From: Ferran Eloi Newsgroups: comp.parallel.mpi Subject: Some problems with permissions in other machines. Date: Mon, 09 Nov 1998 12:23:53 +0100 Organization: Universitat de Lleida (udl.es) Message-Id: <3646D0C8.B68B1994@alumnes.eup.udl.es> References: <364149D3.D67B19CB@scl.ameslab.gov> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit > I have some problems with parallel execution of mpich programs, for > example, First I'm trying to execute the hello_world program in 1 computer and I have not any problem, then I've tried to execute in 2 machines, but the execution gives me one error, exactly a permission denied; my machinefile I think is correct typed,. I'm root on two machines and I've copied the executable to the root directories of 2 machines. The programs runs in two machines , but whe I'm trying to execute in parallel then the error happen. ( the two machines are jupiter-0190 (the local machine) and jupiter-0196 (the remote machine) ). My command line is: [root@jupiter-0190 /root]# mpirun -echo -machinefile /usr/local/mpich/util/machines/machines.LINUX -np 2 hola And the messages displayed are: ++ echo default_arch = LINUX default_arch = LINUX ++ echo default_device = ch_p4 default_device = ch_p4 ++ echo machine = ch_p4 machine = ch_p4 ++ [ 1 -le 5 ] ++ arg=-machinefile ++ shift ++ machineFile=/usr/local/mpich/util/machines/machines.LINUX ++ eval machineFile0=/usr/local/mpich/util/machines/machines.LINUX +++ machineFile0=/usr/local/mpich/util/machines/machines.LINUX ++ shift ++ [ 1 -le 3 ] ++ arg=-np ++ shift ++ np=2 +++ echo 2 +++ sed s/^[0-9]*$// ++ nodigits= ++ [ != -o 2 -lt 1 ] ++ eval np0=2 +++ np0=2 ++ shift ++ [ 1 -le 1 ] ++ arg=hola ++ shift +++ echo hola +++ sed s/%a//g ++ proginstance=hola ++ [ = -a = -a ! -x hola ] ++ [ = -a -x hola ] ++ progname=hola ++ [ 1 -le 0 ] ++ [ 1 -le 0 ] ++ [ ch_p4 = #DEFAULT_MACHINE# ] ++ [ LINUX = #DEFAULT_ARCH# ] ++ [ ch_p4 = #DEFAULT_DEVICE# ] ++ [ = -a hola = ] ++ larch= ++ [ -z ] ++ larch=LINUX ++ [ = dbx -a ( LINUX = alpha -o LINUX = IRIX -o LINUX = SGI -o LINUX = IRIX64 ) ] ++ [ -n sed -e s@/tmp_mnt/@/@g ] +++ pwd +++ sed -e s@/tmp_mnt/@/@g ++ PWDtest=/root ++ [ ! -d /root ] ++ [ -n /root ] +++ echo /root +++ sed -e s@/tmp_mnt/@/@g ++ PWDtest2=/root ++ /bin/rm -f /root/.mpirtmp1475 /root/.mpirtmp1475 +++ eval echo test > /root/.mpirtmp1475 ++ [ ! -s /root/.mpirtmp1475 ] ++ PWD=/root ++ /bin/rm -f /root/.mpirtmp1475 /root/.mpirtmp1475 ++ [ -n /root ] ++ PWD_TRIAL=/root +++ echo hola +++ sed s/\/.*// ++ tail=hola ++ [ hola = ] ++ progname=/root/hola ++ [ = ] +++ /usr/local/mpich/bin/tarch ++ arch=LINUX ++ archlist=LINUX ++ [ ch_p4 = ] ++ [ ch_p4 = p4 -o ch_p4 = execer -o ch_p4 = ch_tcp -o ch_p4 = sgi_mp -o ch_p4 = ch_p4 -o ch_p4 = ch_p4-2 -o ch_p4 = globus -o ch_p4 = globus ] ++ [ = ] ++ MPI_HOST= ++ [ LINUX = ipsc860 ] +++ hostname ++ MPI_HOST=jupiter-0190 ++ [ jupiter-0190 = ] ++ [ /root != ] +++ pwd +++ sed -e s%/tmp_mnt/%/%g ++ PWD_TRIAL=/root ++ [ ! -d /root ] ++ [ 0 = 1 ] + argsset=1 + mpirun_version= + mpirun_version=/usr/local/mpich/lib/LINUX/ch_p4/mpirun.ch_p4 + exitstat=1 + [ -n /usr/local/mpich/lib/LINUX/ch_p4/mpirun.ch_p4 ] + [ -x /usr/local/mpich/lib/LINUX/ch_p4/mpirun.ch_p4 ] + . /usr/local/mpich/lib/LINUX/ch_p4/mpirun.ch_p4 ++ exitstatus=1 ++ [ -z /usr/local/mpich ] ++ [ /usr/local/mpich = #MPIR_HOME# ] ++ [ lib/LINUX/ch_p4 = #MPIRUN_BIN# ] ++ MPIRUN_HOME=/usr/local/mpich/lib/LINUX/ch_p4 ++ [ -z 1 ] ++ [ -n ] ++ [ -n ] ++ . /usr/local/mpich/lib/LINUX/ch_p4/mpirun.pg +++ [ 1 = ] +++ [ 0 = 0 ] +++ narch=1 +++ arch1=LINUX +++ archlist1=LINUX +++ archlocal=LINUX +++ np1=2 +++ [ 0 = 1 ] +++ procFound=1 +++ machinelist= +++ archuselist= +++ nprocuselist= +++ curarch=1 +++ nolocalsave=0 +++ archlocal=LINUX +++ [ 1 -le 1 ] +++ eval arch=$arch1 ++++ arch=LINUX +++ eval archlist=$archlist1 ++++ archlist=LINUX +++ [ -z LINUX ] +++ eval np=$np1 ++++ np=2 +++ [ -z 2 ] +++ eval mFile=$machineFile1 ++++ mFile= +++ [ -n -a -r ] +++ [ -z /usr/local/mpich/util/machines/machines.LINUX ] +++ [ -z /usr/local/mpich/util/machines/machines.LINUX -o ! -s /usr/local/mpich/util/machines/machines.LINUX -o ! -r /usr/local/mpich/util/machines/machines.LINUX ] +++ [ 0 = 0 -o 1 -gt 1 ] ++++ cat /usr/local/mpich/util/machines/machines.LINUX ++++ sed -e /^#/d -e s/#.*^//g ++++ head -2 ++++ tr \012 +++ machineavail=jupiter-0190 jupiter-0196 +++ KeepHost=0 +++ loopcnt=0 ++++ expr 2 - 1 +++ nleft=1 +++ [ 1 -lt 2 ] +++ nfound=0 +++ nprocmachine=1 ++++ expr jupiter-0190 : .*:\([0-9]*\) +++ ntest= +++ [ -n -a != 0 ] +++ [ 0 = 1 -o 0 = 1 -o jupiter-0190 != jupiter-0190 ] +++ [ 1 = 2 ] +++ [ jupiter-0190 = jupiter-0190 ] +++ KeepHost=1 +++ nprocmachine=1 ++++ expr jupiter-0196 : .*:\([0-9]*\) +++ ntest= +++ [ -n -a != 0 ] +++ [ 0 = 1 -o 1 = 1 -o jupiter-0196 != jupiter-0190 ] +++ [ 1 -gt 1 ] +++ [ -z ] +++ nprocmachine=1 +++ machinelist= jupiter-0196 +++ archuselist= LINUX +++ nprocuselist= 1 ++++ expr 1 + 1 +++ procFound=2 ++++ expr 0 + 1 +++ nfound=1 ++++ expr 1 - 1 +++ nleft=0 +++ [ 2 = 2 ] +++ break ++++ expr 0 + 1 +++ loopcnt=1 +++ [ 1 = 0 -a 1 -gt 1 ] +++ [ 2 -lt 2 ] ++++ expr 1 + 1 +++ curarch=2 +++ procFound=0 +++ nolocal=1 +++ machineFile= +++ [ 2 -le 1 ] +++ nolocal=0 +++ [ 1 != 1 ] +++ break ++ prognamemain=/root/hola ++ [ -z ] ++ p4pgfile=/root/PI1475 ++ echo Test ++ [ ! -s /root/PI1475 ] ++ /bin/rm -rf /root/PI1475 ++ procNum=1 +++ echo 1 +++ cut -d -f1 ++ nprocval=1 ++ [ -z 1 ] +++ expr 1 - 1 ++ nprocval=0 ++ archval=LINUX +++ echo /root/hola +++ sed s/%a/LINUX/g ++ proginstance=/root/hola ++ prognamemain=/root/hola ++ [ 0 = 1 ] ++ [ 0 = 0 ] ++ [ -z jupiter-0190 ] ++ nprocval=0 ++ echo jupiter-0190 0 /root/hola ++ cnt=1 ++ [ 1 = 2 ] +++ echo LINUX +++ cut -d -f1 ++ archval=LINUX +++ echo 1 +++ cut -d -f1 ++ nprocval=1 +++ expr 1 + 1 ++ cnt=2 +++ echo /root/hola +++ sed s/%a/LINUX/g ++ proginstance=/root/hola ++ [ 0 = 1 ] ++ echo jupiter-0196 1 /root/hola ++ [ 0 = 0 -a ! -r /root/PI1475 ] ++ [ 0 = 0 -a 0 = 1 ] ++ /bin/sync ++ [ = ] ++ p4workdir=/root ++ startpgm=/root/hola -p4pg /root/PI1475 -p4wd /root ++ [ != ] ++ MPIRUN_DEVICE=ch_p4 ++ export MPIRUN_DEVICE ++ [ 0 = 1 ] ++ doitall= ++ [ 0 = 1 ] ++ /root/hola -p4pg /root/PI1475 -p4wd /root Permission denied. (If I wait some time it cause me and error of Timeout) Anybody can help me ? Thank you. Ferran Eloi.