hello,

we have a 2node sles10 sp3 oes2 cluster with cifs enabled.

Unfortunately, we only have problems with cifs. Nodes hang up, resources go comatose... .

The documentation on cifs is so bad, it's unbelievable. U have to collect your information from everywhere.
...
If we restart cifs with imanager or rcnovell-cifs, this is the exact output on /var/log/messages:

Code:
adminus daemon: CIFS: MNSS_CIFSSetServerConfiguration
CIFS[30499]: WARNING: ENTRY: CIFS server is stopped due to signal 15
CIFS[30499]: WARNING: ENTRY:   CIFS server is shut down 
adminus daemon: The cifsStatus is 0 
adminus daemon: CIFS: MNSS_CIFSGetServerConfiguration 
adminus daemon: CIFS: MNSS_CIFSSetServerConfiguration
checkproc: checkproc: cannot stat /usr/sbin/smbd: No such file or directory
adminus daemon: The cifsStatus is 0 
CIFS[2429]: WARNING: ENTRY:  Interface lo is skipped from populating for name resolution
CIFS[2429]: WARNING: ENTRY:  Interface lo is skipped from populating for name resolution
CIFS[2429]: WARNING: ENTRY:  Interface lo is skipped from populating for name resolution
CIFS[2429]: WARNING: RPC: Cross Protocol File Locking is Enabled.
CIFS[2429]: WARNING: ENTRY: Auditing interface not initialized.
CIFS[2429]: WARNING: ENTRY: cachevalue=2048 dir cache size=2048
CIFS[2429]: WARNING: ENTRY: cachevalue=20000 file cache size=20000
CIFS[2429]: WARNING: ENTRY: cachevalue=50000 subdir cache size=50000
CIFS[2429]: ERROR: ENTRY: CIFSNDSReadFromNDS: Error adding a new share (CIFSNDSPutSharePointInfo): 22101
CIFS[2429]: WARNING: RPC: requestNumber: fadebad1 events: 1ff domainpath /tmp/.ncp2cifs
CIFS[2429]: WARNING: AUTH: thread_processContextsList exiting !!!!
adminus daemon: CIFS: MNSS_CIFSGetServerConfiguration 
adminus daemon: Data Successfully Written 
adminus daemon: Data is successfully passed with the pipe 5 
adminus daemon: NSS:CIFS:RecvDataFrom_Server
CIFS[2429]: ERROR: ENTRY: CIFSNDSGetCIFSServerInfo: isVirtualServer = FALSE, uServerVersion = %U, aServerVersion = e, expected = %U  
adminus daemon:  Command completed successfully
adminus daemon: The function is Fid_CIFS_GetServerConfiguration
adminus daemon: CIFS: AddServerInfo
There is a problem with the 22101 code. In the Cifs admin documentation this problem is mentioned and it says that cifs has to be restarted every time edir is restarted. This doesn't help, since we have restartet it many times after edir runs.

Secondly, some people put a sleep 60 between the mounting of the volume and the cifs-command in the load-script. Also didn't help.

Also there is this "ERROR: ENTRY: CIFSNDSGetCIFSServerInfo" error which doesn't tell me anything.

Can anyone please help on this.
Hugo