On Fri, 2007-03-09 at 20:19 -0600, Mikkel L. Ellertson wrote: > Craig White wrote: > > On Fri, 2007-03-09 at 18:15 -0600, Mikkel L. Ellertson wrote: > >> Cifs supports UNIX extensions. They both support mixed case. It > >> depends on the options you pick. > > ---- > > sure - are you recommending it? Do you do user maps? Are you suggesting > > that cifs mounts are preferable mounts for Linux<=>Linux/Posix<=>Posix > > use? > I would not recommend any network file system without knowing the > network it is going to be used on. There are times where cifs mounts > are going to work out better then NFS mounts. This is especially > true when most clients are Windows machines. If you only have a few > Linux clients, then it makes sense for them to use the same method > as the Windows machines. ---- The only scenario that would make sense would be one where Windows and Linux users might use the same file at the same time at which point, the only way to know that say a Windows users had a write lock on the file would be using it as a cifs mount. I cannot envision a scenario where a cifs mount would work out better than an NFS mount for a Linux client but you are a bright guy and could possibly enlighten me. More to the point, the effort that is necessary to map the uid from Linux to Windows is beyond any simple implementation and if you are going to that level of configuration, it's hardly a stretch to do the native mounts anyway. ---- > > Yeah and then the name is mangled for Windows users since they cannot > > comprehend any difference between 'my file.txt' and 'My File.txt' > > > > So you've relocated the problem from Linux to Windows. > > > You actually have Windows users that type in file names when they > want to open a file? ---- I do run a fairly extensive set of scripts including logon scripts for each user, etc. but no, users mostly won't but the point that I was making was that one the above scenario, one or both of the files would have mangled names to satisfy the limitations of Windows. Craig
