okay@TAFS.MITRE.ORG ("Okay, S J") (08/02/90)
In <9008011858.AA27274@humu.nosc.mil> kent@humu.nosc.mil (Kent K. Kuriyama) writes > >I have heard that the NFS implementation from HP/Apollo does not >support PC-NFS. If this is true has anyone done the port of >PC-NFSD? > Is this true??---If so, can anyone recommend an NFS Product that will work between PCs and Apollos?. We were planning to use PC-NFS to allow us to share volumes and printers between some our PCs and 4500s. This could very well change now... Why can't HP just break down and play with others!?!?!? ------------- Stephen Okay OKAY@TAFS.MITRE.ORG Technical Aide, The MITRE Corporation Shortly after the hold-up, Picasso attempts a police-sketch of the suspects. Disclaimer:I get *MYSELF* in enough trouble with my opinions, Why inflict them on MITRE?
rees@pisa.ifs.umich.edu (Jim Rees) (08/03/90)
In article <9008021356.AA05757@mwunix.mitre.org>, okay@TAFS.MITRE.ORG ("Okay, S J") writes: In <9008011858.AA27274@humu.nosc.mil> kent@humu.nosc.mil (Kent K. Kuriyama) writes >I have heard that the NFS implementation from HP/Apollo does not >support PC-NFS. If this is true has anyone done the port of >PC-NFSD? ... Why can't HP just break down and play with others!?!?!? They did. Apollo has NFS (although I've heard the server side doesn't work too well). Does PC-NFS use a different protocol from regular NFS? I thought all NFS implementations were supposed to play together? Why don't the PC-NFS folks just break down and play with the others?
krowitz@RICHTER.MIT.EDU (David Krowitz) (08/03/90)
PC-NFS requires an additional server, PCNFSD, in addition to the regular NFS servers. I believe it is needed to do user authentication since the PC's don't really have a Unix-style UID/GID database. (I set up a PC-NFS daemon once, about 4 years ago, to allow a PC access to an Alliant FX/1; but I don't remember the details). The PC-NFS daemon sources are readily available. I believe the ones I was using were shipped with the PC's ethernet board and NFS installation media. The problem is that since Apollo's NFS implementation is not a port of Sun's NFS code (which is what almost everyone else's is), the remote procedure call (RPC) library is not accessable to user programs on Apollo workstations. You can compile the code, but you can not link it to the necessary libraries. PC-NFSD is a small program, the version I had was maybe 40-60Kb of source code. Apollo could have easily ported the code provided with the PC-NFS package or have reimplemented the function with their own NFS implementation. Instead, they have chosen to ship an incomplete NFS implementations (no RPC library, no NFSD, no ability to execute or compile objects on remote file systems (fixed on the 3rd or 4th release), etc.); and when the end-user's ask for the functions to the implemented, the standard reply has been "Oh, you don't want to use that, you want to use NCS". NCS is a fine product ... if you are a software systems engineering building a new package ... but telling end-users that what they want to do is to completely rebuild both the client and server sides of a distributed file system on their own is just plain stupid. Every vendor I have had the occasion to check either has PC-NFSD included in their NFS release, or has the ability for the user to compile and install the code theirselves (ie. the RPC calls are available in the system libraries if NFS is installed) except Apollo. -- David Krowitz krowitz@richter.mit.edu (18.83.0.109) krowitz%richter.mit.edu@eddie.mit.edu krowitz%richter.mit.edu@mitvma.bitnet (in order of decreasing preference)
davidb@brac.inmos.co.uk (David Boreham) (08/07/90)
In article <1990Aug2.230727.17560@terminator.cc.umich.edu> rees@citi.umich.edu (Jim Rees) writes: >In article <9008021356.AA05757@mwunix.mitre.org>, okay@TAFS.MITRE.ORG >("Okay, S J") writes: > In <9008011858.AA27274@humu.nosc.mil> kent@humu.nosc.mil (Kent K. >Kuriyama) writes > >I have heard that the NFS implementation from HP/Apollo does not > >support PC-NFS. If this is true has anyone done the port of > >PC-NFSD? > > ... > > Why can't HP just break down and play with others!?!?!? > >They did. Apollo has NFS (although I've heard the server side doesn't work >too well). Does PC-NFS use a different protocol from regular NFS? I >thought all NFS implementations were supposed to play together? Why don't >the PC-NFS folks just break down and play with the others? OK, to try and clear this one up: NFS and PCNFS are not the same thing. To be more specific, there is a PRODUCT, sold by SUN, called PC-NFS. This gives you the ability to do remote file access on a UNIX system from a PC as well as telnet, ftp and so on. In order to function, this product needs to perform USER AUTHENTICATION. It does this by communicating (using RPC) with a server running on a UNIX box, called ``pcnsfd'' (or rpc.pcnfsd). As well as doing the authentication, the pcnfsd server provides remote printing services. Apollo support NFS, and therby support all the functions provided by PC-NFS, EXCEPT user authentication. This is not usually a problem since most sites would have at least one non-apollo machine running yellow pages and the pcnfsd anyway. We run our network in this way and provided you keep the password databases reasonably unified, there are no problems. The only time you need pcnfsd running on an Apollo is where you either have no SUNs or other BSD-type boxes on the network or where you keep separate UID domains (not a very good idea). So, there are two things---PC-NFS which is a bunch of disks and a manual, and pcnfsd which is a daemon running on a host and provides authentication and printing services to users of PC-NFS. David Boreham, INMOS Limited | mail(uk): davidb@inmos.co.uk or ukc!inmos!davidb Bristol, England | (us): uunet!inmos.com!davidb +44 454 616616 ex 547 | Internet: davidb@inmos.com
bep@quintro.uucp (Bryan Province) (08/07/90)
In article <9008021356.AA05757@mwunix.mitre.org> okay@TAFS.MITRE.ORG ("Okay, S J") writes: >In <9008011858.AA27274@humu.nosc.mil> kent@humu.nosc.mil (Kent K. Kuriyama) writes >> >>I have heard that the NFS implementation from HP/Apollo does not >>support PC-NFS. If this is true has anyone done the port of >>PC-NFSD? >> >Is this true??---If so, can anyone recommend an NFS Product that will work >between PCs and Apollos? > >------------- >Stephen Okay >OKAY@TAFS.MITRE.ORG Technical Aide, The MITRE Corporation We are using DPCI between our PCs and Apollos and it works OK except for some pty problems every once in a while. Please no flames about me using an APOLLO product. -- --=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-- Bryan Province Glenayre Corp. quintro!bep@lll-winken.llnl.gov Quincy, IL tiamat!quintro!bep@uunet "Surf Kansas, There's no place like home, Dude."
rmf@media.uucp (Roger Fujii) (08/07/90)
rees@pisa.ifs.umich.edu (Jim Rees) writes: > Why can't HP just break down and play with others!?!?!? >They did. Apollo has NFS (although I've heard the server side doesn't work >too well). Does PC-NFS use a different protocol from regular NFS? I 1) Ha. Apollo's NFS has caused me *no* end of headaches. Saying that it 'doesn't work too well' is putting it kindly (BTW, the client side bits the big one too.....). 2) No. PC-NFS handles the pc-oid services printing and authentication. >thought all NFS implementations were supposed to play together? Why don't They usually do, unless you have apollo's version.... >the PC-NFS folks just break down and play with the others? Why do people make criticisms on topics they no *nothing* about? I suppose that one can take the rpc library in comp.sources.unix and link in the pcnfsd to it (I would try it here, but I don't have a pcnfs client (I do have the pcnfsd source though). I know that SUN's RPC code *does* compile on the apollos will very little modifications (some include file editing, if I remember correctly), and it actually *WORKS* (but I remember that the compiled portmapper didn't work quite right with the apollos (naturally)). -- Roger Fujii - Media Cybernetics Phone: (301)495-3305 Internet: rmf%media@uunet.uu.net UUCP: {uunet,hqda-ai}!media!rmf