[SOLVED] slack13.1 nfsd fatal error

lpallard · 06-27-2010, 09:43 AM

I really wonder how stable I can pretend slackware 13.1 is.... Im getting my share of problems with an "out-of-the-box" supposely "stable" release of the OS...

anyways, I'm trying to start the NFS daemon on the server and I get:

Code:

bash-4.1# /etc/rc.d/rc.nfsd start
FATAL: Error inserting nfsd (/lib/modules/2.6.33.4-smp/kernel/fs/nfsd/nfsd.ko): Invalid module format
Starting NFS server daemons:
  /usr/sbin/exportfs -r
exportfs: /etc/exports:2: syntax error: bad option list
  /usr/sbin/rpc.rquotad
  /usr/sbin/rpc.nfsd 8
  /usr/sbin/rpc.mountd

Now I understand I'm not the only one to have that problem... apparently something to do with a kernel problem or something like that. I've found a website where the site maintainer reported the same exact problem, but his solution worked for him, not for me (of course!)... See http://bluecottagealmonds.com/

Basically, he says that its a problem with modprobe or init and reverting to the older version 3.6 would do it. Here's what he did:

He downloaded & "upgraded" module-init-tools-3.6-i486-2.txz

Code:

upgradepkg module-init-tools-3.6-i486-2.txz

Then:

Code:

/usr/src# depmod -a
/etc/rc.d/rc.nfsd stop
/etc/rc.d/rc.nfsd star

t

And that apparently worked. For me, I get the same error as before...i.e. FATAL.....

Alien Bob · 06-27-2010, 01:29 PM

Are you still running a "huge" kerenel? If so, start using a generic kernel (including an initrd) and then check again.

About the "exportfs: /etc/exports:2: syntax error: bad option list" messge - you did not post your exports file but the error indicates that you made a syntactical mistake in there.

Eric

lpallard · 06-27-2010, 01:41 PM

Eric, thanks for replying!

1- when I installed slack, I only press ENTER when the setup ask for a kernel, and then installed like all the other computers.... except my laptop where I compiled a custom kernel last year (when I was running 12.2), I always kept the default kernels.

typing Uname -a will give:

Code:

Linux lhost2 2.6.33.4-smp #2 SMP Wed May 12 22:47:36 CDT 2010 i686 AMD Athlon (tm) 64 X2 Dual Core Processor 4600+ AuthenticAMD GNU/Linux

So does it means that I don't have the proper kernel running?

For the /etc/exports, I've found the typo... I put an underscore instead of a , between the options....

Alien Bob · 06-27-2010, 02:03 PM

Slackware only installs a huge kernel, you'll have to add a generic kernel to lilo.conf yourself.

Looking at your two posts, you have the matching kernel and kernelmodules instaleld (2.6.33.4 SMP) but the huge kernel is notorious (was like that in previous Slackware releases too) for refusing to load some kernel modules. The generic kernel will not have that issue.

Eric

lpallard · 06-27-2010, 02:23 PM

Quote:

you'll have to add a generic kernel to lilo.conf yourself.

How do I do that?

Doing so will it screw up everything else? partitions, files, settings, etc??

samac · 06-27-2010, 02:44 PM

Have a look for the file README.initrd in the root directory of your install CD/DVD. All is explained there. If you follow the intructions you will not screw up anything. Remember to run /sbin/lilo after adding a new stanza to /etc/lilo.conf

samac

lpallard · 06-27-2010, 07:57 PM

Eric, what you described was 100% accurate and correct. many thanks!!

samac, also thank you for your help, the README.initrd file was helpful...

Thanks guys!

roragnarok · 10-20-2010, 09:01 PM

hi all,
Same problem here, but as I tried to mkinitrd after reboot it will stuck in the vmware workstation complaining about unable to mount /mnt
do you know what is wrong ?

lpallard · 10-20-2010, 09:12 PM

If I understand correctly, it stops whenever it tries to load the vmware module? If so, have you included all necessary modules for vmware in your initrd? you can use mkinitrd.conf in /etc/ or within the mkinitrd command.

Is /mnt located on a special device? Raid? reiserFS?

roragnarok · 10-20-2010, 10:06 PM

hi lpallard thanks for reply.
It is a virtual hdd sda2 partition in ext4 format.
http://img194.imageshack.us/img194/1...rdnodevice.png

lpallard · 10-21-2010, 05:35 PM

I looked at your screenshot, classic. it happened to me 100 times... my guess is that you miss a (or some) drivers in your initrd. Which kernel are you using?

basically, udev tries to find the media where your / partition is located but cant mount due to missing driver (fs, disk controller, etc). not having / mounted successfully leads to troubles after the initrd image has been loaded and will halt there for sure.

You mentioned something about a virtual hard drive on /dev/sda2... sda is your first SATA HDD. How did you create this "virtual drive"?

roragnarok · 10-21-2010, 11:37 PM

I was wondering which features or modules of the kernel which I'm missing to leading into this trouble.
I'm using the stock kernel + the update "[slackware-security] 64-bit kernel (SSA:2010-265-01)"
It is as the default choice for virtual hard drive using SCSI.
I heard some saying vmware workstation isn't officially supported by slackware.
But I'm a huge fan of Slackware64, and I'm wondering why isn't NFSD module loading successfully for slackware64 13.1.
The Slackware64 13.0 however doesn't have this problem..

Thanks~