog so I*issue the command as follows:

SERVERNAME:/# dmesg | tail
***** heads= 255, sectors= 32, cylinders= 52299

*cciss/c0d1: p1 < >
***** blocks= 426759840 block_size= 512
***** heads= 255, sectors= 32, cylinders= 52299

*cciss/c0d1: p1 < >
attempt to access beyond end of device
cciss/c0d1p1: rw=0, want=4, limit=2
EXT3-fs: unable to read superblock

*

I thought the "superblock" was the 5% of space reserved by the system so I can perform root maintenance should the disk become full, the data corrupt, etc.

Anyway,*would someone please (and be kind...) tell me what I am doing wrong, how I can fix it, and provide the steps I need to follow in order to successfully mount the 200GB storage partition?

Thanks for much!!!

Dov

*



*
 
Old 01-05-2009, 07:24 PM
"Boyd Stephen Smith Jr."
 
Default Partitioning and allocating a logical disk....

On Monday 2009 January 05 12:46:03 Dov Oxenberg wrote:
> The Server has five 72GB
> hot swap SCSI disks and I configured two arrays - one
> logical disk 72GB RAID1 to house the O/S and the remaining 200GB array as a
> second logical disk which I intend to use as storage for FTP
> upload/download.

RAID-0? I assume this is only temporary storage then, yes? If one drive
fails in a RAID-0, the entire array is toast.

> What I need help with is
> in using the available space on the second logical disk. I suspect I fudged
> the partitioning - I THOUGHT I partitioned the 200GB drive as EXT3 and the
> space was available for use, but here is what is happening.

What tool did you use? There's no partition type specifically for ext3,
AFAIK. Even if there was, you'd still need to create the filesystem. To do
that from the command-line something like "mkfs.ext3 /dev/cciss/c0d1p1" would
be required.

> I created a
> directory under /home named "secureftp." I also added this entry to my
> FSTAB, contents of which look like this:

> SERVERNAME:/# cat /etc/fstab#
> /etc/fstab: static file system information.## <file system> <mount point>
> <type> <options> <dump> <pass>proc /proc proc
> defaults 0 0/dev/cciss/c0d0p1 / ext3
> defaults,errors=remount-ro 0 1/dev/cciss/c0d0p5 none swap
> sw 0 0/dev/hda /media/cdrom0 udf,iso9660
> user,noauto 0 0/dev/fd0 /media/floppy0 auto
> rw,user,noauto 0 0/dev/cciss/c0d1p1 /securedata ext3 defaults
> 0 0

I assume it's not missing all the newlines and your mailer just screwed that
up? If not, that might be part of your problem.

Also, your prose above indicates you wanted it mounted at /home/secureftp, but
your fstab entry says you want it at /securedata. Make sure you are
consistent.

> I run
> the command "mount -t ext3 /dev/cciss/c0d1p1 /securedata" the system
> responds with "mount: wrong fs type, bad option..."

Basically, that means that it can't find the filesystem you've specified.

> SERVERNAME:/# dmesg | tail heads= 255, sectors= 32,
> cylinders= 52299 cciss/c0d1: p1 < > blocks= 426759840 block_size= 512
> heads= 255, sectors= 32, cylinders= 52299 cciss/c0d1: p1 < >attempt to
> access beyond end of devicecciss/c0d1p1: rw=0, want=4, limit=2EXT3-fs:
> unable to read superblock

Attempt to access beyond the end of device is a bit unusual. I suppose that
might simply have been the ext3 driver looking for one of the extra recovery
superblocks. For now, I'll ignore it and focus on "unable to read
superblock" which means the filesystem is either corrupt or hasn't been
created yest.

> I thought the "superblock" was the 5% of space reserved by the system so I
> can perform root maintenance should the disk become full, the data corrupt,
> etc.

No, the superblock is the area of the filesystem (usually stored toward the
beginning of the drive) that contains information about the filesystem as a
whole. For example, it generally stores what size the filesystem thinks it
is, what block size is in use (if the fs supports multiple block sizes), any
options or sub-options that are expected to be persistent, maybe a free space
list or bitmap, etc. It also stores things like UUID and LABEL. Most, if
not all, filesystems have one.

The reserved space and inodes are fairly ext[234] specific -- most other
filesystems will let you shoot yourself in the foot quite easily. Also, even
on ext[234] that percentage is adjustable and can be eliminated entirely if
need be. UNIX and Linux systems traditionally behave very poorly, if at all,
when no filesystem can be written to -- a lot of programs/scripts require at
least some scratch space. This reservation by ext[234] can make it quite a
bit easier to recover from that state.
--
Boyd Stephen Smith Jr. ,= ,-_-. =.
bss@iguanasuicide.net ((_/)o o(\_))
ICQ: 514984 YM/AIM: DaTwinkDaddy `-'(. .)`-'
http://iguanasuicide.net/ \_/
 
Old 01-05-2009, 07:47 PM
Dov Oxenberg
 
Default Partitioning and allocating a logical disk....

Hello Stephen,

Thank you for taking the time to reply.

The O/S is on a mirrored array, and the storage array is RAID5.

The command*I used to create the partition is exactly as you wrote "mkfs.ext3 /dev/cciss/c0d1p1"

Indeed, I created a directory under /home and named it "secureftp."* I want to mount just the "secureftp" directory.

Is there something wrong with the command syntax I used?* I entered the command exactly as I wrote in my original e-mail.

Ok, so the system can't find my target file system, would you please explain why so I may know where I went wrong and how I can*fix it?

*

Thanks

*



> From: bss@iguanasuicide.net
> To: debian-user@lists.debian.org
> Subject: Re: Partitioning and allocating a logical disk....
> Date: Mon, 5 Jan 2009 14:24:15 -0600
>
>
> RAID-0? I assume this is only temporary storage then, yes? If one drive
> fails in a RAID-0, the entire array is toast.
>
>
> What tool did you use? There's no partition type specifically for ext3,
> AFAIK. Even if there was, you'd still need to create the filesystem. To do
> that from the command-line something like "mkfs.ext3 /dev/cciss/c0d1p1" would
> be required.
>
> > I created a
> > directory under /home named "secureftp." I also added this entry to my
> > FSTAB, contents of which look like this:
>
> > SERVERNAME:/# cat /etc/fstab#
> > /etc/fstab: static file system information.## <file system> <mount point>
> > <type> <options> <dump> <pass>proc /proc proc
> > defaults 0 0/dev/cciss/c0d0p1 / ext3
> > defaults,errors=remount-ro 0 1/dev/cciss/c0d0p5 none swap
> > sw 0 0/dev/hda /media/cdrom0 udf,iso9660
> > user,noauto 0 0/dev/fd0 /media/floppy0 auto
> > rw,user,noauto 0 0/dev/cciss/c0d1p1 /securedata ext3 defaults
> > 0 0
>
> I assume it's not missing all the newlines and your mailer just screwed that
> up? If not, that might be part of your problem.
>
> Also, your prose above indicates you wanted it mounted at /home/secureftp, but
> your fstab entry says you want it at /securedata. Make sure you are
> consistent.
>
> > I run
> > the command "mount -t ext3 /dev/cciss/c0d1p1 /securedata" the system
> > responds with "mount: wrong fs type, bad option..."
>
> Basically, that means that it can't find the filesystem you've specified.
>
> > SERVERNAME:/# dmesg | tail heads= 255, sectors= 32,
> > cylinders= 52299 cciss/c0d1: p1 < > blocks= 426759840 block_size= 512
> > heads= 255, sectors= 32, cylinders= 52299 cciss/c0d1: p1 < >attempt to
> > access beyond end of devicecciss/c0d1p1: rw=0, want=4, limit=2EXT3-fs:
> > unable to read superblock
>
> Attempt to access beyond the end of device is a bit unusual. I suppose that
> might simply have been the ext3 driver looking for one of the extra recovery
> superblocks. For now, I'll ignore it and focus on "unable to read
> superblock" which means the filesystem is either corrupt or hasn't been
> created yest.
>
> > I thought the "superblock" was the 5% of space reserved by the system so I
> > can perform root maintenance should the disk become full, the data corrupt,
> > etc.
>
> No, the superblock is the area of the filesystem (usually stored toward the
> beginning of the drive) that contains information about the filesystem as a
> whole. For example, it generally stores what size the filesystem thinks it
> is, what block size is in use (if the fs supports multiple block sizes), any
> options or sub-options that are expected to be persistent, maybe a free space
> list or bitmap, etc. It also stores things like UUID and LABEL. Most, if
> not all, filesystems have one.
>
> The reserved space and inodes are fairly ext[234] specific -- most other
> filesystems will let you shoot yourself in the foot quite easily. Also, even
> on ext[234] that percentage is adjustable and can be eliminated entirely if
> need be. UNIX and Linux systems traditionally behave very poorly, if at all,
> when no filesystem can be written to -- a lot of programs/scripts require at
> least some scratch space. This reservation by ext[234] can make it quite a
> bit easier to recover from that state.
> --
> Boyd Stephen Smith Jr. ,= ,-_-. =.
> bss@iguanasuicide.net ((_/)o o(\_))
> ICQ: 514984 YM/AIM: DaTwinkDaddy `-'(. .)`-'
> http://iguanasuicide.net/ \_/
 
Old 01-05-2009, 08:56 PM
"Boyd Stephen Smith Jr."
 
Default Partitioning and allocating a logical disk....

On Monday 2009 January 05 14:47:28 Dov Oxenberg wrote:
> Hello Stephen,
> Thank you for taking the time to reply.
> The O/S is on a mirrored array, and the storage array is RAID5.

The math doesn't work on that. You said 5x 72G disks, 2 in mirrored array.
That leaves 3x 72G disks for the storage array. In RAID-5 that's (3 - 1) *
72G = 144G of storage. In RAID-0 that's 3 * 72G = 216G. You said you had
200G of storage.

Now, the "access beyond end of device" error worries me. I think that (at
least) your storage array was created incorrectly or has become corrupt.

> The command I used to create the partition is exactly as you wrote
> "mkfs.ext3 /dev/cciss/c0d1p1"

That's the command you used to create the *filesystem*. You probably created
the *partition* with fdisk, gparted, or a similar tool. The ADC might even
let you do that.

partition != filesystem. Most of my filesystems are not on partitions, most
of my partitions do not hold filesystems. Instead, most of my partitions are
md component devices and most of my filesystems on are logical volumes.

> Indeed, I created a directory under /home and
> named it "secureftp." I want to mount just the "secureftp" directory. Is
> there something wrong with the command syntax I used? I entered the
> command exactly as I wrote in my original e-mail.

You wrote "mount -t ext3 /dev/cciss/c0d1p1 /securedata". If you want to mount
the filesystem at /home/secureftp, you want to use the command "mount -t
ext3 /dev/cciss/c0d1p1 /home/secureftp". The second pathname argument to
mount is the "mount point", the pathname tied to the root of the filesystem.

> Ok, so the system can't
> find my target file system, would you please explain why so I may know
> where I went wrong and how I can fix it?

fsck.ext3 should be able to put the filesystem in a consistent state. It may
be able to salvage data as well, but I don't think you've written any data to
the filesystem yet, right?

> Thanks

[snip: hard to read, poorly re-wordwrapped, unedited original mail.]

If your mail client can't quote correctly, please don't quote anything.
However, it would be preferable for you use get a good client or learn how to
use it and quote properly. http://en.wikipedia.org/wiki/Posting_style talks
about the different types. While I prefer interleaved/replying inline,
there's no mandated style on this list, AFAIK. Whatever style you use,
please trim irrelevant text.
--
Boyd Stephen Smith Jr. ,= ,-_-. =.
bss@iguanasuicide.net ((_/)o o(\_))
ICQ: 514984 YM/AIM: DaTwinkDaddy `-'(. .)`-'
http://iguanasuicide.net/ \_/
 
Old 01-05-2009, 09:15 PM
Dov Oxenberg
 
Default Partitioning and allocating a logical disk....

Hello Stephen,

Thank you for your prompt reply.

Please forgive me, I actually have six 72GB drives and the second array is seen by the O/S as having about 203GB of space, sorry for the mix-up.*

And I believe you are correct, I used fdisk to create the partition, again my apologies, and currently there is no data on that partition so even if I needed to reformat the storage array and start over it should be no problem.

Just to be clear, I wish the mount point to be just "/secureftp."**I just wanted to point out where I created the directory in case it makes any difference.

Hopefully I have clarified all the questionable items - that being the case, how can I now mount the storage array and make it available for data storage?

*

Thanks!

Dov


*
 
Old 01-05-2009, 09:35 PM
"Boyd Stephen Smith Jr."
 
Default Partitioning and allocating a logical disk....

On Monday 2009 January 05 16:15:10 you wrote:
> Hello Stephen,
> Thank you for your prompt reply.
> Please forgive me, I actually have six 72GB drives and the second array is
> seen by the O/S as having about 203GB of space, sorry for the mix-up.

Ah, very good then. That should be fine.

> And I
> believe you are correct, I used fdisk to create the partition, again my
> apologies, and currently there is no data on that partition so even if I
> needed to reformat the storage array and start over it should be no
> problem.

Run mkfs.ext3 again, and check dmesg afterward for any errors/warnings, copy
and paste them in your reply if possible.

> Just to be clear, I wish the mount point to be just "/secureftp."

Okay, make sure that directory exists. What's the output of:
ls -ld /secureftp

> I just wanted to point out where I created the directory in case it makes
> any difference.

It doesn't make a difference as long as you are *consistent*. You
can't "mkdir /home/secureftp", edit /etc/fstab to contain "/securedata", and
then mount with "mount -t ext3 /dev/cciss/c0d1p1 /secureftp" -- they all need
to have the same path to the point point.

> how can I now mount the storage array and make it
> available for data storage?

You seem to have the steps basically down:
1. Use the array-specific tools to create a RAID disk. In you case something
that show up to linux like /dev/cciss/c0d1.
2. Partition that disk with fdisk (or similar tools). This should make a
device like /dev/cciss/c0d1p1 appear.
3. Make filesystem(s) with mkfs.
4. Make mount point(s) with mkdir.
5. Edit /etc/fstab to tie the partition holding the filesystem to the mount
point.
6. Issue a mount command like: mount -t <fs_type> <partition> <mount_point>.

I don't have any experience with the cciss stuff in specific. They might
require a reboot after step 1 or 2.

Watch for errors during the process. Not just from the commands themselves,
but from in dmesg.
--
Boyd Stephen Smith Jr. ,= ,-_-. =.
bss@iguanasuicide.net ((_/)o o(\_))
ICQ: 514984 YM/AIM: DaTwinkDaddy `-'(. .)`-'
http://iguanasuicide.net/ \_/
 

Thread Tools




All times are GMT. The time now is 10:11 AM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org