shfs-user — General mailing list.

[Shfs-user] Reading >4K files locks shfs mounts since 2.6.19.*

[<zou...@gm...>]

Since I installed the 2.6.19, 2.6.19.1 and 2.6.19.2 kernels, I've had problems with my shfs mounts.
My friend (different architecture, different software, but 2.6.19) has exactly the same problem.

Reading >4K files locks the mount.

example :

mount -t shfs machine: /mnt
cd /mnt
dd if=/dev/zero of=pouet bs=1M count=10
less pouet 
# lock





PS : I hope the project isn't going to be abandonned since it has obvious advantages over sshfs
 (no need for sftp..)

[Shfs-user] amarok + shfs = no metadata on mp3s

[Tejing <tej...@ya...>]

I originally posted this as a bug on the amarok bug tracker, since other
mp3 players had no issue, but the bug was closed as invalid, and I was
told to send this your way. Here's the text of the bug I sent in to amarok:

I was attempting to play some mp3 files from my sister's computer, using
an shfs mount, and amarok played them, but no information about the file
was available (id3 tag info, bitrate, even the duration was unknown).
the very same files showed full information when I copied them to my
local machine, and they showed full information from either location
with several other programs that read id3 tags (I checked with id3v2
(the command line tool) and mplayer). I haven't tested any other network
filesystems, so it may or may not be shfs-specific.

a little help telling whether/why this is an amarok bug, or if it is
indeed caused by shfs, would be appreciated.

Tejing
-- 
There are 10 types of people in the world: Those who understand binary
and those who don't.

-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GCS/M/P d- s++: a19 C++++>$ UL++++ P+ L+++(++++) E++ W+ N O? o? K?
w--- M- V-- PS(--) PE+() Y+ PGP++ t+ 5? X+ R+ !tv b++(++++) DI+ D G++
e* h!>++ !r !y-
------END GEEK CODE BLOCK------

[Shfs-user] Is it possible to get the same speed on shfs as with scp?

[Mikael S. <mi...@po...>]

If I do

scp -c blowfish -C -o Compression=no us...@ip...:/path/bigfile.bin .

I get over 1mb/s a second on my ADSL2+ connection to a specific host.



Using shfs with the command:

shfsmount -t perl --cmd="ssh %u@%h -x -c blowfish -C -o Compression=no 
/bin/bash" us...@ip...:/path /mnt/smb -o cachesize=256 cachemax=20

and cp /mnt/smb/bigfile.bin

I get ~650kb/s for the same file. I've tried both with and without perl 
mode.

How can I tune cachesize and cachemax to get the same speed as scp 
(which should be possible??) Or is there some other wat to improve the 
speed? The remote machine is a Quad PIII 550, which uses 50% cpu on the 
ssh connection when I copy the file.

Regards,
-m

[Shfs-user] My timestamps are all wrong

[Du <ed...@mi...>]

Hi folks, when I create a file under SHFS and do 'ls -l'  on it, it show
me the timestamp almost right, but the seconds are alway equal zero,
Does anybody have an idea???


Thanks!!!

Re: [Shfs-user] v0.35 modified to work with 2.6.17 & 2.6.18

[dsred <ds...@gm...>]

runs perfekt, thank you very much,
dsred


Steve Redler IV wrote:
> 
> For those interested, I've been fixing shfs as the kernel changes, heres a
> link to a few of my
> recent versions. 
> 
>   http://server.linuxsys.net/oss/shfs/
> 
> --
> Regards,
> Steve Redler IV, SR Technology
> Tcl2006 Conference Program Chairman, http://www.tcl.tk/community/tcl2006/
> 
> 
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share
> your
> opinions on IT & business topics through brief surveys -- and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> _______________________________________________
> Shfs-user mailing list
> Shf...@li...
> https://lists.sourceforge.net/lists/listinfo/shfs-user
> 
> 

-- 
View this message in context: http://www.nabble.com/v0.35-modified-to-work-with-2.6.17---2.6.18-tf2362328.html#a6582053
Sent from the Shfs - General mailing list archive at Nabble.com.

[Shfs-user] v0.35 modified to work with 2.6.17 & 2.6.18

[Steve R. I. <st...@sr...>]

For those interested, I've been fixing shfs as the kernel changes, heres a link to a few of my
recent versions. 

  http://server.linuxsys.net/oss/shfs/

--
Regards,
Steve Redler IV, SR Technology
Tcl2006 Conference Program Chairman, http://www.tcl.tk/community/tcl2006/

Re: [Shfs-user] bug in inode.c

[Tranzmatt <tra...@gm...>]

> Hi shfs-users
>   since I have installed kernel 2.1.14 (FC4) shfs is not more
> stable. Both with shfs-0.35 and cvs version I get segfault and
> even complete system blockage. At the end of the mail
> there is an extract from /var/log/messages.
> I suppose that something is changed in the kernel. Any idea?

I'm getting something similar with shfs-0.35 on my 2.6.17-1.2142_FC4.
Mounting and accessing the partition are fine, but when I try to
shfsmount, it blows up:

Aug 16 09:28:21 gollum kernel:  <0>------------[ cut here ]------------
Aug 16 09:28:21 gollum kernel: kernel BUG at fs/inode.c:251!
Aug 16 09:28:21 gollum kernel: invalid opcode: 0000 [#4]
Aug 16 09:28:21 gollum kernel: last sysfs file: /block/hdc/size
Aug 16 09:28:21 gollum kernel: Modules linked in: shfs(U) nls_utf8
ivtv_fb(U) i915 drm parport_pc lp parport autofs4 rfcomm l2cap
bluetooth sunrpc dm_mod video button battery ac ipv6 uhci_hcd ehci_hcd
msp3400 saa7127 saa7115 tda9887 tuner v4l2_common ivtv(U) i2c_algo_bit
v4l1_compat tveeprom videodev i2c_i801 i2c_core snd_intel8x0
snd_ac97_codec snd_ac97_bus snd_seq_dummy snd_seq_oss
snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss
snd_pcm snd_timer snd soundcore snd_page_alloc e1000 ext3 jbd ata_piix
libata sd_mod scsi_mod
Aug 16 09:28:21 gollum kernel: CPU:    0
Aug 16 09:28:21 gollum kernel: EIP:    0060:[<c046dfd4>]    Tainted: GF     VLI
Aug 16 09:28:21 gollum kernel: EFLAGS: 00210202   (2.6.17-1.2142_FC4 #1)
Aug 16 09:28:21 gollum kernel: EIP is at clear_inode+0x28/0x10d
Aug 16 09:28:21 gollum kernel: eax: cf73edc0   ebx: cf73edc0   ecx:
00000000   edx: c8f41000
Aug 16 09:28:21 gollum kernel: esi: d0c6a845   edi: c0a688a4   ebp:
00000000   esp: c8f41ec0
Aug 16 09:28:21 gollum kernel: ds: 007b   es: 007b   ss: 0068
Aug 16 09:28:21 gollum kernel: Process shfsumount (pid: 25904,
threadinfo=c8f41000 task=cba15000)
Aug 16 09:28:21 gollum kernel: Stack: cf73edc0 d0c6a845 c046e62d
cf73edc0 c0a6889c c046d624 c0a688d8 c046c207
Aug 16 09:28:21 gollum kernel:        00000001 c0a688d0 c67b5308
c0a6889c c06e1e7c c046c331 cf488e64 00000001
Aug 16 09:28:21 gollum kernel:        cf488ea8 cf488e64 c7a33c00
cf488e64 d0c73780 c8f41000 c045d425 00000018
Aug 16 09:28:21 gollum kernel: Call Trace:
Aug 16 09:28:21 gollum kernel:  <d0c6a845> shfs_delete_inode+0x0/0xd7
[shfs]  <c046e62d> generic_delete_inode+0xa7/0x113
Aug 16 09:28:21 gollum kernel:  <c046d624> iput+0x64/0x66  <c046c207>
prune_dcache+0x170/0x1a8
Aug 16 09:28:21 gollum kernel:  <c046c331>
shrink_dcache_parent+0x15/0x141  <c045d425>
generic_shutdown_super+0x21/0x152
Aug 16 09:28:21 gollum kernel:  <c045d58c> kill_anon_super+0x9/0x30
<c045d753> deactivate_super+0x7c/0x91
Aug 16 09:28:21 gollum kernel:  <c0470cf2> sys_umount+0x22a/0x234
<c043a8c0> audit_syscall_entry+0x104/0x12b
Aug 16 09:28:21 gollum kernel:  <c0470d07> sys_oldumount+0xb/0xe
<c0402bb3> syscall_call+0x7/0xb
Aug 16 09:28:21 gollum kernel: Code: 5b 5e c3 56 ba f8 00 00 00 53 89
c3 b8 4d f8 60 c0 e8 f1 8f fa ff e8 ae c5 17 00 89 d8 e8 fc dc fe ff
83 bb 2c 01 00 00 00 74 08 <0f> 0b fb 00 4d f8 60 c0 8b 83 c4 01 00 00
a8 10 75 08 0f 0b fc
Aug 16 09:28:21 gollum kernel: EIP: [<c046dfd4>]
clear_inode+0x28/0x10d SS:ESP 0068:c8f41ec0

This 2.6.17-1.2142_FC4 came "stock" (via yum) with no twiddles.  Also,
I had to patch the shfs-0.35 with the d_entry-2.6.16.diff and
gcc4-compilefix.patch

Any ideas?

Thanks

Matt

[Shfs-user] mounting problem: cannot see contents except broken symlinks

[Paul G. <pau...@gm...>]

Greetings,

I am having trouble using shfs to connect from my linux box (Mandriva
2006) to Solaris.  I have previously set up passwordless
authentication between these machines.  When I mount without the "-s"
option, the only items I can see are broken symlinks:

> shfsmount pvg@amachine:/users/pvg mnt
> ls mnt
ls: -> courses/cs120F06: No such file or directory
ls: -> /local/www/homepages/pvg: No such file or directory

Those two items are the destinations of two symlinks on amachine:/users/pvg.

If I mount with the "-s" flag, then I see nothing at all on an ls.

Verbose output on shfsmount doesn't show anything abnormal as far as I
can tell.  Can someone give me a hand on this?  (I'd appreciate a
direct CC since I'm not a subscriber on the list.)

Thanks,
 -Paul

[Shfs-user] problem to compile shfs on debian "sarge"

[Raphael B. <li...@gm...>]

Dear Users, 

I have problems to compile shfs on debian "sarge". I installed the
"usual" deb-packages on any testing-server... 

I don't have any idea, what's going wrong... I compiled many times this
source, but it seems with kernel 2.6.15 and shfs 0.35-6 it doesn't work
anymore... 

Any ideas? Or also positive feedback?

Thanks a lot...

here is the compile-log:

...
dh_clean
make -C Linux-2.6 clean
make[1]: Entering directory `/usr/src/modules/shfs/Linux-2.6'
rm -rf linux-2.6.15 linux-2.6.15.orig;
rm -f linux-2.6.15.diff
rm -f *.o *.ko *.mod.c .*o.cmd
make[1]: Leaving directory `/usr/src/modules/shfs/Linux-2.6'
/usr/bin/make  -f debian/rules kdist_clean kdist_config binary-modules
make[1]: Entering directory `/usr/src/modules/shfs'
dh_clean
make -C Linux-2.6 clean
make[2]: Entering directory `/usr/src/modules/shfs/Linux-2.6'
rm -rf linux-2.6.15 linux-2.6.15.orig;
rm -f linux-2.6.15.diff
rm -f *.o *.ko *.mod.c .*o.cmd
make[2]: Leaving directory `/usr/src/modules/shfs/Linux-2.6'
/usr/bin/gcc-4.0
for templ in /usr/src/modules/shfs/debian/shfs-module-_KVERS_.postinst /usr/src/modules/shfs/debian/shfs-module-_KVERS_.postinst.backup /usr/src/modules/shfs/debian/shfs-module-_KVERS_.postinst.modules.in; do \
    cp $templ `echo $templ | sed -e 's/_KVERS_/2.6.15/g'` ; \
  done
for templ in `ls debian/*.modules.in` ; do \
    test -e ${templ%.modules.in}.backup || cp ${templ%.modules.in} ${templ%.modules.in}.backup 2>/dev/null || true; \
    sed -e 's/##KVERS##/2.6.15/g ;s/#KVERS#/2.6.15/g ; s/_KVERS_/2.6.15/g ; s/##KDREV##/neptun.0.7/g ; s/#KDREV#/neptun.0.7/g ; s/_KDREV_/neptun.0.7/g' < $templ > ${templ%.modules.in}; \
  done
dh_testdir
dh_testroot
dh_clean -k
env MODVERSIONS=detect make -C Linux-2.6 KERNEL_SOURCES=/lib/modules/2.6.15/source KERNEL=linux-2.6.15
make[2]: Entering directory `/usr/src/modules/shfs/Linux-2.6'
make -C /lib/modules/2.6.15/source SUBDIRS=/usr/src/modules/shfs/Linux-2.6 modules
make[3]: Entering directory `/usr/src/linux-source-2.6.15'
  CC [M]  /usr/src/modules/shfs/Linux-2.6/dcache.o
/usr/src/modules/shfs/Linux-2.6/dcache.c: In function 'shfs_invalidate_dircache_entries':
/usr/src/modules/shfs/Linux-2.6/dcache.c:71: error: 'struct dentry' has no member named 'd_u'
/usr/src/modules/shfs/Linux-2.6/dcache.c:71: warning: type defaults to 'int' in declaration of '__mptr'
/usr/src/modules/shfs/Linux-2.6/dcache.c:71: warning: initialization from incompatible pointer type
/usr/src/modules/shfs/Linux-2.6/dcache.c:71: error: 'struct dentry' has no member named 'd_u'
/usr/src/modules/shfs/Linux-2.6/dcache.c: In function 'shfs_dget_fpos':
/usr/src/modules/shfs/Linux-2.6/dcache.c:104: error: 'struct dentry' has no member named 'd_u'
/usr/src/modules/shfs/Linux-2.6/dcache.c:104: warning: type defaults to 'int' in declaration of '__mptr'
/usr/src/modules/shfs/Linux-2.6/dcache.c:104: warning: initialization from incompatible pointer type
/usr/src/modules/shfs/Linux-2.6/dcache.c:104: error: 'struct dentry' has no member named 'd_u'
make[4]: *** [/usr/src/modules/shfs/Linux-2.6/dcache.o] Fehler 1
make[3]: *** [_module_/usr/src/modules/shfs/Linux-2.6] Fehler 2
make[3]: Leaving directory `/usr/src/linux-source-2.6.15'
make[2]: *** [default] Fehler 2
make[2]: Leaving directory `/usr/src/modules/shfs/Linux-2.6'
make[1]: *** [binary-modules] Fehler 2
make[1]: Leaving directory `/usr/src/modules/shfs'
make: *** [kdist_build] Fehler 2

THE END!


other system-infos:
- Debian testing (actual)
- SHFS-Util-Version 0.35-6 (deb-package)

Re: [Shfs-user] shfs + chroot does not work

[Tomas Pospisek's M. L. <tp...@so...>]

On Tue, 16 May 2006, Arie Huijzer wrote:

> Is shfs (userspace?) possible in combination with chroot?
>
> I want to mount a slow machine over the network to a fast one.
> Chroot and compile some packages with the performance of the fast machine.
> But unfortunately nothn happens after I do " chroot /mnt/P3 /bin/bash"
> Mounting with shfs is no problem chroot with nfs works, so shfs is suspected.
>
> This is the script I use to mount/chroot-------::
>
> sshfs -o allow_other root@192.168.123.119:/ /mnt/P3
> mount -t proc none /mnt/P3/proc
> mount --bind  /var/tmp/portageP3  /mnt/P3/var/tmp/portage
> chroot /mnt/P3 /bin/bash
> env-update
> source /etc/profile

As far as I understand shfs allows only one simultaneous file access (at a 
time). If there's another process (or thread for that matter) trying to 
access a different file at the same time then it will block untill the 
first file operation is done. Thus I wonder whether you won't run into 
such a problem... (strace would probably tell...).
*t

--
-----------------------------------------------------------
   Tomas Pospisek
   http://sourcepole.com -  Linux & Open Source Solutions
-----------------------------------------------------------

[Shfs-user] shfs + chroot does not work

[Arie H. <ari...@gm...>]

Hello,

Is shfs (userspace?) possible in combination with chroot?

I want to mount a slow machine over the network to a fast one.
Chroot and compile some packages with the performance of the fast machine.
But unfortunately nothn happens after I do " chroot /mnt/P3 /bin/bash"
Mounting with shfs is no problem chroot with nfs works, so shfs is suspecte=
d.

This is the script I use to mount/chroot-------::

sshfs -o allow_other root@192.168.123.119:/ /mnt/P3
 mount -t proc none /mnt/P3/proc
 mount --bind  /var/tmp/portageP3  /mnt/P3/var/tmp/portage
 chroot /mnt/P3 /bin/bash
 env-update
 source /etc/profile
--------- end of script.

[Shfs-user] Re: .35 won't compile on 2.6.16

[Carlos <cfc...@re...>]

I'm having sane problem, the only clue i've found  is in:

http://shfs.sourceforge.net/install.html

in relation with the difference in gcc and kernel compliler

The compiler:
gcc (GCC) 3.4.5 (Gentoo 3.4.5-r1, ssp-3.4.5-1.0, pie-8.7.9)

Kernel compiler
Linux version 2.6.16-gentoo-r6 (root@livecd) (gcc version 3.4.4 (Gentoo
3.4.4-r1, ssp-3.4.4-1.0, pie-8.7.8)) #2 SMP PREEMPT Mon May 8 15:23:00 COT 2006

but can't find the way out

[Shfs-user] persistent & ssh-askpass

[Benko <ben...@gm...>]

hello!

i'm trying to automatically mount a partition on a server to a client
without X. as the server is a failover cluster the mount could be
interrupted in case of a failover from the first to the second node
(both share the same ip). 
in this case, shfs would fail on the client
and a remount would be required. 

luckily, there's the "persistent"-option in
shfs that should bypass such a situation, unfortunately it doesn't seem
to work for my setup which should run without human intervention. 

when i add persistent to the options in the fstab-line where i've
already configured the server-partition and ip, i receive a error
message when i mount the shared partition, saying that i'm missing
ssh-askpass. since only version of ssh-askpass(x11-ssh-askpass resp.
gtk2-ssh-askpass) which require a X-server seem to be available, i'm
really out of luck here. 

is there a way to have a persistent shfs-mount
with authorized_keys-entries or keychains possible? thanks for your
input!!!

best regards
christian

[Shfs-user] Re: "persistent" for sshfs

[Marko N. <man...@gm...>]

Sorry, apparently wrong mailing list.

Marko

[Shfs-user] "persistent" for sshfs

[Marko N. <man...@gm...>]

Hi,

is it possible to define that the client should retry the connection
to the ssh server in case there are network failures? This is
apparently done with --persistent with shfsmout but I've only got the
sshfs fuse client.

Marko

[Shfs-user] .35 won't compile on 2.6.16

[John H. <mis...@gm...>]

worked till i upgraded:)

make
make[1]: Entering directory `/tmp/shfs-0.35/shfs'
make[2]: Entering directory `/tmp/shfs-0.35/shfs/Linux-2.6'
make -C //lib/modules/2.6.16/build
SUBDIRS=3D/tmp/shfs-0.35/shfs/Linux-2.6 modulesmake[3]: Entering
directory `/usr/src/linux-2.6.16'
  CC [M]  /tmp/shfs-0.35/shfs/Linux-2.6/dcache.o
/tmp/shfs-0.35/shfs/Linux-2.6/dcache.c: In function
'shfs_invalidate_dircache_entries':
/tmp/shfs-0.35/shfs/Linux-2.6/dcache.c:71: error: 'struct dentry' has
no member named 'd_child'
/tmp/shfs-0.35/shfs/Linux-2.6/dcache.c:71: warning: type defaults to
'int' in declaration of '__mptr'
/tmp/shfs-0.35/shfs/Linux-2.6/dcache.c:71: warning: initialization
from incompatible pointer type
/tmp/shfs-0.35/shfs/Linux-2.6/dcache.c:71: error: 'struct dentry' has
no member named 'd_child'
/tmp/shfs-0.35/shfs/Linux-2.6/dcache.c: In function 'shfs_dget_fpos':
/tmp/shfs-0.35/shfs/Linux-2.6/dcache.c:104: error: 'struct dentry' has
no member named 'd_child'
/tmp/shfs-0.35/shfs/Linux-2.6/dcache.c:104: warning: type defaults to
'int' in declaration of '__mptr'
/tmp/shfs-0.35/shfs/Linux-2.6/dcache.c:104: warning: initialization
from incompatible pointer type
/tmp/shfs-0.35/shfs/Linux-2.6/dcache.c:104: error: 'struct dentry' has
no member named 'd_child'
make[4]: *** [/tmp/shfs-0.35/shfs/Linux-2.6/dcache.o] Error 1
make[3]: *** [_module_/tmp/shfs-0.35/shfs/Linux-2.6] Error 2
make[3]: Leaving directory `/usr/src/linux-2.6.16'
make[2]: *** [default] Error 2
make[2]: Leaving directory `/tmp/shfs-0.35/shfs/Linux-2.6'
make[1]: *** [all] Error 2
make[1]: Leaving directory `/tmp/shfs-0.35/shfs'
make: *** [module] Error 2

Re: [Shfs-user] shfs to an internal host?

[Koen V. <k.a...@tn...>]

On Mon, 2006-03-20 at 09:07 +0100, Tomasz Chmielewski wrote:
> If it's not possible, you can probably "redirect" ports on the 
> "intermediate machine" using some programs which bind to one port, and 
> redirect all requests to some other IP/port.
> I don't have the names of such programs, but search for the things like 
> "bouncer" etc.

ssh comes to mind :-) Try:

ssh -L 1234:no_public_IP:22 intermediate_machine

This forwards local port 1234 to port 22 on machine no_public_IP. Test
it by trying to connect to the no_public_IP machine by running

ssh --port=1234 localhost

If that works, just let shfs use localhost:1234 as well, and you should
be in business.

Koen

Re: [Shfs-user] shfs to an internal host?

[Tomasz C. <ma...@wp...>]

Josh Bailey wrote:
> I would like to mount a directory on machine which has no public IP.  I 
> must first connect to an intermediate machine, which has a public IP.  I 
> have only user privileges on the intermediate machine.
> Is this possible?

Can this machine with no public IP connect to the outside world 
(NAT/masquerade)?

If yes, then the easiest would be to build a VPN connection (OpenVPN 
would be just great here) from the machine with no public IP (OpenVPN 
client) to your machine with a public IP (OpenVPN server).

If it's not possible, you can probably "redirect" ports on the 
"intermediate machine" using some programs which bind to one port, and 
redirect all requests to some other IP/port.
I don't have the names of such programs, but search for the things like 
"bouncer" etc.


-- 
Tomasz Chmielewski

Silent software deployment with Samba
http://wpkg.org

[Shfs-user] shfs to an internal host?

[Josh B. <moj...@bi...>]

I would like to mount a directory on machine which has no public IP.  I 
must first connect to an intermediate machine, which has a public IP.  I 
have only user privileges on the intermediate machine. 

Is this possible?

-Josh Bailey

[Shfs-user] Cannot mount fs images when on remote ssh box

[Michael H. <mic...@as...>]

Hi list,

I want to use shfs for my thin clients instead of nfs.
The problem is: when i try to exec /sbin/init (the whole fs is mounted using 
shfs) I see a kernel bug (on qemu, my box and my thin client; fs/inode.c; see 
below).
So I tried to mount a ext2-fs-image on the shfs share to avoid that problem (I 
mount shfs and that contains only one file: an ext2-image).
But when I try to mount the fs-image I just receive this error:
ioctl: LOOP_SET_FD: Invalid argument
I can mount other fs-images, so it is not a loop problem.

Any ideas? I'm waiting hopefully for your answers ;-)

Thanks,

Michael

-----------------------------------
box thinclient # uname -a
Linux box 2.6.15.3-box #2 PREEMPT Fri Feb 17 18:36:52 GMT 2006 i686 AMD 
Athlon(tm) XP 2200+ AuthenticAMD GNU/Linux

SHell File System, (c) 2002-2004 Miroslav Spousta
------------[ cut here ]------------
kernel BUG at fs/inode.c:252!
invalid operand: 0000 [#1]
PREEMPT 
Modules linked in: shfs smbfs ext2 usblp ntfs vfat fat ohci_hcd uhci_hcd 
ehci_hcd usbcore tulip 8139too snd_cmipci snd_opl3_lib snd_hwdep 
snd_mpu401_uart snd_rawmidi
CPU:    0
EIP:    0060:[<c015b86f>]    Not tainted VLI
EFLAGS: 00210202   (2.6.15.3-box) 
EIP is at clear_inode+0x16/0xa5
eax: df0fc1c8   ebx: df0fc1c8   ecx: d654a874   edx: df0fc1c8
esi: e0954ac7   edi: df0fc1c8   ebp: c9830000   esp: c9831edc
ds: 007b   es: 007b   ss: 0068
Process umount (pid: 3794, threadinfo=c9830000 task=cd02e0f0)
Stack: df0fc1c8 e0954ac7 c015c5e7 df0fc1c8 df0fc1c8 d46c2228 c015c830 df0fc1c8 
       c9830000 c0159bf5 df0fc1c8 d46c2228 d654a800 e095bf20 c014b90d d46c2228 
       d654a88c d46c2228 c9830000 0000000f 0804d9af c014c1f2 d654a800 d654a800 
Call Trace:
 [<e0954ac7>] shfs_delete_inode+0x0/0xe2 [shfs]
 [<c015c5e7>] generic_delete_inode+0x6c/0xea
 [<c015c830>] iput+0x63/0x66
 [<c0159bf5>] dput+0x1af/0x1f0
 [<c014b90d>] generic_shutdown_super+0x34/0x119
 [<c014c1f2>] kill_anon_super+0x16/0x42
 [<c014b830>] deactivate_super+0x51/0x67
 [<c015e710>] sys_umount+0x6b/0x73
 [<c013cd2b>] sys_munmap+0x31/0x4b
 [<c015e723>] sys_oldumount+0xb/0xe
 [<c0102651>] syscall_call+0x7/0xb
Code: c7 42 04 04 af 31 c0 89 15 04 af 31 c0 ff 0d e8 af 39 c0 5b c3 56 53 8b 
5c 24 0c 53 e8 39 c7 fe ff 83 bb c8 00 00 00 00 58 74 08 <0f> 0b fc 00 9d 4d 
2d c0 8b 83 24 01 00 00 a8 10 75 08 0f 0b fe

[Shfs-user] Re: system freezes while using shfs

[Tomasz C. <tc...@wp...>]

> server without problems. Sometimes it is impossible to work with it. But 
> the worse problem is that my system sometimes completly crashes/freezes 
> without any error in /var/log/messages or on the screen. The only thing 
> I can do is the reset button. This only happens while I read or write 
> via shfs. It really happens often if I run portsentry and I have no idea 
> why. But it also happens without portsentry. On the other hand I have to 
> say that shfs is a very useful tool (as long as it works)!

Hi,

Same happens here.

Sometimes I can work a couple of days and nothing happens, and sometimes 
the system freezes soon after boot, when I do something like:

cd /mnt/shfs

Keyboard doesn't work, mouse doesn't work, the machine doesn't reply to 
pings... is dead.

Nothing in logs, only reset helps.

-- 
Tomasz Chmielewski
http://lkmi.sourceforge.net

[Shfs-user] system freezes while using shfs

[Tobias B. <to...@to...>]

Hi,

I have big problems with shfs. I mount a server via internet in fstab of 
a Fedora Core release 3. Often connections hang and also the editor or 
script currently using it crash though I have another ssh open to that 
server without problems. Sometimes it is impossible to work with it. But 
the worse problem is that my system sometimes completly crashes/freezes 
without any error in /var/log/messages or on the screen. The only thing 
I can do is the reset button. This only happens while I read or write 
via shfs. It really happens often if I run portsentry and I have no idea 
why. But it also happens without portsentry. On the other hand I have to 
say that shfs is a very useful tool (as long as it works)!

[Shfs-user] Is this possible?

[Mike L. <ml...@ve...>]

Hey everyone,

I haven't yet set up SHFS but wanted some feedback to wether or not the 
following setup is possible.

Is there any way to implement an on-demand SHFS connection for user 
directories?

This is for an IMAP server that is located on a DMZ. The user's mail is 
to be stored on an local intranet server and we were thinking of using 
SHFS to mount the user's mail directory when they connect via IMAP 
(secure IMAP actually).

I guess another way to implement is by using scripts in the MTA.

Mike

[Shfs-user] Re: static declaration of "shfs_dentry_operations" follows non-static declaration

[<jz...@vo...>]

If you'v got this error during compilation it means you are used gcc
4 . In this case you should patch
http://atrey.karlin.mff.cuni.cz/~qiq/src/shfs/shfs-0.35/gcc4-compilefix.pat=
ch.

Jirazy


--=20
Neomezen=E9 vol=E1n=ED na pevn=E9 linky po cel=E9 =C8R za pau=B9=E1l 399 K=
=E8
m=ECs=ED=E8n=EC. Jedin=EC se slu=BEbou VOLN=DD TelefoNet.
http://telefonet.volny.cz

[Shfs-user] Re: shfs fstab problem

[<jz...@vo...>]

Markus Tauchnitz  writes:

>=20
> Hi,
>=20
> I"m using shfs version 0.35 and I want to use shfs-mount with rsa
key  > authentication. There is no problem as long I"m using "
> mount -t shfs -o cmd=3D"ssh -i /USER/.ssh/shfs.key 172.25.1.22"=20
> 172.25.1.22:/raid /home/backup" with a key without a passphrase.=20
> Everything works fine. But how do I note that in /etc/fstab??
> The line
>=20
> USER@...:/raid  /home/backup   shfs  ro,cmd=3D"ssh -i=20
> /USER/.ssh/shfs.key 172.25.1.22" 0   0
>=20
> doesn"t work.  Mount tells that "line xx in /etc/fstab is bad".
It is  > because of the cmd option I"ve noticed. If a miss that
option,=20
> everything works. But I have to pass the identity file (ssh -i)
to ssh,  > because the default rsa key file has a passphrase and I
want to do it  > without user interaction.
> Is there any way, to get it work?
>=20
> Bye,
>=20
> Markus

As olli <olli_vs_lemmings@gm...> writes
>...
> The problem is the spaces in your option cmd, not the cmd-option
> itself. Linux parses the fstab and wherever it finds a space, it
> writes the next variable. ...

Good solution is also write \040 instead of space in fstab-fields.

Jirazy


--=20
Kompletn=EC =F8e=B9en=E9 matematick=E9 testy z p=F8ij=EDmac=EDch zkou=B9ek =
na V=A9:
http://www.zkousky-nanecisto.cz