Noel.Hardy@newcastle.ac.uk (09/10/90)
Back in May a few messages went around about problems with the Meiko svcs daemon hanging up for no apparent reason. I've recently acquired an M-10 unit (I haven't actually seen the thing since I'm running remotely) and am running CStools on a SparcStation using SUNOS 4.0.3c. I've now got the same problems! Programs which run okay most of the time will suddenly hang for no reason. Once I've breaked from the program, the svcs daemon has frozen and must be restarted before using mrun again. Since this problem has obviously been around for months, has anyone found or been given a solution? While I'm on the subject of Meiko, I've come to realise (since I can't find it mentioned the documention) that byte orders need reversing when sending data (using csn_rx and csn_tx) between a process on the host (Sun) and a process on a transputer (by the way, I'm using FORTRAN but I've tried C and it's the same). I get round it by swapping the bytes before sending. Has anyone found an easy way round this, or is there something I've missed in the documentation? Thanks Noel Hardy JANET: Noel.Hardy @ uk.ac.newcastle Dept. of Marine Technology University of Newcastle upon Tyne NE1 7RU Tel. 091 - 2226000 x6726
bruce@seismo.gps.caltech.edu (Bruce Worden) (09/12/90)
In article <emu-ul20.cg.1990.0910.115710.nvi3@uk.ac.ncl.mts> Noel.Hardy@newcastle.ac.uk writes: > [ svcsd hangs sometimes ] I have never had this problem on our unit, but if I did I would harass Meiko until they fixed it. I would recommend that you do so, too. > [ thinks the byte order is reversed from the Sun host ] Yes, the byte order is different. Meiko supports the xdr functions (as does Sun) which is the (supposedly) portable way of dealing with the issue. There is no easy way around it. -------------------------------------------------------------------------- C. Bruce Worden bruce@seismo.gps.caltech.edu 252-21 Seismological Laboratory, Caltech, Pasadena, CA 91125