Newsgroups: comp.parallel.pvm
From: alsenoy@sch2.uia.ac.be (Kris.VanAlsenoy)
Subject: can't get memory
Organization: U.I.A.
Date: Mon, 10 Apr 1995 08:45:47 GMT
Message-ID: <1995Apr10.084547.9687@reks.uia.ac.be>


Dear PVM'ers,

I encountered a problem and got the following set of error messages:
libpvm [t40036]: fr_new() can't get memory
libpvm [t40036]: mxfer() mxinput bad return on pvmd sock
libpvm [t40036]: pvm_recv(): Can't contact local daemon
libpvm [t40036]: pvm_bufinfo(): Bad parameter
libpvm [t40036]: pvm_upkdouble(): No current buffer
libpvm [t40036]: pvm_upkdouble(): No current buffer
libpvm [t40036]: fr_new() can't get memory
libpvm [t40036]: mxfer() mxinput bad return on pvmd sock
libpvm [t40036]: pvm_recv(): Can't contact local daemon
libpvm [t40036]: pvm_bufinfo(): Bad parameter
libpvm [t40036]: pvm_upkdouble(): No current buffer
libpvm [t40036]: pvm_upkdouble(): No current buffer
libpvm [t40036]: fr_new() can't get memory
bat[15]: 15594 Segmentation fault(coredump)

Analysing the core, I found that the problem was was caused by
segmentation violation in enc_foo_init at 0x10014998
which I traced back to one of my calls to pvmfmcast

All problems probably come from the first message 'can't get memory',
however there is plenty of paging-space.
Small jobs run without a problem, larger ones cause problems.

Any explanations/suggestions ?
may thanks in advance,
      kris


