If I understand correctly, an NSS volume is automatically presented to cifs (as listed by "novcifs -sl") when mounted.

I need some clustered volumes not to be visible at all as a cifs share (by accessing the hostname/IP only (from a Mac) of the virtual server of the volume / the physical node / an other virtual server published as a cifs share on the same node).
In order to hide a cifs share, is it adequate to use "novcifs -srn" in the cluster volume load script ?
Remark 1) : "novcifs -srn" is it dedicated to local volumes only !?
Remark 2) : "novcifs -srn" is added at the end of the load script and a sleep command (loop in a shell script) is required if the volume size is not very small (otherwise the "novcifs -srn" is run before the volume is presented to cifs (not yet displayed by "novcifs -sl")).

I need some clustered volumes to be published with a different cifs share name.
In order to rename a cifs share, is it adequate to use "novcifs -sap -n" and "novcifs -srn" in the cluster volume load script ?
Above remarks 1) and 2) also apply.


The sleep command (loop in a shell script) mentioned above was previously required during the mixed cluster (Netware and Linux nodes), to prevent the ""novcifs --add" to be run before the volume is presented to cifs, apparently the load scripts commands where (are ?) not fully executed sequentially !?

Yet, my clustered volumes work great in production ("cluster commit" executed), but I still have a small doubt about this way to be correct and compatible in the future.

Any advice is welcome.

Servers releases : SLES10SP3 and OES2SP3 fully patched.

Best regards.