possible partition error suse linux boot

English Install/Boot/Login
#1

We had a perfectly working SUSE OES netware server for some months, and wanted to create a new image backup. Using Acronis Backup and Recovery Advanced Server Software. This worked a treat on our SBS 2008 Server. We went to run it for Novell OES and a system lock up occured and services such as Iprint then disabled. A server restart recovered the server after a few attempts.

We then learnt that the server could not write files to an NTFS formatted USB drive. This was the ideal destination to test the backup. However we could not even write a simple file to the drive without an error message about permissions.

We then learnt that Linux needs a package called NTFS-3G to enable NTFS writing permissions. This did fix the writing issues. We attempted to run another back up which again failed, bringing down the server. We again attempted to reboot the server this time we fail to get to the graphical user login.

At the suse Linux Enterprise boot screen we select “boot from hard disc”. Which fails immediately with “GDM could not write a new authorisation entry to disk…Error no space left on drive”. When the boot eventually continues, a number of things fail with the same message about “No space left on device”.

The next message starts “Could not start the X server due to some internal error.”

Further to the above, starting “recovery” instead, the checks of partition fail and it seems obvious that the partition definitions have become corrupted. In the expert partion manager the lines for partitions seem correct except that there are no entries for ‘mount point’ and ‘mount by’ columns. Trying to edit these lines does not appear to allow entries for mount point or mount by - they are disabled.

It seems likely the server is not mounting some partitions at boot which is why the kernel thinks the server is full. Is there some simple way we can use say fdisk to repair the partition definition without loosing anything of the server OS and data ?

#2

I don’t understand what you’re doing.

I think you’re doing somethong like this: you boot the Linux machine (using the acronis cd to boot it) and use the acronis software on the cd to copy an image of the Linux machine across to an attached USB NTFS drive. Then you restore the image to the Linux drive and it won’t work.

Is that right. If it’s wrong, you will see my confusion and perhaps tell me what it is you’re doing so I can visualise the operations better.

#3

You forgot to specify which operating system you use and it’s version. But from a glance at your post I deduct that this is abouit SLES. You may get help here on the openSUSE Forums (we are not that xenophobe), but you mat be better serverd at SUSE Linux Enterprise Server (SLES)

#4

Swerdna - I’ll have to come back to you on that, it was not me doing the running of the Acronis backup. All I can say is to the best of my knowledge, this is two failed efforts to copy an image of the linux machine to an attached USB storage, resulting in the Linix machine not booting (as described with possible partitioning error). I am virtually certain this is not a restored image we are discussing - this was a live server, now it is not - we dont know why.

It may be possible the server drive is actually full if it was incorrectly used as a target and maybe the image would loop, getting bigger and bigger all the time. A pretty stupid thing to do, I cannot believe anyone would try to backup a server onto itself, but I can just understand it happenening by accident.

Also hccv - this is SLES netware as I understand. I have posted in the netware forums without response. The very fact that I get two virtually instant responses on **this **forum is telling me something very important.:slight_smile:

#5

Well, it is all nice that you like us answering, but that still does not give you a clue.

In fact, personaly I do not understand your problem fully (understatement) and I do not know Acronis.

Isn’t there anything you can show us? Aren’t there any error messages? And again it would be wise to tell us exactly which SLES you use. Also the output of

uname -a

might tell us a bit of what you are running.

#6

On 2010-10-13 21:06, davidm01 wrote:
>
> We had a perfectly working SUSE OES netware server for some months, and
> wanted to create a new image backup. Using Acronis Backup and Recovery
> Advanced Server Software. This worked a treat on our SBS 2008 Server. We
> went to run it for Novell OES and a system lock up occured and services
> such as Iprint then disabled. A server restart recovered the server
> after a few attempts.

This is impossible to understand.

First, specify exactly what operating system you are using. Then what exact operations you did, one
by one, step by step, exact error messages and exact actions. Detailed and clearly explained.

Please note that if this is SLES or SLED, as it seems, you have to ask in their specific forums, not
here - as we are not familiar with their problems. We can try to help, if you are clear in the details.

–
Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)

#7

Ok, have talked to one of the two people who were doing this, and am getting a moderately good picture of what decisions have been made, and what has happened. I should emphasise right away that what caused us to get into the situation we are may or may not get us back to a working system, but here goes.

This was an experienced and loyal Novell user. To move forward, it was wanted to continue with a familiar system if possible, and the nearest was SLES netware. We wanted simple user rights management, nothing much except file and iPrint services, NSS file system and Salvage. Extensive attempts to achieve this with SUSE 11.1 and 11.2 failed to support Netware. Success was achieved with SUSE 10.2.0.91, and there was a good system running tests for weeks. There are weeks of effort in this system concerning setting up users, data, facilities etc. All this is going to waste right now.

Before going live, the final tests of imaging and daily backup were begun, using Acronis Backup & recovery 10 Advanced server. This needs to be installed under yast in order to be scheduled, which was done. Attempts were made to backup to a USB drive, one could set the drives to backup and set the drive to backup to, but execution failed with errors ‘could not create archive’. Discovering a need to run linux software to enable writing to NTFS formatted drives, so this was installed.

A successful write of a small test file was completed, so the main server drives backup was begun, resulting in the same sort of ‘could not create archive’ error as before, plus other critical errors such as 0x00000FF1 and 0x0000FF2. FF1 seems to have to to with permissions. FF2 is unknown. At this point the server was frozen, and after a hard reset could not be restarted. Which is where we are now. The server seems to be full, the partitions don’t seem to be completely described.

I know just enough to run a few listing commands. Here are the results.
uname -a
linux linux 2.6.16.60-0.54.5-default #1 …

fdisk -l
Disk /dev/sda: 770.0Gb 7308…08 bytes
255 heads, 63 sectors/track, 88849 cylinders
Units = cylinders 0f 16065 * 512 = 8225208 bytes

device boot start end blocks Id system
/dev/sda1 * 1 9 72261 83 Linux
/dev/sda2 19 1966 15719602+ 8e Linux LUM
/dev/sda3 1967 88849 697887697+ 65 Novell Netware 386

Disk /dev/dm-0: 10.7Gb 10737418240 bytes
255 heads, 63 sectors/track. 1305 cylinders
Units = cylinders of 16065 * 512 = 8225208 bytes

Disk /dev/dm-0 doesn’t contain a valid partition table

Disk /dev/dm_1: 2147 Mb, 2147473648 bytes
two more lines

Disk /dev/dm-1 doesn’t contain a valid partition table

(my comment - no idea what the last two devices are, but the first three seem ok)

df -h
/dev/loop0 255M used 255M 100% mounted /mounts/instsys

mounts
rootfs on / type rootfs (rw)
tmpfs on / type tmpfs (rw)
tmpfs on / type tmpfs (rw)
proc on /proc type proc (rw)
sysfs on /sys type sysfs (rw)
usbfs on /proc/bus/usb type usbfs (rw)
/dev/loop0 on /mounts/instsys type cramfs (ro)
devpts on /dev/pts type devpts (rw)

Right now, the console screens say
F1 screen
about 30 lines ending
Found a Linux console termial on /dev/console…
*** Starting YAST2 ***

F2 screen
a command screen with the results shown about (oh my poor fingers)

F3 screen
some sort of install screen log probing hardware, finishing
Loading image “/var/adm/mount/boot/x86_64/root”…
mount /download/image0: we need a loop device
mount: using /dev/loop0
starting setctsid ‘showconsole’ inst_setup yast

F4 screen
some logging every 30 minutes all saying Oct 13 14:29:53 linux – MARK –
and other times, then
Oct 14 09:44:49 linux kernel: Adding 2097144k swap on /dev/system/swap. Priority:-4 extents:1 across:2097144k
and some more times

F5,F6 screen empty

F7 screen
Yast repair Tool Box GUI screen

F8 screen
A panic screen, continuously refreshing lines
ACPI: Trying to reopen
ACPI: opening device

(few million records, I suspect)
need to be stopped.

#8

This is going to be a bit difficult I am afraid. Apparently you did not (or could not) copy/paste the computer output to your post, but wrote/typed it down. That means we have crippled output (I e.g. can not understand that there are three mounts on /). Also not putting computer text between CODE tags here on the Forums (like my *uname -a *statement in post #5) makes everything difficult to interprete.

In any case there is one thing I am wondering about from your first post. It seems that you back up to a file system on a removable medium, but that you have a non Linux file system type there. When that is correct, how do you ever think that restoring from there will give you everything that does not exist (like ownership of user/group and access-bits) on an NTFS file system, back?

For the rest of your post, I think some of us need some time to understand this, especialy as all the logical screens seem to have a differnet usage in SLES then in openSUSE.

#9

So sorry, it is true I have no way of copy-paste the SUSE screens. Would it help to enclose in

(if that is how you do it).

Very grateful if you can have patience with me !!

As to the NTFS formatted backups. You are probably right. So we would need to create a reiser-formatted partition on the USB attached device? Note one of the partitions is Netware NSS. Also the backup package blurb says you cannot restore particular files (meaning selected?) from ReiserFS3 upwards, only ext2/3/4 so this software probably won’t be any use anyway…ho hum.

Right, also just restarted the server, selected Rescue, tried and failed to mount the first partition temporarily, and ran fsck.reiserfs --check /dev/sda1
This is what it says


resierfs_open: the reiserfs superblock cannot be found on /dev/sda1.
Failed to open the filesystem
If the parttition table has not been changed, and the partition is valid and it really contains a reiserfs, then the superblock is corrupted and you need to run this utility with --rebuild-sb

I am not going to do that yet !

#10

On the fs type. As said, I do not know the backup utility you use, nor what it does, but I doubt (that is an understatement from me again) that using an fs type that is in fact completely strange to the OS will help in making good backups. And when you use Reiserfs in general, Reiserfs seems to me the correct choice. Reiserfs is still supported on openSUSE, but I think most of us have long converted to ext3 and because ext4 is the default from 11.2 (or earlier?) that is standard now. But as long as your SLES system is using Reiserfs as default, I would advice to stick to it. I do not know anything about Netware NSS, thus can not advicse on that.

On the error from fsck. That makes me very pessimistic. OTOH, are you sure there must be a Reiserfs on it? /dev/sda1 is nowhere used in your earlier information (mounts). To me it is not clear what is there and how it is used.

Most of us are no professionals (I was one), but the documentation your system manager made about how the system is installed, changed, working, etc. is scanty. Professional organisations are gone bancrupt on such lousy behaviour. When you by any change are trying to get this running within a reasonable amount of time for a commercial company, I would advice to hire a Linux/SLES prof from not to far away that can realy come to your site and see what this is all about. Not that we will not try to help you, but not being able laying hands on the system makes it realy difficult.

#11

On 2010-10-14 12:36, davidm01 wrote:
>
> Ok, have talked to one of the two people who were doing this, and am
> getting a moderately good picture of what decisions have been made, and
> what has happened. I should emphasise right away that what caused us to
> get into the situation we are may or may not get us back to a working
> system, but here goes.
>
> This was an experienced and loyal Novell user. To move forward, it was
> wanted to continue with a familiar system if possible, and the nearest
> was SLES netware. We wanted simple user rights management, nothing much
> except file and iPrint services, NSS file system and Salvage. Extensive
> attempts to achieve this with SUSE 11.1 and 11.2 failed to support
> Netware. Success was achieved with SUSE 10.2.0.91, and there was a good
> system running tests for weeks. -There are weeks of effort in this
> system concerning setting up users, data, facilities etc. All this is
> going to waste right now.-

Netware is a commercial product of Novell, and I’m not familiar with it, except it existence. I had
to look it up in the wikipedia. I don’t see “SLES netware” there.

What you have might be this:

http://en.wikipedia.org/wiki/Netware
+++********
Open Enterprise Server
Main article: Novell Open Enterprise Server
2.0

OES 2 was released on October 8, 2007. It includes NetWare 6.5 SP7, which supports running as a
paravirtualized guest inside the Xen hypervisor and new Linux based version using SLES10.
********+±

Or this:

http://en.wikipedia.org/wiki/Novell_Open_Enterprise_Server

+++********
Novell Open Enterprise Server (OES) is the successor product to Novell, Inc.'s NetWare operating
system, based on Suse Linux Enterprise Server. Originally released in March 2005, the current (2009)
release is OES 2 SP2.

Summary

Novell Open Enterprise Server (OES) is best thought of as a platform for delivery of -level shared
network services (file, print, directory, clustering, backup, storage management, PKI, web
applications, etc.) and common management tools. OES can run atop either a Linux or a NetWare
kernel. Clustered configurations can include nodes with either kernel types, and most services can
migrate freely between the platforms. Thus, customers can deploy the platform selection that best
suits their needs, as opposed to being locked into a single platform.

OES-Linux

When installed using a Linux kernel, the product is known as OES-Linux. This uses SUSE Linux
Enterprise Server (SLES) as its platform. Atop the SLES install, daemons are added to provide NCP,
eDirectory, NSS, iPrint and other services delivered by OES.

OES-NetWare

When installed using a NetWare kernel, the product is known as OES-NetWare. This uses NetWare v6.5
as its platform. Atop the NetWare install, NLMs are added to provide Apache web server, Tomcat,
OpenSSH, NCP, eDirectory, NSS, iPrint and other services delivered by OES.

********+±

I have absolutely no experience with this setup (and I don’t have clear which one of those it is) -
you have to ask your vendor. At least, ask on the Novell SLES/SLED forums, not the openSUSE forums.

In your first post you mentioned “SUSE OES netware server”, which means the OES-NetWare above ->
netware kernel. This has nothing to do with these forums, you can not ask us to help you if that is
what you have.

Thus I’ll assume you have OES-Linux as above description.

> Before going live, the final tests of imaging and daily backup were
> begun, using Acronis Backup & recovery 10 Advanced server. This needs
> to be installed under yast in order to be scheduled, which was done.
> Attempts were made to backup to a USB drive, one could set the drives to
> backup and set the drive to backup to, but execution failed with errors
> ‘could not create archive’. Discovering a need to run linux software to
> enable writing to NTFS formatted drives, so this was installed.

Acronis is a comercial product with which I’m not familiar. I can’t help.

About NTFS writing, I have to assume that you got it from official channels…

In any case, using ntfs to backup some kind of linux system, is generally a mistake.

> A successful write of a small test file was completed, so the main
> server drives backup was begun, resulting in the same sort of ‘could not
> create archive’ error as before, plus other critical errors such as
> 0x00000FF1 and 0x0000FF2. FF1 seems to have to to with permissions. FF2
> is unknown. At this point the server was frozen, and after a hard reset
> could not be restarted. Which is where we are now. The server seems to
> be full, the partitions don’t seem to be completely described.

I can’t even give educated guesses.

> I know just enough to run a few listing commands. Here are the
> results.

There are partitions, of a type I never used. And two raids (dm-0, 1)

> device boot start end blocks Id system
> /dev/sda1 * 1 9 72261 83 Linux
> /dev/sda2 19 1966 15719602+ 8e Linux LUM
> /dev/sda3 1967 88849 697887697+ 65 Novell Netware 386

Ufff.

> Right now, the console screens say
> F1 screen
> about 30 lines ending
> Found a Linux console termial on /dev/console…
> *** Starting YAST2 ***

You are running the install CD, I guess. You did not say.

Sorry, I can’t help you.

–
Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)

#12

Henk - I entirely take your points. I am a part-time systems man myself, some Novell. The company is quite small and my advice (similar to yours) as to how to move forward was actually to move to MS 2008 server, but that was not adopted on grounds of cost (not only the server hardware, OS licences but also a lot of PC’s still being Win98 meant they could not attach…wow. ).

So sticking with Novell in some shape or form seemed sensible in some ways. As you have possibly guessed, it was up and working very well for quite a long time. Now this…

I am away until next Monday now. I hope to resume with you then. I can then boot the server from the CD, enter the install option and then select Repair Installed system and finally the automatic repair. This has been done, apparently, and I understand it gives more information that might be useful, such as the expert partitioning tool will tell me exactly what type of partitions are there. See you then…

many thanks for your attention so far.

#13

On 2010-10-14 14:36, hcvv wrote:

> On the error from fsck. That makes me very pessimistic. OTOH, are you
> sure there must be a Reiserfs on it? /dev/sda1 is nowhere used in your
> earlier information (mounts). To me it is not clear what is there and
> how it is used.

It is a small boot partition - educated guess.


> fdisk -l
> Disk /dev/sda: 770.0Gb 7308....08 bytes
> 255 heads, 63 sectors/track, 88849 cylinders
> Units = cylinders 0f 16065 * 512 = 8225208 bytes
>
> device  boot  start  end   blocks Id system
> /dev/sda1 *  1  9            72261  83  Linux

72261 blocks. How big is a block, 512 bytes? Then it would be about 35 megs. Or 1 K? Then it would
be 70 megs. That makes it impossible to be reiserfs, it will be ext2.

Then the system partition is probably that “Linux LUM” - what is that?

–
Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)

#14

This could be a simple out of disk space problem. a df command will show partition usage. Many backup methods use /tmp to marshal/compress the data before moving it to the finial destination.

It would help to know exactly what copy/backup methods/programs where used in that last try. Also exactly what version of Suse.

#15

@Carlos. Good guess about that /boot partition. But I am not sure that in the old days when Reiserfs was normaly used, the boot partition had to be ext2. I assume GRUB can function with Reiserfs.

That partition type 8e means Unix LVM (Logical Voluime Manager). That shows how trustworthy read and type instead of copy and paste is >:( .

I must admit that I do not understand much of the output of his df -h and his “mounts”. But as he does not tell what he is doing (how can anything of those disks be mounted when he booted from a CD when he did not do this manualy?) all speculation is futile.

#16

On 2010-10-14 17:36, hcvv wrote:
>
> @Carlos. Good guess about that /boot partition. But I am not sure that
> in the old days when Reiserfs was normaly used, the boot partition had
> to be ext2. I assume GRUB can function with Reiserfs.

No, because I happen to know that minimum size for a reiserfs is 100 M, and that one has 70 >:-)

And no, because since ever ext2 is the partition type of choice for /boot (if it exists). WHen
reiserfs came, you could have / as reiserfs (or ext2), or with a separate boot, which in that case
it was always ext2 - except for experiments. It is the best choice - why? Because no journal means
less space “wasted” (it is already small), and small means ext2 checks in seconds, no need for a
journal.

> That partition type 8e means Unix LVM (Logical Voluime Manager). That
> shows how trustworthy read and type instead of copy and paste is >:( .

AH!!! LVM, of course. X’-)

> I must admit that I do not understand much of the output of his df -h
> and his “mounts”.

Because it was done from a live CD. Notice yast running in console 1 or there abaouts, and the logs
in others… so it is the install CD running the install… I hope they did not install again.

> But as he does not tell what he is doing (how can
> anything of those disks be mounted when he booted from a CD when he did
> not do this manualy?) all speculation is futile.

I guess it is the installation disk, stopped to get to a VT with bash and run those commands for us,
and copy paste with pencil and paper :slight_smile:

On 2010-10-14 16:36, gogalthorp wrote:
>
> This could be a simple out of disk space problem. a df command will
> show partition usage. Many backup methods use /tmp to marshal/compress
> the data before moving it to the finial destination.

Indeed.

Or that the destination disk was not mounted, or that the destination path was not correctly
written. Something of that sort, 95% sure.

–
Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)

#17

On 2010-10-14 16:06, davidm01 wrote:
>
> Henk - I entirely take your points. I am a part-time systems man
> myself, some Novell. The company is quite small and my advice (similar
> to yours) as to how to move forward was actually to move to MS 2008
> server, but that was not adopted on grounds of cost (not only the server
> hardware, OS licences but also a lot of PC’s still being Win98 meant
> they could not attach…wow. ).
>
> So sticking with Novell in some shape or form seemed sensible in some
> ways. As you have possibly guessed, it was up and working very well for
> quite a long time. Now this…

Notice that you are asking for help in the wrong forum.

We are trying, but this is not our “field”. This is openSUSE, but you are using one of the
propietary products, and a special one.

You have to ask in specific forums for your product - it is the same server (nntp), but different
forums. You have to look the novel.support.* forums, and I see there are a few specific for iprint,
netware, and such. It is there where they know what you are using and what those errors mean. I do
not know how active they are, whether novel staff attends there or only users (like us here), or a mix.

Please don’t give up without asking there first.

–
Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” at Telcontar)

#18

You can boot the machine off a “live” CD and then run a graphical utility that displays the partitions and how full they are. If the main partition is too full now, you can mount it in the filesystem of the live cd and delete the files in /tmp directory to enable a boot, or take other measures depending on what the arrangement of the partitions is. It needs a screenshot I suppose to get a good picture back here.

Many CDs will facilitate this. One good one is the Ubuntu live CD or the Knoppix CD.

#19

Ok, back again. Unfortunately on my own for at least two days - that’s me, who had nothing whatever to do with installing this system !!

First let me encourage you to believe that you are the best people to work this one out, because we ARE talking about OES-linux - that is

When installed using a Linux kernel, the product is known as OES-Linux. This uses SUSE Linux Enterprise Server (SLES) as its platform. Atop the SLES install, daemons are added to provide NCP, eDirectory, NSS, iPrint and other services delivered by OES.

Now the problem **must **be entirely confined to the SUSE linux booting process, and I am confident you can feel at home there !

Your guesses in the last few posts were spot on. I am booting from the install CD and then doing various things to try to determine the problem, like starting a rescue system, or dropping to a bash shell and so on - as you have guessed correctly.

SUSE linux Enterprise server 10 SP3 Kernel 2.6.16.60-0.54.5-smp
partition 1 is ext2
partition 2 is LVM
partition 3 is a Novell NSS-formatted volume

Right now, I have booted the recovery system, chosen Install, other, and am looking at the Yast screen which offers (1) automatic repair (2) customised repair and (3) expert tools.

I have chosen option 3, which offers
Install new boot loader
Start partitioning tool
Repair file system
Recover lost partitions
and a couple of irrelevant options.

Looking at the partitions first
/dev/sda1 70.5Mb Linux native (no mount point or mount by)
/dev/sda2 14.9Gb Linux LVM (no mount point or mount by)
/dev/sda3 665.5Gb Novell netware
/dev/system 14.9Gb LVM2 system (no mount point or mount by, and no start or end either)
/dev/system/root 10.0Gb LV (no mount point or mount by, and no start or end either)
/devsystem/swap 2.0Gb LV (no mount point or mount by, and no start or end either)

I am reluctant to make ANY changes to the system without knowing what I am doing…your advice will be extremely valuable.

The problem seems to be that a volume **may **be full, or fstab file corrupt, or MBR bootloader over-written.

What can I do to help you detect which it is ?

#20

Of course nothing from sda is mounted, because you are running a live CD system. Any mounting must be done by hand.

I would suggest that you try to mount sda1, so we can see if it is corrupted and/or has the correct contents (at least we hope we will be able to come to a conclusion here).

Thus:

mount /dev/sda1 /mnt

and we want of course you to report any output from that. If it fails you could go to a second step by explicitly specifying it is ext2 (though the system should be able to detect that itself):

mount -t /dev/sda /mnt

same again, please report everything and when possible without typos rotfl!

When mounting succeeds, please do

cd /mnt
ls -l

and report that output.