pjh@mccc.UUCP (Pete Holsberg) (07/23/88)
I have Starlan and RFS running between a pair of 3B2/400s which are
running SysV R3.1.2. Machine mc3 is advertising resource mc3-root to
the primary (mccc). That resource is on mccc's automatic mount list,
yet mccc says that there are no resources available for mounting. ???
What's happening? How do I get mccc to mount mc3's resource? Thanks.
(I've been using the sysadm menus to do all this. Is there a bug
somewhere?) I should point out that at one point mc3-root was indeed
mounted on mccc's /mnt, but since rebooting mccc, I can't get mccc to
mount mc3-root.
Pete Holsberg UUCP: {rutgers!}princeton!mccc!pjh
Technology Division CompuServe: 70240,334
Mercer College GEnie: PJHOLSBERG
Trenton, NJ 08690 Voice: 1-609-586-4800les@chinet.chi.il.us (Leslie Mikesell) (07/25/88)
In article <734@mccc.UUCP> pjh@mccc.UUCP (Pete Holsberg) writes: > >I have Starlan and RFS running between a pair of 3B2/400s which are >running SysV R3.1.2. Machine mc3 is advertising resource mc3-root to >the primary (mccc). That resource is on mccc's automatic mount list, >yet mccc says that there are no resources available for mounting. ??? > >What's happening? How do I get mccc to mount mc3's resource? Thanks. >(I've been using the sysadm menus to do all this. Is there a bug >somewhere?) I should point out that at one point mc3-root was indeed >mounted on mccc's /mnt, but since rebooting mccc, I can't get mccc to >mount mc3-root. When you reboot the primary name-server, you lose all the resource names although any current mounts would continue to work. The solution is to designate mc3 as a secondary name server or always re-boot it after re-booting mccc. Or, you can fix things manually: If you go to mc3 and "unadv resource_name" followed by "adv resource_name directory", it should clear things up, and "nsquery" will show the resource. At that point you can go to mccc and mount it with the command "mount -d resource_name directory". There may be an automatic recovery process attempting to do the mount, though. If so you should either wait till it succeeds or kill it after mounting manually. If you have designated a secondary name server, note that it takes over automatically if the primary goes down, but you must manually switch back before anything new can be advertised (using "rfadmin -p" at the secondary). Also note that no process can be using the mount point (or below) as its current directory when the mount is done (but this would give a different error message). Les Mikesell
rwhite@nusdhub.UUCP (Robert C. White Jr.) (07/28/88)
in article <734@mccc.UUCP>, pjh@mccc.UUCP (Pete Holsberg) says: > I have Starlan and RFS running between a pair of 3B2/400s which are > running SysV R3.1.2. Machine mc3 is advertising resource mc3-root to > the primary (mccc). That resource is on mccc's automatic mount list, > yet mccc says that there are no resources available for mounting. ??? Several Questions: (1) Are both mc3 and mccc in teh same domain? (2) If 1 above is false, have you used the DOMAIN.resource naming convention? (3) If 1 above is true, are you shoure you have one and only one of them defined as the primary name server, and the other has had "rfadmin -q <primary name server name>" command run on it at least once. (4) What does nsquery have to say. (5) When you created mc3-root did you restrict access to a spesific client list? (6) If 5 above is true, did you spell it right? Obvious checks: (7) Are RFS and STARLAN both set to run automatically on system startup? (8) Do other STARLAN Functions work? (9) Is your domain name unique to your network? The biggest and easiest thing to get wrong with RFS is defining two or more machines as the primary name server for the same domain name. If you do this, and circumstances permit it to pass unnoticed, you will effectively have two isolated domains with the same name. Each machine on the network will only be able to access the other machines which beleive in the same primary. If an out-of-domain request comes through for that domain name, all hell will break loose. If you want each machine to serve as its own domain, then all your mount and inquiry requests will have to start with the domain name of the machine as well as the resource name (NUION.pubdev), and the domain names have to be unique. If you have accidently dual-domained your systems then do the following: (1) log into the machine which you want to be the real primary and use "sysadm rfsmgmt" to add the other machine name to that primary and give it a password. you may want to list the second most important machine as a secondary if you have more than two machines in the domain. (2) log into the other machine and use "sysadm rfsmgmt" to re-setup the local machine and make it a "non-primary" durring the setup. When the dependant machine is started it will atuomatically receive the details of it's role from the primary. Once the primary is configured correctly, issuing "rfadmin -q primary-name" on any of the dependant machines will usually correct any problems on that machine (except a bad password). Rob.