About Us    
    Join Us    
    Contact Us    

GeoTomo USA
15995 N Barkers Landing, Ste 110
Houston, TX 77079
Tel:  281-597-1429
Fax: 281-597-1201  www.geotomo.com     

GeoTomo Hefei
Suite 601, Bldg. A3
800 Wangjiang West Rd
Hefei, Anhui, China

India Representative for GeoTomo:

Essem High Tech

196 A, Hsidc
Udyog Vihar
Gurgaon - 122 001
Haryana, India

Russian Federation Exclusive distributor for GeoTomo products in Russia:

Sovfransgeo LLC

Office X, 56A
Nagatinskaya emb., Moscow, 115407
Tel: +7(495) 982-37-27
Email: sfg@sovfransgeo.ru
Web: www.sovfransgeo.ru


GeoTomo System Installation FAQ
Revised 23-October-2015




If you have already downloaded the package, please make sure to fully read the “Quick-Start Guide” and any other documentation that might relate to your application. 

Please verify the system requirements and any additional library requirements prior to installation.  Your systems administrator should be able to assist you with installation questions, however, please email support@geotomo.com if you have any questions, comments, or concerns.



Our 32bit and 64bit line of products come with shared libraries that may or not be compatible with your distribution.  We recommend using CentOS 5.x 64bit, NFS version 3 (for non-intensive file I/O) to unify mount paths, central authentication if you are deploying a job to multiple machines, passwordless SSH or RSH, and that your users use bash or sh.  If you use an alternate shell such as ksh, csh, or zsh you might have to substitute certain commands in the scripts to correctly work with your setup (our scripts are tested in environments where the user shell is sh or bash).

Our software is designed to run with OpenMPI and MPICH2 for mutiple cpu jobs and also has support for job schedulers such as Torque or PBS to help you manage your system resources.



If you use an alternative shell or a distribution other than CentOS 5.x series you may have to install some libraries for compatibility.  Please refer to your system documentation on the correct installation procedures, just to make sure.

CentOS 4.7 and up / Redhat 5 and up

·         yum install xterm libXext numactl libnuma1 compat-f2c redhat-lsb

CentOS 6.3+                                                                                                                 

·         yum install xterm libXext numactl redhat-lsb compat-libgfortran-41                                                                          

Ubuntu 10.x/11.x/12.x/13.x/14.x

·         apt-get install lsb-core numactl libf2c2 libnuma1 xterm

*** Earlier versions of Ubuntu will require you to install gcc-3.4 libs and correctly install libg2c0 for your distribution ***


If the software shipped with a license dongle, you will need to install drivers. If you use Windows, the drivers should automatically install, unless you are using Windows XP. In that case, there should be drivers located in the license folder which should be run/installed with administrator privelages.

If the license dongle is purple, you will need to install the HASP4hl drivers for Linux. Use the appropriate link for your distribution:

RedHat / CentOS / SuSE
Download from here
Ubuntu / Debian
Download from here

If the license dongle is green, you will need to install the FlexNet drivers for Linux. Use the appropriate link for your distribution:

RedHat / CentOS / SuSE
Download from here

If you are using a Virtual Machine with a license dongle, you will install the drivers in the guest OS and disconnect the dongle from the host OS, if using Windows as the hosting operating system. For Linux, the device should just pass through to the guest operating system.

The installation process is as easy as running (for the .rpm file) "rpm --install (filename).rpm" or (for the .deb file) "dpkg -i (filename).deb". You may have to install certain required packages during the time of installation, which may be different for each machine. On RedHat/CentOS, you may also utilize "yum provides (filename)" to determine the name of the package that provides certain libraries to install with "yum". The drivers are actually 32bit, so you will need to install the i686 versions. It will tell you what packages are missing, so you can run (for RedHat/CentOS) "yum install (required package name).i686", (for SuSE) "yast2 install (required package name)-32bit", or (for Ubuntu/Debian) "apt-get install (required package name):i386".

You may need to consult your operating system documentation or contact us for assistance.




The licenses we send are network-enabled licenses, which pose an issue with stand-alone workstations that are not connected to a network. If you attempt to start the license and your machine and receive an error message from the GUI like "operation in progress" or "cannot contact license server", first verify the actual error. You may do this with running "(APPLICATION_PATH)/license/lmdown", and manually trying to start the license with "lmgrd -c (license_file).lic". If it complains about not being able to find or resolve the host in the license (usually will say "license specifies THIS_NAME, but host reports the hostname is SOMETHING_ELSE", please follow the below steps:

1. If you run "ifconfig" and there is no "eth0" or "em0" or "p7p1" adapters, but it shows up when running "ifconfig -a", you will need to type "hostname" and then edit /etc/hosts to include "(hostname)". Verify this hostname is on the first line of the license file. Make sure the network adapter is disconnected, then run "ifconfig eth0 up". You should be able to run "ping -c 1 (hostname)" and get a response from your local machine. Bring down ("lmdown") the license and try to manually start it again.If you have manually brought up the network adapter and will place it on a network, you must make sure the IP address does not conflict with other devices on the network.
2. You may edit the hostname of the license file. Anything else will cause the license to become invalid. Try modifying the license file's hostname to "localhost". The location for the hostname is in the first line of the license file. The format is "SERVER (hostname) (hardware address of network adapter/disk serial number/dongle address". The hostname "localhost" means "this machine" in layman's terms. The name "localhost" is also what you will use in the machine list when you start processing data / running projects.

If you have any problems or questions with this, please contact us for assistance.



The version of FlexNet we ship our software with relies on the network adapter to have the prefix of eth. If your network adapter is not eth0, eth1, etc, then you will need to update your version of the license daemon. Examples of network adapters that need the updated daemon are "em1" and "p7p1". You can see your network configuration with "ifconfig -a". In order to update the license daemon, please change to the license folder in the application directory, make a folder called "old" and move "lm*" and "geotomo" into this folder. You will then download and extract the contents of this archive into the license folder with the commands " cd (APPLICATION_FOLDER)/license; wget http://geotomo.com/flexnet64_11.12.1.tar; tar xvf flexnet64_11.12.1.tar". If you place the tar file into the license folder, you can skip the "wget" portion of the command. If you get "permission denied" running any of the commands, ensure you run them as the user that owns the folder/files or run it as root.

If you have any problems or questions with this, please contact us for assistance.



After the new user can successfully login to their machine, you will need to make sure they have access to the machine and ports for the license server. 

Quick-Start:  Login as that user and open a terminal in X-Windows and run:

    ssh-keygen -t dsa
  cd ~/.ssh
  cat id_dsa.pub >> authorized_keys
  cat id_dsa.pub >> authorized_keys2  **
  chmod 600 ~/.ssh/auth*

This sets up the keys, permissions, and verifies ssh to each node without a password (which is very important to deploy jobs).  After that you should be able to “ssh localhost” and verify they don’t need a password.  If you are unable to do so, please double check your steps and retry.  Also make sure the ssh service is running (“/etc/init.d/sshd start” or “/etc/init.d/openssh-server start”) and you have correct permissions on the files in “~/.ssh”. 

** This is dependent on your current setup.  Please check your system configuration and select the appropriate choice or check with the systems administrator.



Please make sure the latest video drivers, X11 drivers, Linux distribution updates (including mesa-utils and any open source drivers you use [if you don’t use manufacturer drivers]), and user profile have been installed and updated correctly.  If the machine running VNC you must connect to display :0 to render the graphics or you will get a blank screen. 

You may also try to force indirect and software rendering (depending on your shell):

** If you have an NVidia card these LIBGL flags will have no effect.

In Linux, the method for 3DViewM to work is:

GPU in the machine -> ssh –X (machine with GPU)

No GPU in the machine -> ssh –X (machine without GPU)

GPU in the machine -> ssh –X (machine without GPU) will not work correctly



If the computer’s clock gets changed after the daemon has been run, you’ll encounter this message – it is a license protection scheme.  You’ll want to check the time on the nodes you’re trying to run the software on and the computer the license server is running on.  Please contact GeoTomo’s support if you cannot remove this message by rebooting and restarting the daemon.



If the software doesn’t run on a new installation of a machine, please check with support@geotomo.com for any updated documentation and assistance on installation.

Operating Systems:

·         CentOS 5.7 & 5.8 have compatibility/shared lib issues so if you can install a different version, please do so to avoid potential conflicts.  This doesn’t affect RedHat Enterprise Linux 5 though.


Cannot access my NFS shares from my Redhat-based kernel.

There is a known bug in kernel 2.6.9 and was fixed in 2.6.18 with the way it tries to access NFS shares on other machines.  You might try NFS with –o noatime,nodiratime,vers=3,proto=tcp but it’s not guaranteed to work with all combinations of Linux distributions.  With our computing nodes we run CentOS 5.5, but utilize Ubuntu 10.04LTS on data, home, and NIS server. 


My cluster keeps losing connection to my NFS host.

If you get system “pauses” while cd or ls an NFS directory and the /var/log/messages reports something about a timeout while access the hostname of the NFS server, this will require a little tweaking of your NFS server.  If you have over 50 computers that will initiate an NFS mount to the same machine, you’ll need to have more than the system’s default 8 concurrent NFSd instances running as well as adequate bandwidth in your network. 


My NFS mount command keeps returning permission denied.

First make sure you’re allowed in the /etc/exports, the gid/uid settings of the mount, the nfs daemon is running on the machine, and that either portmap and/or other related daemons are allowed access in /etc/hosts.allow.  If you continue to get permission denied, check your kernel version, and that /proc/sys/nfsd exists.  If not, run “mount –t nfsd nodev /proc/sys/nfsd” on debian/Ubuntu or on redhat-based kernels you should check for “proc/fs/nfsd” and run “mount –t nfsd nodev/proc/fs/nfsd”.  Also note that by default, nfs uses the option of root_squash, to disallow the root user of any machine to write to those shares, a way to prevent anyone from doing things as root where they aren’t supposed to. Please make sure the appropriate person in the IT department is performing the troubleshooting, as there might be other underlying issues.


MPI job hangs or isn’t running properly.

There can be a number of reasons why your job may not run correctly.  After verifying the mounts and you’ve made sure there are no system problems, the next step to try is editing the runtime script generated (usually created in the data folder of the project and includes the word “_runScript”) and placing following text after “mpirun” but before “-np”:

--mca btl_base_verbose 30 –-mca plm_base_verbose 100

This will increase the verbosity of mpirun and include information that will assist in troubleshooting in the job’s log file.  Please send the log file, screen shots, and any relative information to support when contacting us about the issue.


Linux keeps showing very little to no free physical memory and I have plenty.

Most of the time the kernel will cache files and keep them resident for a certain amount of time, but sometimes doesn’t release the cache.  A safe method to flush the cache is to run “sync” and then “echo 3 > /proc/sys/vm/drop_caches” which writes all data to disk and then flushes it.  Please note that the next time you access a large file, the kernel will cache it again, and it may take longer than before the cache was flushed.  Check your documentation on assistance with changing the mbcache modules and settings. 


I get the error “Could not create the Java virtual machine”.

Follow these steps:

·         Verify there is enough free space on the drive (df –h)

·         Verify memory reserves / Clean up running processes

·         Change the option “MaximumHeapSize” in ~/.geotomo/config.xml to a lower number and restart the launcher


I get the error “Could not load background” when trying to start the software.

Make sure you have space available in /tmp (or whichever folder the TMP and/or TMPDIR export points to).  By default, the system should use /tmp but the systems administrator can change this.


libg2c.so not found

You may consult the (installation path)/contrib/reqlib/README file for instructions and the RPMs for RHEL4, CentOS 5, CentOS 6, or CentOS 7. We've also included the .deb package for Debian/Ubuntu systems

This library is required to run some of the Fortran code in our software. You will need to search for the RPM if you use RHEL/CentOS/SuSE and install it (as root) with "rpm -i (rpmname).rpm". For Debian/Ubuntu, please use the command "dpkg -i --force-all libg2c0_3.4.6-8ubuntu2_amd64.deb"."

To find the RPM for your specific distribution, please visit http://rpm.pbone.net and search for

Version 10 and below of Ubuntu, you can find the package in the repository as "libf2c2" and install it with the command:
"apt-get install libf2c2".