adnano
asked on
mounting Windows (all version) using CIFS on kernel 2.4
Hi,
It is very important for me to get CIFS VFS running on my system. I have compiled CFS version 1.02b in the kernel (not as a module). I have also compiled mount.cifs from the Samba vers 3.0.2rc2 source and using it. I have also tried it with the mount.cifs that is on the website. All gives me the same result.
When I mount either a Windows 2000 or Windows XP (home or professional server)using the following command my mount command is
mount -t cifs //192.168.1.35/e$ /mnt -o user=administrator,passwor
I get the message that
mount error 22 = Invalid argument
my /var/log/syslog gave me this message
Mar 18 15:57:36 rv2400 kernel: CIFS VFS: Error -32 sending data on socket to server. Mar 18 15:57:36 rv2400 kernel: CIFS VFS: cifs_mount failed w/return code = -5
So I looked more into the documentation and enabled debug (change the
contents of /proc/fs/cifs/cifsFYI to 1).
And I using the same mount command above, I get the same message but the message in /var/log/syslog is
Mar 18 16:47:28 rv2400 kernel: connect.c: CIFS VFS: in cifs_mount as Xid: 4 with uid: 0 Mar 18 16:47:28 rv2400 kernel: connect.c: Username: administrator
Mar 18 16:47:28 rv2400 kernel: connect.c: UNC: \\192.168.1.35\c$
Mar 18 16:47:28 rv2400 kernel: connect.c: Socket created
Mar 18 16:47:28 rv2400 kernel: connect.c: Demultiplex PID: 1394 Mar 18 16:47:28 rv2400 kernel: connect.c: tcp session abended prematurely (after SMBnegprot) Mar 18 16:47:28 rv2400 kernel: connect.c: About to exit from demultiplex thread Mar 18 16:47:28 rv2400 kernel: connect.c: Existing smb sess not found
Mar 18 16:47:28 rv2400 kernel: connect.c: CIFS VFS: leaving cifs_mount (xid = 4) rc = -2 Mar 18 16:47:28 rv2400 kernel: CIFS VFS: cifs_mount failed w/return code = -2
My environment:
Mandrake Linux 9.0
Kernel version: 2.4.19-16mdksmp
Using 3-ware SATA card with 1.3TB on the SATA drives
Oh, the smbfs mounts the drive without any problem. The reason I am using CIFS is because of speed and it doesn't have 2GB file limit.
Thanks a lot in advance
Adnan
It is very important for me to get CIFS VFS running on my system. I have compiled CFS version 1.02b in the kernel (not as a module). I have also compiled mount.cifs from the Samba vers 3.0.2rc2 source and using it. I have also tried it with the mount.cifs that is on the website. All gives me the same result.
When I mount either a Windows 2000 or Windows XP (home or professional server)using the following command my mount command is
mount -t cifs //192.168.1.35/e$ /mnt -o user=administrator,passwor
I get the message that
mount error 22 = Invalid argument
my /var/log/syslog gave me this message
Mar 18 15:57:36 rv2400 kernel: CIFS VFS: Error -32 sending data on socket to server. Mar 18 15:57:36 rv2400 kernel: CIFS VFS: cifs_mount failed w/return code = -5
So I looked more into the documentation and enabled debug (change the
contents of /proc/fs/cifs/cifsFYI to 1).
And I using the same mount command above, I get the same message but the message in /var/log/syslog is
Mar 18 16:47:28 rv2400 kernel: connect.c: CIFS VFS: in cifs_mount as Xid: 4 with uid: 0 Mar 18 16:47:28 rv2400 kernel: connect.c: Username: administrator
Mar 18 16:47:28 rv2400 kernel: connect.c: UNC: \\192.168.1.35\c$
Mar 18 16:47:28 rv2400 kernel: connect.c: Socket created
Mar 18 16:47:28 rv2400 kernel: connect.c: Demultiplex PID: 1394 Mar 18 16:47:28 rv2400 kernel: connect.c: tcp session abended prematurely (after SMBnegprot) Mar 18 16:47:28 rv2400 kernel: connect.c: About to exit from demultiplex thread Mar 18 16:47:28 rv2400 kernel: connect.c: Existing smb sess not found
Mar 18 16:47:28 rv2400 kernel: connect.c: CIFS VFS: leaving cifs_mount (xid = 4) rc = -2 Mar 18 16:47:28 rv2400 kernel: CIFS VFS: cifs_mount failed w/return code = -2
My environment:
Mandrake Linux 9.0
Kernel version: 2.4.19-16mdksmp
Using 3-ware SATA card with 1.3TB on the SATA drives
Oh, the smbfs mounts the drive without any problem. The reason I am using CIFS is because of speed and it doesn't have 2GB file limit.
Thanks a lot in advance
Adnan
did you tried with the precompiled binaries?
er.. I meant... compile as a module, download the binaries and try.
er... reading in the README, I found
Use instructions:
================
Once the CIFS VFS support is built into the kernel or installed as a module
(cifs.o), you can use mount syntax like the following to access Samba or Windows
servers:
mount -t cifs //9.53.216.11/e$ /mnt -o user=myname,pass=mypasswor
after -o the following cifs vfs specific options are supported:
user=<username>
pass=<password>
domain=<domain name>
so, what happend if you use pass= instead password= ?
Use instructions:
================
Once the CIFS VFS support is built into the kernel or installed as a module
(cifs.o), you can use mount syntax like the following to access Samba or Windows
servers:
mount -t cifs //9.53.216.11/e$ /mnt -o user=myname,pass=mypasswor
after -o the following cifs vfs specific options are supported:
user=<username>
pass=<password>
domain=<domain name>
so, what happend if you use pass= instead password= ?
ASKER
Sorry, it was a typo in my question.....
I have tried pass , not password (sorry again)
Moreover, I will try what u guys have suggested and see what happens tomorrow.
As far as binaries are concerned, I have already use the mount.cifs binary that is on the website...I also believe that you want me to download the kernel cifs binary that is on the website (is that right??) .
Thanks everybody for your input......really appreciate it.
Adnan
I have tried pass , not password (sorry again)
Moreover, I will try what u guys have suggested and see what happens tomorrow.
As far as binaries are concerned, I have already use the mount.cifs binary that is on the website...I also believe that you want me to download the kernel cifs binary that is on the website (is that right??) .
Thanks everybody for your input......really appreciate it.
Adnan
yeap. and try using modules also (not compiled in the kernel)
hey
followed you here from the other thread
Are we sure that Windows 2K / XP / 2K3 file servers inherently support CIFS? (Glenn?)
followed you here from the other thread
Are we sure that Windows 2K / XP / 2K3 file servers inherently support CIFS? (Glenn?)
Yes Erik. At least according to all sources I can find:-). BTW, did you (Erik) replace the kernels CIFS with something newer?
-- Glenn
-- Glenn
Glenn > No. I configured the built in CIFS support in K2.6.4.
ASKER
I have recieved this reposnse from Steve French (the original coder of CIFS) and my response ....
Any thoughts??????
Message from Steve........
The symptoms shown by your debug log entries match what happens when the cifs client attempts to mount older servers to Windows9x or Windows NT (prior to Windows 2000). I have not seen this happen to Windows 2000 or later but perhaps the servers will do if the list of allowed workstations is restricted through the Windows admin gui since without the rfc1001 session initialize a clients can not identify their
(netbios) workstation name. Older operating systems (such as NT4)
which do not support "native TCP" require that the first frame sent be RFC1001 ("Netbios over TCP") netbios session initialize (before the SMB negotiate protocol request) which the cifs client does not sent. This would not be too hard to add though.
Other possible less likely reasons could be hitting a tcp session. Let me know if you are restricting the hostnames/workstation names that can connect to your system since the symptom indicates that the server is taking down the tcp session before responding even to the first SMB request.
My response.............
I am trying to Connect to Windows 2000 Server, XP (home and Professional) and Windows 2003 Server. I checked the Local Security Settings of Windows XP Professional and it is not restricting any Workstations. Also to verify that I connect to this XP machine. I mounted a share of this machine using
Smbclient (samba 3.0.2rc2) to my Linux machine (Mandrake 9.0). Moreover, I can also connect to this share using other Windows Machines from My Network Places.
I have also run,
Smbclient -L "192.168.1.35" -N and I see the list of machines that are registered to that Workgroup/Domain.
Hope this additional information helps.
Any thoughts??????
Message from Steve........
The symptoms shown by your debug log entries match what happens when the cifs client attempts to mount older servers to Windows9x or Windows NT (prior to Windows 2000). I have not seen this happen to Windows 2000 or later but perhaps the servers will do if the list of allowed workstations is restricted through the Windows admin gui since without the rfc1001 session initialize a clients can not identify their
(netbios) workstation name. Older operating systems (such as NT4)
which do not support "native TCP" require that the first frame sent be RFC1001 ("Netbios over TCP") netbios session initialize (before the SMB negotiate protocol request) which the cifs client does not sent. This would not be too hard to add though.
Other possible less likely reasons could be hitting a tcp session. Let me know if you are restricting the hostnames/workstation names that can connect to your system since the symptom indicates that the server is taking down the tcp session before responding even to the first SMB request.
My response.............
I am trying to Connect to Windows 2000 Server, XP (home and Professional) and Windows 2003 Server. I checked the Local Security Settings of Windows XP Professional and it is not restricting any Workstations. Also to verify that I connect to this XP machine. I mounted a share of this machine using
Smbclient (samba 3.0.2rc2) to my Linux machine (Mandrake 9.0). Moreover, I can also connect to this share using other Windows Machines from My Network Places.
I have also run,
Smbclient -L "192.168.1.35" -N and I see the list of machines that are registered to that Workgroup/Domain.
Hope this additional information helps.
Well.... Steve should really know what he's talking about... We might theorize, but he's the Master in this case:-).
You should try to work closely with him on this, since it sounds more like a genuine bug than anything else. If we're lucky, the "first probable cause" might explain Eriks problem too. Polite but tenacious would be my strategy:-).
-- Glenn
You should try to work closely with him on this, since it sounds more like a genuine bug than anything else. If we're lucky, the "first probable cause" might explain Eriks problem too. Polite but tenacious would be my strategy:-).
-- Glenn
You could also try posting to
linux-cifs-client at lists.samba.org
i received no response though.
What's Steve's email address?
linux-cifs-client at lists.samba.org
i received no response though.
What's Steve's email address?
ASKER
To answer both the questions above....
Yes, I am trying to work with Steve on this. Hopefully, he will respond with my answers.
Next, I have also posted to linux-cifs-client at list.samba.org, but with not much response like other user-groups :(
Steve email address is: smfltc@us.ibm.com
Thanks
Adnan
Yes, I am trying to work with Steve on this. Hopefully, he will respond with my answers.
Next, I have also posted to linux-cifs-client at list.samba.org, but with not much response like other user-groups :(
Steve email address is: smfltc@us.ibm.com
Thanks
Adnan
Just one stupid thought :
mount -t cifs //192.168.1.35/e\$ /mnt -o user=administrator,passwor
(note the escaping of the "$" for the shell).
Would this make any difference ?
mount -t cifs //192.168.1.35/e\$ /mnt -o user=administrator,passwor
(note the escaping of the "$" for the shell).
Would this make any difference ?
ASKER
good point Alf666 :=)
This was just an example, I have tried other shares too....like
mount -t cifs "//192.168.1.35/downloads"
I have been able to mount all those shares using smbclient as well as smbmount.
Strange....isn't it?
My network is based on Netbios over TCP/IP and I thought CIFS may pose a problem with it. But Steve (original coder) of CIFS has confirmed that it should run over Netbios over TCP/IP.
Thanks for your help
This was just an example, I have tried other shares too....like
mount -t cifs "//192.168.1.35/downloads"
I have been able to mount all those shares using smbclient as well as smbmount.
Strange....isn't it?
My network is based on Netbios over TCP/IP and I thought CIFS may pose a problem with it. But Steve (original coder) of CIFS has confirmed that it should run over Netbios over TCP/IP.
Thanks for your help
Have you tried sniffing (using tethereal or tcpdump) ?
Escaping the $ isn't necessary... The shell will not see it as a variable indicator when followed by ws, so will just leave it be.
I think we can be fairly certain we're looking at a bug... Sniffing it out is a good idea though;-).
-- Glenn
I think we can be fairly certain we're looking at a bug... Sniffing it out is a good idea though;-).
-- Glenn
Hey guys
I've done some research and found that the SMB/CIFS implementation in OS-X seems to have a similar authentication issue. May want to raise this to Steve to aid his troubleshooting.
I've done some research and found that the SMB/CIFS implementation in OS-X seems to have a similar authentication issue. May want to raise this to Steve to aid his troubleshooting.
ASKER
I have posted additional comment at listserv.
The other problem is that I get an kernel OOPS when I try to mount the windows share. But as soon as I increase the debugging by
echo 1 > /proc/fs/cifs/cifsFYI
OOPS goes away and I get the message that I posted above.
Steve's comment:
There is a known oops (at least in the older 2.6 code, this was fixed prior to 1.04 version of the cifs vfs) in which the kernel can oops after the server abnormally terminates the tcp session (e.g. the above log entry reminded me of that case)
..........................
I am not getting any response from Steve yet on this issue other than the one I posted before. Now I am thinking maybe Mandrake or Redhat has some special modification to Steve's original code. I am gonna do a google on it.
Moreover, I also ran with Mandrake 10.0 (kernel 2.6.3), I still am unable to mount but the error number was different. I do not have that machine now, but next week I will install a test Mandrake 10.0 and post the error message there and here.
Whew............
Thanks all
Adnan
The other problem is that I get an kernel OOPS when I try to mount the windows share. But as soon as I increase the debugging by
echo 1 > /proc/fs/cifs/cifsFYI
OOPS goes away and I get the message that I posted above.
Steve's comment:
There is a known oops (at least in the older 2.6 code, this was fixed prior to 1.04 version of the cifs vfs) in which the kernel can oops after the server abnormally terminates the tcp session (e.g. the above log entry reminded me of that case)
..........................
I am not getting any response from Steve yet on this issue other than the one I posted before. Now I am thinking maybe Mandrake or Redhat has some special modification to Steve's original code. I am gonna do a google on it.
Moreover, I also ran with Mandrake 10.0 (kernel 2.6.3), I still am unable to mount but the error number was different. I do not have that machine now, but next week I will install a test Mandrake 10.0 and post the error message there and here.
Whew............
Thanks all
Adnan
While you're at it guys... have you tried the "smoking fresh" patches? (Not the ones bundled with the kernel)
-- Glenn
-- Glenn
v 1.0.2b for K2.4.X is August '03
v 0.94 is the one that is built in to the 2.6 kernel. This is the most current 2.6 compatible CIFS VFS release
v 0.94 is the one that is built in to the 2.6 kernel. This is the most current 2.6 compatible CIFS VFS release
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
It's "hidden" in the "sidebar" about 2.6 kernels. Perhaps not the best webpage design one's seen:-).
As usual... Good Luck (to both of you)!
-- Glenn
As usual... Good Luck (to both of you)!
-- Glenn
ASKER
Thanks...will try that and let you guys know of the results
Hard to tell if anything really helped, true. Only Adnan (and Erik) can tell for sure.
-- Glenn
-- Glenn
ASKER
Sorry, I never responded earlier to this thread until I recieved a message this morning. Due to this problem not being resolved, we changed our OS from Linux to FreeBSD and my original issue regarding 2GB was solved.
I do not know if Erik got his issue resolved. Whomover, puts the correct information regarding the fix to the solution I will accept that as an answer and give him/her the points.
Thanks
I do not know if Erik got his issue resolved. Whomover, puts the correct information regarding the fix to the solution I will accept that as an answer and give him/her the points.
Thanks
SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
-- Glenn (who mostly post to say "Listneing":-)