On Thursday, 07/27/2006 at 08:58 ZE3, Shimon Lebowitz <[EMAIL PROTECTED]> wrote: > My first posting in that old thread said OBEYFILE, and Alan > had some juicy comment at that idea. > He suggested NETSTAT OBEY, which I have not looked at yet, > but is apparently an easier way to deal with OBEYFILE. > > I mentioned that nowadays I use IFCONFIG <if> DOWN or UP. > These commands require that you have a R/W mdisk currently > accessed, but I don't remember running into TCPIP link problems. > (Or maybe that is just because VM:Secure gives TCPIP r/o rights > to my mdisk?) > > In any case, I believe there is an option to pass a password (<g>). > I am home... so I am not checking this right now. > > I definitely remember in the past tearing my hair out trying > to get TCPIP to do something, and having it come up against > this 'link to the mdisk' problem. What it really needs is a better > command interface, such as SENDFILE, where it can accept > and run the file (if it comes from an authorized user), > or perhaps just accepting SMSG commands. > > And what is so terrible about SFS? Who has an MDISK???
The TCPIP server must be authorized to LINK to the minidisk that holds the file specified on the OBEYFILE command. If you have an ESM, then you need to put a rule in. If you don't, then you need to specify the minidisk password (if it isn't ALL, which it shouldn't be). You can't use SFS because (a) the stack and SFS don't get along very well, and (b) the mechanisms underlying the OBEYFILE interface would have to significantly change. With the addition of NETSTAT OBEY, the need for OBEYFILE is considerably reduced. Someday Miguel and I would like get ifconfig changed to use NETSTAT OBEY, but it hasn't yet bubbled to the top of the priority list. Alan Altmark z/VM Development IBM Endicott
