Enhanced autofs and CIFS

[Ruan_3]

Hi all

I have installed Enhanced AutoFS on one of my servers running HP-UX 11.11. I configured the /etc/auto.direct.cifs map file as per the admin guide. The CIFS filesystem mounts correctly using autofs, but when the CIFS filesystem is unmounted after the inactivity period has expired, the filesystem is not unmounted correctly.

Here is the output which I got in the trace file /var/adm/automountd.log after switching on tracing for autofs:

Jun 25 14:02:43 (thread 21) UNMOUNT REQUEST:
Jun 25 14:02:43 (thread 21) dev=4300000f rdev=0 direct
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x40000003 from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x40000001 from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x40000008 from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x40000007 from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x40000009 from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x40000004 from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x4000000a from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x40000006 from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x40000005 from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x43000000 from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x43000001 from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x43000002 from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x43000003 from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x4300000e from mnttab file
Jun 25 14:02:43 (thread 21) get_devid(): Got devid 0x4300000f from mnttab file
Jun 25 14:02:43 (thread 21) unmount /FTP/CIFS OK
Jun 25 14:02:43 (thread 21) UNMOUNT REPLY : status=0

But the cifslist command still shows the filesystem as mounted:

=========================================================================
server dmc_plz_csfs01:
=========================================================================
Remote Username: CIFSGUEST Local Username: root

Share: \\DMC_PLZ_CSFS01\CIFS
rw /FTP/CIFS

The problem is that when the CIFS filesystem needs to be mounted the next time a user accesses the /FTP/CIFS mountpoint and does an ll the output is:

. not found

The entry in the cifs trace is as follows

Successful initial mount:


06-25 15:11:45.991 [2] ipcserverstub: connectServer(0,3): dmc_plz_csfs01 (null s
tring) (null string) -1 0
06-25 15:11:45.991 [0] Establishing connection to server dmc_plz_csfs01: connect
Timeout: 5000, requestTimeout: 60000
06-25 15:11:45.992 [2] ipcserverstub: connectUser(0,3): dmc_plz_csfs01 CIFSGUEST
0 0
06-25 15:11:45.993 [2] user: setting up session with encrypted password
06-25 15:11:45.994 [10] cifs trace: cifsSessionSetup()
06-25 15:11:45.994 [10] smb: startCommand: 4
06-25 15:11:46.009 [10] cifs trace: [4] rval = 0/0
06-25 15:11:46.009 [0] Server OS: Windows NT 4.0
06-25 15:11:46.010 [0] Server LM: NT LAN Manager 4.0
06-25 15:11:46.010 [2] userConnectProc: smbUsername=CIFSGUEST, unixUid=0, cstatu
s=2, server=dmc_plz_csfs01, smbUid=2049
06-25 15:11:46.011 [2] ipcserverstub: connectShare(0,3): dmc_plz_csfs01 \\dmc_pl
z_csfs01\CIFS (null string)
06-25 15:11:46.011 [2] ipcserverstub: mountShare(0,3): dmc_plz_csfs01 \\dmc_plz_
csfs01\CIFS /FTP/CIFS
06-25 15:11:46.019 [2] ipcserver: closing client connection

Failing second mount:

06-25 15:11:47.735 [2] ipcserver: accepted client connection
06-25 15:11:47.737 [2] ipcserverstub: connectServer(0,3): dmc_plz_csfs01 (null s
tring) (null string) -1 0
06-25 15:11:47.738 [0] Establishing connection to server dmc_plz_csfs01: connect
Timeout: 5000, requestTimeout: 60000
06-25 15:11:47.738 [2] ipcserverstub: connectUser(0,3): dmc_plz_csfs01 CIFSGUEST
0 0
06-25 15:11:47.739 [2] ipcserverstub: connectShare(0,3): dmc_plz_csfs01 \\dmc_pl
z_csfs01\CIFS (null string)
06-25 15:11:47.740 [2] ipcserverstub: mountShare(0,3): dmc_plz_csfs01 \\dmc_plz_
csfs01\CIFS /FTP/CIFS

Anybody know what the problem could be?

[Ruan_3]

Anyone?

[Todd Whitcher]

Hello Ruan,

A few questions I can think of?

Does the automount debug show anything at the failed mount time?

After the cifs filesystem is umounted do you see it listed in /etc/mnttab?

What is the output of the mount -v command when the cifs filesystem is umounted?

Can you mount the cifs filesystem outside of autofs?

What version of CIFS are you using?

What is your ONC Cumulative patch version?

I've not heard of this issue before with CIFS but the support for CIFS and autofs is fairly new.

Todd

[Ruan_3]

Hi Todd,

Thanks for your reply! The answers to your questions:

Does the automount debug show anything at the failed mount time?

automount trace:

Jun 29 14:32:36 (thread 1) MOUNT REQUEST : name=/FTP/CIFS map=/etc/auto.direct.cifs opts= path=/FTP/CIFS
Jun 29 14:32:36 (thread 1) PUSH /etc/auto.direct.cifs
Jun 29 14:32:36 (thread 1) POP /etc/auto.direct.cifs
Jun 29 14:32:36 (thread 1) mount: dmc_plz_csfs01:/CIFS /FTP/CIFS cifs username=CIFSGUEST,password=general1
Jun 29 14:32:36 (thread 1) fork_exec: /sbin/fs/cifs/mount -o username=CIFSGUEST,password=general1 -- dmc_plz_csfs01:/CIFS /FTP/CIFS
Jun 29 14:32:36 (thread 1) fork_exec: returns exit status 0
Jun 29 14:32:36 (thread 1) MOUNT REPLY : status=0

CIFS trace:

06-29 14:32:36.904 [2] ipcserver: accepted client connection
06-29 14:32:36.907 [2] ipcserverstub: connectServer(0,3): dmc_plz_csfs01 (null string) (null string) -1 0
06-29 14:32:36.907 [0] Establishing connection to server dmc_plz_csfs01: connectTimeout: 5000, requestTimeout: 60000
06-29 14:32:36.908 [2] ipcserverstub: connectUser(0,3): dmc_plz_csfs01 CIFSGUEST 0 0
06-29 14:32:36.909 [2] ipcserverstub: connectShare(0,3): dmc_plz_csfs01 \\dmc_plz_csfs01\CIFS (null string)
06-29 14:32:36.909 [2] ipcserverstub: mountShare(0,3): dmc_plz_csfs01 \\dmc_plz_csfs01\CIFS /FTP/CIFS
06-29 14:32:36.913 [2] ipcserver: closing client connection
06-29 14:32:56.500 [2] share \\DMC_PLZ_CSFS01\CIFS: total files=4, open files=0

After the cifs filesystem is umounted do you see it listed in /etc/mnttab?

No, only the entry for autofs
/etc/auto.direct.cifs /FTP/CIFS autofs ignore,direct,dev=4300003e 0 0 1088511733

What is the output of the mount -v command when the cifs filesystem is umounted?
/etc/auto.direct.cifs on /FTP/CIFS type autofs ignore,direct,dev=4300003e on Tue Jun 29 14:22:13 2004

Can you mount the cifs filesystem outside of autofs?

Yes, using the same command as listed in the autofs trace - /sbin/fs/cifs/mount -o username=CIFSGUEST,password=general1 -- dmc_plz_csfs01:/CIFS /FTP/CIFS

ONC patch version:
PHNE_29783 1.0 ONC/NFS General Release/Performance Patch

CIFS:
B8724AA A.01.09 CIFS Client

AutoFS:
ENHAUTO B.11.11.0402.1 Enhanced AutoFS, February 2004


Thanks

Ruan

[Todd Whitcher]

I dont see any known issues so far for this problem description.

The debug below is during the failed autofs mount and not the successful mounts?
automount trace:

Jun 29 14:32:36 (thread 1) MOUNT REQUEST : name=/FTP/CIFS map=/etc/auto.direct.cifs opts= path=/FTP/CIFS
Jun 29 14:32:36 (thread 1) PUSH /etc/auto.direct.cifs
Jun 29 14:32:36 (thread 1) POP /etc/auto.direct.cifs
Jun 29 14:32:36 (thread 1) mount: dmc_plz_csfs01:/CIFS /FTP/CIFS cifs username=CIFSGUEST,password=general1
Jun 29 14:32:36 (thread 1) fork_exec: /sbin/fs/cifs/mount -o username=CIFSGUEST,password=general1 -- dmc_plz_csfs01:/CIFS /FTP/CIFS
Jun 29 14:32:36 (thread 1) fork_exec: returns exit status 0
Jun 29 14:32:36 (thread 1) MOUNT REPLY : status=0

That status = 0 means success as far as autofs is concerned so its a bit puzzling.

Todd

[Ruan_3]

Hi Todd,

Yes, the log extract is from the mount that does not seem to be working.

Here are the entries from the CIFS trace for the mounts:

Successful:

06-29 14:52:15.765 [2] ipcserver: accepted client connection
06-29 14:52:15.769 [2] ipcserverstub: connectServer(0,3): dmc_plz_csfs01 (null string) (null string) -1 0
06-29 14:52:15.770 [0] Establishing connection to server dmc_plz_csfs01: connectTimeout: 5000, requestTimeout: 60000
06-29 14:52:15.773 [10] cifs trace: cifsNegotiate()
06-29 14:52:15.774 [10] smb: startCommand: 0
06-29 14:52:15.775 [10] cifs trace: authenticationLevel "ntlm" negotiated by server dmc_plz_csfs01
06-29 14:52:15.776 [10] cifs trace: [0] rval = 0/0
06-29 14:52:15.776 [2] Negotiate result: maxXmit=4356, serverGmtOffset=-120
06-29 14:52:15.776 [2] server: setting window size to 16
06-29 14:52:15.777 [2] ipcserverstub: connectUser(0,3): dmc_plz_csfs01 CIFSGUEST 0 0
06-29 14:52:15.778 [2] user: setting up session with encrypted password
06-29 14:52:15.779 [10] cifs trace: cifsSessionSetup()
06-29 14:52:15.779 [10] smb: startCommand: 1
06-29 14:52:15.814 [10] cifs trace: [1] rval = 0/0
06-29 14:52:15.814 [0] Server OS: Windows NT 4.0
06-29 14:52:15.814 [0] Server LM: NT LAN Manager 4.0
06-29 14:52:15.815 [2] server: serverClasses: trying match with OS="" LM=""
06-29 14:52:15.815 [2] server: serverClasses: trying match with OS="*[Ww]indows[nN][tT]*" LM="*"
06-29 14:52:15.815 [0] server: serverClasses: match found with OS="*[Ww]indows [nN][tT]*" LM="*"
06-29 14:52:15.816 [2] userConnectProc: smbUsername=CIFSGUEST, unixUid=0, cstatus=2, server=dmc_plz_csfs01, smbUid=4097
06-29 14:52:15.816 [2] ipcserverstub: connectShare(0,3): dmc_plz_csfs01 \\dmc_plz_csfs01\CIFS (null string)
06-29 14:52:15.817 [10] cifs trace: cifsTreeConnect()
06-29 14:52:15.817 [10] smb: startCommand: 2
06-29 14:52:15.819 [10] cifs trace: [2] rval = 0/0
06-29 14:52:15.822 [2] ipcserverstub: mountShare(0,3): dmc_plz_csfs01 \\dmc_plz_csfs01\CIFS /FTP/CIFS
06-29 14:52:15.822 [2] nfs: set RX buffer size: want=131072, default=65535
06-29 14:52:15.823 [0] nfs: set RX buffer to 131072 bytes
06-29 14:52:15.823 [2] nfs: set TX buffer size: want=131072, default=65535
06-29 14:52:15.823 [0] nfs: set TX buffer to 131072 bytes
06-29 14:52:15.832 [2] srvUserForUid: Uid is 0
06-29 14:52:15.834 [2] srvUserForUid: Uid is 0
06-29 14:52:18.609 [2] ipcserver: closing client connection
06-29 14:52:18.616 [2] srvUserForUid: Uid is 0

Unsuccessful:

06-29 15:00:04.619 [2] ipcserver: accepted client connection
06-29 15:00:04.622 [2] ipcserverstub: connectServer(0,3): dmc_plz_csfs01 (null string) (null string) -1 0
06-29 15:00:04.622 [0] Establishing connection to server dmc_plz_csfs01: connectTimeout: 5000, requestTimeout: 60000
06-29 15:00:04.623 [2] ipcserverstub: connectUser(0,3): dmc_plz_csfs01 CIFSGUEST 0 0
06-29 15:00:04.624 [2] ipcserverstub: connectShare(0,3): dmc_plz_csfs01 \\dmc_plz_csfs01\CIFS (null string)
06-29 15:00:04.624 [2] ipcserverstub: mountShare(0,3): dmc_plz_csfs01 \\dmc_plz_csfs01\CIFS /FTP/CIFS
06-29 15:00:04.628 [2] ipcserver: closing client connection

When I do a manual CIFS mount using /sbin/fs/cifs/mount -o username=CIFSGUEST,password=general1 -- dmc_plz_csfs01:/CIFS /FTP/CIFS after the autofs umount, I get the same result as the failing CIFS mount attempt. The same entry as above is displayed in the CIFS trace file.

Thanks

[Todd Whitcher]

I cant find a known issue, I'd suggest logging a support case for cifs support.

[Todd Whitcher]

I just checked the release notes for CIFS. It states cifs and autofs 2.3 is only supported with 11.i version 2 which is 11.23 not 11.11.

Are you running 11.11 or 11.23?

http://docs.hp.com/hpux/onlinedocs/B8724-90055/B8724-90055.html

[Todd Whitcher]

OK, so further news from the CIFS support team states that it is actually supported on 11.i version 1 ( 11.11 ) but the release notes have not been updated yet. So this should all work for you. You may want to update to the latest Autofs patch along with the dependencies since it appears the initial release of Enhanced Autofs did not provide CIFS support at 11.11.

PHNE_29883 s700_800 11.11 ONC/NFS General Release/Performance Patch

Sorry for the confusion, I was just going by the release notes.

Todd

[Ruan_3]

Hi Todd,

I installed the patch, but the symptoms are still the same. I have logged a call with the HP South Africa call centre, but the support engineer assigned to the case was not able to help me and told me to log a message on ITRC, which is what I did in the first place!

Thanks

[melvyn burnard]

I would suggest this is not acceptable, I suggest you contact the manager of that office and request this be re-opened and handled in the correct way.

[Todd Whitcher]

Ruan,

That is absolutely unacceptable, I would suggest you log another case or re-open your previous case and ask them to elavate to another resource, if you get ANY resistance ask for a manager. Your configuration is supported by HP. The problem you are having sounds like it will need to get a little deeper debug from the code which will involve the support engineer elavating the case to his/her next level of support. If you have any issues with the support they provide please feel free to email me with the case id and I will follow up. If you would like, email me w/ your previous case id and I will follow up with that as well.


Todd
todd.whitcher@hp.com