Condor 2008 Fall

From CDOT Wiki
Jump to: navigation, search

Oct 28, 2008:

  • Downloaded condor 7.05 source tarball onto Ireland system (running rawhide)
  • Unpacked tarball, changed to src directory and ran ./build_init (seemed to run ok)
  • Tried to run configure but got error (something dealing with build_external)
    • to fix this I ran ./configure --disable-glibc-version-check but this produced another error - I needed to install csh
  • later in the evening Chris installed Fedora 9 so I'll have to start over.

Oct 29, 2008:

  • Now that Ireland is running Fedora 9, I did yum install condor - worked ok
    • ran condor_master
    • ran condor_status
    • ran condor_off
    • ran condor_off -master
  • Also, on China (running Fedora 8) I downloaded condor 7.05 source tarball.
    • installed csh
    • ran ./build_init
    • ran ./configure --disable-glibc-version-check
    • ran make - got error "asm/page.h no such file

Oct 30, 2008:

  • In the Ireland system, Edit and compile a simple "Hello World" c program
  • Edit a submit file named hello.sub, submit the file to condor (condor_submit hello.sub)
  • Run condor_q to check on the status of my job

Nov 4, 2008:**

  • Check the submit job on Ireland machine, it hasn't worked.
  • On Korea machine(Fedora 9), run yum install condor -- Can't install, remind me lack of gsoap
    • run install gsoap -- Can't install
    • manual download gsoap-2.7.9-0.1.I.i386.rpm, then run rpm -i gsoap-2.7.9-0.1.I.i386.rpm -- OK
    • yum install condor -- OK
  • On China machine(Fedora 8), reconfigure using the command "./configure --disable-glibc-version-check --disable-glibc-full-port", then run make it's ok.
  • On China machine, create a condor1 virtual machine for Fedora 9, and install condor on virtual machine
    • Solve network problem:
      • 1.Stop firewall
      • 2.Enable ipforward: echo 1 > /proc/sys/net/ipv4/ip_forward
      • 3.run "iptables -t nat -A POSTROUTING -o eth0 -j MASQUERADE"
    • scp 192.168.122.1:DOWNLOAD/*.gz .
    • run yum install csh
    • untar the condor source file
    • ran ./build_init
    • ran ./configure --disable-glibc-version-check ---- ERROR: build_krb5-1.4.3 ERROR: unable to build krb5-1.4.3

Nov 7, 2008:

Nov 10, 2008:

  • Download condor-7.0.5_1.fc9.src.rpm package from http://rpm.pbone.net on my laptop
  • Install Condor
  • Configure condor
    • ran condor_configure --install=/usr/etc --type=execute,submit,manager --local-dir=/home/condor
    • Error

Nov 11, 2008:

  • Install Fedora 9 32-bit on virtual machine on China and Ireland machine respectively.

Nov 12, 2008:

  • On Ireland system
    • yum remove condor
    • Download condor-7.0.5_1.fc9.src.rpm package from http://rpm.pbone.net
    • $ rpmdev-setuptree
    • $ cp condor-7.0.5_1.fc9.src.rpm ~/rpmbuild/SOURCES/
    • $ cd ~/rpmbuild/SOURCES/
    • $ rpmbuild --rebuild condor-7.0.5-1.fc9.src.rpm
      • error: Failed build dependencies:
      • pcre-devel is needed by condor-7.0.5-1.fc9.x86_64
      • postgresql-devel is needed by condor-7.0.5-1.fc9.x86_64
      • openssl-devel is needed by condor-7.0.5-1.fc9.x86_64
      • gsoap-devel >= 2.7.10-2 is needed by condor-7.0.5-1.fc9.x86_64
      • classads-devel is needed by condor-7.0.5-1.fc9.x86_64
    • # yum install pcre-devel postgresql-devel openssl-devel gsoap-devel classads-devel
    • $ rpmbuild --rebuild condor-7.0.5-1.fc9.src.rpm
    • # cd /home/kliang10/rpmbuild/RPMS
    • # rpm -i condor-7.0.5-1.fc9.x86_64.rpm
    • # cd /usr
    • # /usr/sbin/condor_configure --install
      • Missing release directories (looking in /usr):
      • sql
      • man
      • Unable to find some release directories.
      • Please specify the correct location of these with
      • --install=<path to release directories> or make sure it's
      • in the current directory
    • # yum install planner
    • # ln -s /usr/share/planner/sql /usr/sql
    • # ln -s /usr/kerberos/man /usr/man
    • # /usr/sbin/condor_configure --install
      • Setting up Condor in /usr
      • cp: cannot stat `/usr/etc/examples/condor_config.generic': No such file or directory
      • Unable to create condor_config file: /usr/etc/condor_config!
    • # ln -s /usr/share/doc/condor-7.0.5/examples /usr/etc/examples
    • # /usr/sbin/condor_configure --install
      • Setting up Condor in /usr
      • WARNING: Multiple network interfaces detected. Condor might not work
      • properly until you set NETWORK_INTERFACE = <interface IP>
      • WARNING: Unable to contact central manager: localhost.localdomain localhost!
      • Condor has been installed into:
      • /usr
      • Configured condor using these configuration files:
      • global: /usr/etc/condor_config
      • local: /usr/local.ireland/condor_config.local
    • # /usr/sbin/condor_configure --install --type=execute,submit,manager --local-dir=/home/condor
      • Setting up Condor in /usr
      • Condor has been installed into:
      • /usr
      • In order for Condor to work properly you must set your CONDOR_CONFIG
      • environment variable to point to your Condor configuration file:
      • /usr/etc/condor_config before running Condor commands/daemons.
      • Created scripts which can be sourced by users to setup their
      • Condor environment variables. These are:
      • sh: /usr/condor.sh
      • csh: /usr/condor.csh
    • Startup Condor
      • # /usr/sbin/condor_master
  • Submit the first Condor job
    • Write a small program in C, file name is simple.c stored in /home/condor, code is as below
      • #include <stdio.h>
      • main(int argc, char **argv)
      • {
      • int sleep_time;
      • int input;
      • int failure;
      • if (argc != 3) {
      • printf("Usage: simple <sleep-time> <integer>\n");
      • failure = 1;
      • } else {
      • sleep_time = atoi(argv[1]);
      • input = atoi(argv[2]);
      • printf("Thinking really hard for %d seconds...\n", sleep_time);
      • sleep(sleep_time);
      • printf("We calculated: %d\n", input * 2);
      • failure = 0;
      • }
      • return failure;
      • }
    • Compile the simple.c
      • $ gcc -o simple simple.c
    • Create submit file named simple.sub as below
      • Universe = vanilla
      • Executable = simple
      • Arguments = 4 10
      • Log = simple.log
      • Output = simple.out
      • Error = simple.error
      • Queue
    • Submit the job
      • $ condor_submit simple.sub
    • Check the queue
      • $ condor_q

Nov 13, 2008:

Nov 14, 2008

  • Finished the v0.2 release
  • Send a bug report to bugzilla.redhat.com

Nov 19, 2008

  • Install Fedora 9 for X86_64 on the virtual machine of China system(Named condor1)
  • Install git
    • #yum install git
  • Download the source codes of Condor 7.0.5 using git
  • Create a new branch
    • $cd mrg-grid
    • $git branch
    • $git branch -a
    • git-checkout -b steven origin/V7.0.5
  • Modify the configure file
    • $cd config
    • $vi LINUX.cf
    • Increase two lines "# elif defined(IS_GLIBC28)" and "define GlibCFlag -DGLIBC=GLIBC -DGLIBC27=GLIBC28"
    • $git status
    • $git-commit -a
  • Install Condor
    • $cd ../src
    • $./build_init //--OK
    • $./configure --disable-gcc-version-check --disable-glibc-version-check --disable-full-port --enable-proper //--Fail
    • #yum install libpqxx libpqxx-devel csh

Nov 25, 2008

    • $./configure --disable-gcc-version-check --disable-glibc-version-check --disable-full-port --enable-proper //--OK
    • $make //--OK
    • $make release