ECLIPSE

Version 2017.2

Installation Guide
ECLIPSE Installation Guide

Copyright Notice
Copyright © 2017 Schlumberger. All rights reserved.

This work contains the confidential and proprietary trade secrets of Schlumberger and
may not be copied or stored in an information retrieval system, transferred, used,
distributed, translated or retransmitted in any form or by any means, electronic or
mechanical, in whole or in part, without the express written permission of the copyright
owner.

Trademarks & Service Marks

Schlumberger, the Schlumberger logotype, and other words or symbols used to identify
the products and services described herein are either trademarks, trade names or
service marks of Schlumberger and its licensors, or are the property of their respective
owners. These marks may not be copied, imitated or used, in whole or in part, without
the express prior written permission of Schlumberger. In addition, covers, page headers,
custom graphics, icons, and other design elements may be service marks, trademarks,
and/or trade dress of Schlumberger, and may not be copied, imitated, or used, in whole
or in part, without the express prior written permission of Schlumberger. Other company,
product, and service names are the properties of their respective owners.

ECLIPSE® is a mark of Schlumberger.

An asterisk (*) is used throughout this document to designate other marks of

Schlumberger.

Security Notice
The software described herein is configured to operate with at least the minimum
specifications set out by Schlumberger. You are advised that such minimum
specifications are merely recommendations and not intended to be limiting to
configurations that may be used to operate the software. Similarly, you are advised that
the software should be operated in a secure environment whether such software is
operated across a network, on a single system and/or on a plurality of systems. It is up
to you to configure and maintain your networks and/or system(s) in a secure manner. If
you have further questions as to recommendations regarding recommended
specifications or security, please feel free to contact your local Schlumberger
representative.
ECLIPSE Installation Guide

Table of Contents
1 Introduction .................................................................................................. 1
Simulator computing configurations .............................................................................. 1
Supported platforms and software .................................................................................. 2

2 Install the simulator on a PC ....................................................................... 5

Prepare for the installation ............................................................................................... 5
Install the simulator on a Windows PC ........................................................................... 6
Install Intel MPI on a Windows PC ................................................................................... 7
Complete the installation ................................................................................................. 7
Test the installation .......................................................................................................... 7
Test using Simulation Launcher ...................................................................................... 7
Test using the command line ........................................................................................... 9

3 Install the simulator on Microsoft HPC .................................................... 11

The HPC installation procedure ..................................................................................... 11
Check the Infiniband driver firmware ............................................................................ 12
Install simulation software on the HPC cluster ............................................................ 12
Configure the HPC cluster ............................................................................................. 13
Configure cluster information on the head node ............................................................ 13
Configure the HPC web services interface on the head node ....................................... 14
Configure nodes ............................................................................................................ 15
Install the VCC runtimes on compute nodes .............................................................. 16
Define the file system parameters ............................................................................. 16
Set the Windows power configuration ........................................................................ 16

Configure PCs to access the cluster ............................................................................. 16

Complete the HPC installation ....................................................................................... 17

4 Install the simulator on Linux ................................................................... 18

The Linux installation procedure ................................................................................... 19
Configure InfiniBand drivers .......................................................................................... 19
Configure job schedulers ............................................................................................... 20
About job schedulers ..................................................................................................... 20
Configure LSF to work with Infiniband ........................................................................... 20
Configure LSF to work with Intel MPI ............................................................................ 21
Configure LSF HPC ....................................................................................................... 21
Install and configure the integration kits ........................................................................ 22
Using the cluster without a scheduler ............................................................................ 22
Install the simulation software on a Linux cluster ....................................................... 23

i
ECLIPSE Installation Guide

Install ECLIPSE in a Linux cluster ................................................................................. 23

Configure Intel MPI or Platform MPI for Linux ............................................................. 25
Check and configure ssh for use with the MPIs ............................................................ 25
Intel MPI settings ........................................................................................................... 26
Platform MPI settings .................................................................................................... 27
Set up user accounts in the Linux cluster .................................................................... 28
Install ECLRUN in the Linux cluster .............................................................................. 29
Configure the schedulers to work with ECLRUN ........................................................... 29
Test the installation ........................................................................................................ 29

A Legacy macros and scripts ....................................................................... 32

Macros: command line flags .......................................................................................... 34
Examples ......................................................................................................................... 35

B Installation DVD content ........................................................................... 36

ii
ECLIPSE Installation Guide

1
Introduction
Using the ECLIPSE or Intersect simulation software, you can create models of your reservoir or field to
match historical data and to evaluate different production and recovery scenarios. You can use the
simulators with Schlumberger programs such as Petrel or the ECLIPSE pre and post-processor applications
(EPP) to model, assemble, visualize and analyze your data.
The simplest simulator configuration is a standalone installation on a Windows computer, but if your
organization has a computing cluster, you can use it to run your simulation jobs. This chapter provides an
overview of the standalone and cluster installations and lists the supported platforms and software.

Simulator computing configurations

On a standalone system, you run the simulator software on your workstation, PC or laptop. To do this, you
install the software on your computer and set up a link to the license server which controls access to the
software. The software is supplied on a DVD and uses an installation wizard similar to those used widely
for Windows software installations.

Figure 1.1. Standalone simulator installation

In this case, all of the applications and data required for running simulation jobs are on the workstation or
computer. The licensing server(s) is usually in the network to provide an access point for all users, but it
can be installed on individual Windows computers too.
You can also use network connections to computing clusters to process the simulation jobs. In this case,
you might choose to run smaller simulation jobs on your computer and larger, more complex jobs on the
cluster. These clusters use either Microsoft Windows or Linux operating systems but share many features.
A head node or control node assigns simulation requests/jobs to the available compute nodes. To do this,
the cluster requires its own software and data which is kept on local high-speed storage. This includes

Introduction
1
ECLIPSE Installation Guide

scheduling software which controls how and when the simulation requests are processed. The cluster also
has the simulation software installed and uses it to process the submitted requests. The cluster and your
computer need to share the simulation data so that the cluster can run the simulation and return the results
for analysis.

Figure 1.2. Cluster installation

With cluster configurations, there is a great deal of flexibility in where the software components are
installed. For example, the license server could be installed on the head node, or be a separate machine in
the network. There could be a single installation of the simulation software, with the cluster starting up
multiple copies (instances) when needed, or there could be several installations across the cluster. This
allows IT professionals to choose a setup that fits with the available cluster equipment.
Installation on a cluster requires a number of steps to configure it, install the software and to set up
connections from each Windows computer to the cluster. The installation is normally carried out by IT
professionals familiar with network environments and Windows and Linux computing clusters.

Supported platforms and software

This section describes the supported platforms for a PC/workstation, a Windows HPC cluster, a Linux
cluster and for the licensing server.

Windows PC or workstation
Operating systems
Windows 7 Professional 64-bit, Service Pack 1.
Windows 10 64-bit Pro and Enterprise.
No 32-bit systems are supported.
Required software
Microsoft .NET 3.5 and Microsoft Installer 3.0
Intel MPI 5.1.3. This is used for parallel processing. It is included on the installation DVD and
you install it separately from the simulation software.
Carnac 1.2b145 to support the EPP program FloGrid. Installed automatically.
Other requirements
Minimum screen resolution of 1024x768

Introduction
2
ECLIPSE Installation Guide

Windows HPC cluster

For simulator job submission.
Operating systems
Windows Server 2012 R2 Update 2 x64.
Windows Server 2016 x64.
Requires the HPC pack.
The Windows HPC operating systems have their own Message Passing Interface (MPI) and job
scheduler.
Required software
Microsoft C runtimes V9, V10 and V11 which are in the 3rdparty directory on the DVD.
These runtimes are installed automatically on the cluster headnode when you install the
simulation software. You must install them manually the other nodes in the Windows HPC
cluster.
Optional software
The License Aware Activation Filter. You can use this to manage license allocation on a
Windows HPC cluster. It is supplied as a utility on the installation DVD in the 3rdparty
directory.

Linux cluster
For simulator job submission. Linux systems must be based on the Intel Xeon x86_64 chips.
Operating systems:
Red Hat Enterprise Linux Server 6, Update 6 (x86_64), or later.
Red Hat Enterprise Linux Server 7, Update 2 (x86_64), or later.
CentOS Linux Server 7, Update 2 (x86_64), or later.
Job schedulers/queueing systems:
LSF 9.1
PBS Pro 12.x
UGE 8.1.3
These are not supplied on the simulator installation DVD and you must purchase and install
them separately.
The installation DVD contains an integration kit for each of these schedulers which provides an
interface between the license server and the scheduler. Each integration kit enables jobs to be
started only when licenses are available and to queue jobs when no licenses are available.
Message Passing Interfaces (MPI)
Intel MPI 5.1.3.
Platform MPI 9.1.2
Both of these are installed automatically during Linux installation if you select tools. After
installation you select and configure your chosen MPI.

Introduction
3
ECLIPSE Installation Guide

Licensing platforms

Note: For the 2015 or later versions of the simulators you must install or upgrade your license server to use
Schlumberger Licensing 2015, or later (2017). This is available on the installation DVD for Windows-
based licensing and is in the Tools directory for Linux licensing.

Windows licensing platform operating systems

Windows 7 and Windows 10 64-bit
Windows Server 2012 R2 and 2016
Linux licensing platform operating systems
Red Hat Enterprise Linux Server 6 Update 3 (x86_64) or later.
Licensing software
2015 license server for Windows and Windows clusters. This must be installed separately from
the other software.
2015 license server on Linux. This is installed automatically during Linux installation if tools
are selected.

Introduction
4
ECLIPSE Installation Guide

2
Install the simulator on a PC
In a typical standalone installation, you use the simulator software with other applications to model,
visualize and characterize data, for example with Petrel which provides full field and reservoir management
tools for processing data from seismic survey through to simulation. ECLIPSE is also supplied with a
number of legacy pre and post-processor (EPP) applications to help with the construction of simulation
datasets. The simulator and associated applications are accessed through the Simulation Launcher. The
software components are shown in the following figure.

Figure 2.1. Windows installation

The Simulation Launcher and Petrel, if it is installed, send requests to a management and control program
called ECLRUN which starts the simulator or EPP application. Intel MPI supports parallel processing and
so ECLRUN interacts with it when running simulation jobs which use the multiple processors/parallel
processing available on personal computers or workstations.

Prepare for the installation

Before installing the software, make sure that you are prepared:
• Have administrator privileges on your Windows computer.
• Know the identifiers of your drives, for example C: for the hard drive and D: for the DVD drive.
• Remove previous versions of the simulator, if installed. For example, if you have a Beta version,
remove it before starting this installation.
• Temporarily disable your virus checking software while installing the software.
• Read the documentation provided you should have a PDF reader such as Adobe Reader or Foxit.

Install the simulator on a PC

5
ECLIPSE Installation Guide

• Activate .NET framework 3.5 in the Control Panel if you have not done so already for another
application. Simulation Launcher requires .NET framework 3.5 to function. Windows 10, Windows
Server 2012 R2, and Windows Server 2016 all require that the user manually activates .NET
framework 3.5. The Microsoft support website contains guidance on activating .NET framework 3.5
for these operating systems.
If you need any advice on these items, consult your IT department.
The installation DVD contains all of the software that you need: the simulator software, associated
programs and utilities, the Simulation Launcher, Intel MPI and licensing. You use normal Windows-based
installation procedures to install the software and then contact your IT administrator to connect to the
license server. Petrel is available separately.
There are several stages to the installation:
1. Install the simulator software and reboot your PC.
2. Install Intel MPI (message passing interface) to allow parallel processing and use with other
applications.
3. Complete the installation by setting up the license connection, performing some installation checks,
and running some test simulations.
You may need help from your IT department to complete the installation, for example with setting up
your connection to the license server.

Install the simulator on a Windows PC

The installer uses a standard installation wizard. In most cases, clicking OK, Next or Continue moves you
through the installation process.
1. Insert the DVD in the DVD drive.
In most cases, this starts the installation automatically. If it doesn't, use Windows Explorer to access
the DVD and double click the file setup.exe.
This displays the Install Menu window.
If you want to access the documentation before proceeding with the installation, click on Release
Notes or Installation Guide.
2. Click Install Products and then click the simulator.
3. Click on Next or Yes in each window to carry on with the installation.
4. Select the location for the installation and then click Next.
Either accept the default location of c:\ecl, or change to another directory. Note: directory names
cannot contain spaces, for example Program Files is not an acceptable directory.
5. Continue the installation by clicking Next or Yes in the following windows, accepting the default
options and leaving any settings or options unchanged.

6. If prompted leave this setting and click Finish to complete the

installation.
A reboot may be required to complete the installation, after which you must install Intel MPI.

Install the simulator on a PC

6
ECLIPSE Installation Guide

Install Intel MPI on a Windows PC

The MPI installer uses a standard Windows style installation wizard.
1. Insert the DVD in the DVD drive.
This displays the Install Menu, or you can start it by browsing the DVD for the setup.exe
file.
2. Click Install Products, click Intel MPI, and then click Extract to extract the program files. Do not
alter any selections in this window.
This opens in MPI installation wizard.
3. Click Next and step through the installation procedure accepting the license conditions.
The MPI installation wizard closes when the installation is complete.
4. After the MPI installation is complete, click Exit in the Install Menu dialog box.

Note: To complete the installation, you need a connection to the Schlumberger licensing server.

Complete the installation

After installing the simulator and Intel MPI, you need a connection to the license server. Contact your IT
department to either:
• Add/update the SLBSLS_LICENSE_FILE user variable through the Control Panel.
• Install the licensing software and set up your PC as a license server.
During installation, two system variables are added to your PC: \ecl\HOME and \ecl\MACROS in this
order. You can check the system variable settings using the Control Panel, or have an IT professional
check for you.

Test the installation

Once you have a license connection, you can test your installation by running a simulation using one of the
example datasets in your installation directory. You can run the tests using the Simulation Launcher
or the command line.
At any stage after installing the simulation software, you can modify or repair it using the installation DVD.
In the Control Panel, find the software in your installed programs, and select the change option. Follow the
prompts to install media and to change or repair selected components.

Test using Simulation Launcher

Once you have a license connection, you can test your installation by running a simulation using one of the
example datasets in your installation directory. You can run the tests using the Simulation Launcher
or the command line.

Install the simulator on a PC

7
ECLIPSE Installation Guide

Note: It is recommended that you test the installation with one of the example datasets before running your
own models.

Open the Simulation Launcher

From the Windows Start menu, select Simulation Launcher.

Test the ECLIPSE installation

The ECLIPSE datasets are in your installed directory, for example in C:\ecl\<ECLIPSE_version>
\eclipse\data for ECLIPSE 100. To run an ECLIPSE simulation:
1. Select ECLIPSE, from the Simulators group.
2. Select a dataset from the eclipse\data directory and click Run.
You can perform similar tests for ECLIPSE 300 and FrontSim if you wish, using datasets in the
e300\data directory and selecting the E300 simulator or using datasets in frontsim\data and
selecting the FRONTSIM simulator.

Test the INTERSECT installation

A sample dataset is provided with the INTERSECT installation in the C:\ecl
\<INTERSECT_version>\IX\Examples\Example_2\EX2.DATA directory. To open the dataset:
1. Select Migrator in the INTERSECT group.
2. Click Add Dataset, select EX2.DATA and click Run.
This runs the migration. Wait until that has finished before going on to the next step.
3. Select INTERSECT in the INTERSECT group.
4. Click Add Dataset, select the file EX2.afi and click Run.
The should run and end as shown in the example below:

Install the simulator on a PC

8
ECLIPSE Installation Guide

Alternatively, if you already have an AFI file then you can skip the first two steps and go directly to
running INTERSECT.
If you have an ECLIPSE dataset that you want to simulate in INTERSECT, use the Migrator option in the
same way as above to convert the ECLIPSE dataset to AFI format.

Test using the command line

Create a batch file for testing

As an alternative to running jobs one at a time from the Simulation Launcher you can create a batch
file and use it to submit several jobs one after the other. The batch file contains commands for ECLRUN to
run the simulation job(s). To do this create a file called RUN.BAT and put in the commands to run each job,
for example, for ECLIPSE:

cd \data\dir1
eclrun -v 2012.2 eclipse DATASET1
cd \data\dir2
eclrun -v 2013.1 eclipse DATASET2
cd \data\dir3
eclrun -v 2014.1 e300 DATASET3
cd \data\dir4
eclrun -v 2014.2 e300 DATASET4
eclrun eclipse DATASET5

To start the runs, double-click on the RUN.BAT in Windows Explorer. The example shows how you can
specify particular release versions, or leave out the version number to run the latest version. You can use
similar commands but with other simulators, such as frontsim, ix, and ecl2ix.

Test the Intersect installation

The Migrator can convert an ECLIPSE keyword file into an AFI file, which can then be read by
INTERSECT.
1. Use the cd command to change to the directory containing the example ECLIPSE dataset.
For example, change to the directory C:\ecl\<INTERSECT_version>\IX\Examples
\Example_2\ .

Install the simulator on a PC

9
ECLIPSE Installation Guide

2. Type eclrun ecl2ix <basename> where <basename> is the root name of the input dataset,
for example EX2.
The command generates the AFI file but does not run or simulate it.
3. After the migration is complete, run the simulation by typing eclrun ix <basename>.
If the input ECLIPSE dataset is altered, the AFI file has to be regenerated from the dataset first before you
can simulate it using INTERSECT.
An AFI file consists of {basename}.afi and its associated files basename_ECL2IX.gsg,
basename_ECL2IX_IX.ixf, basename_ECL2IX_FM.ixf, basename_fm_edits.ixf and
basename_reservoir_edits.ixf.

Install the simulator on a PC

10
ECLIPSE Installation Guide

3
Install the simulator on Microsoft
HPC
This section describes how to set up and configure a Microsoft Windows Server 2012 HPC cluster, to allow
simulation jobs to be run on it. This involves some cluster-side and PC-side configuration. An example
network is shown in the following figure.

Figure 3.1. Windows HPC installation

The installed software on the personal computer or workstation is the same as for the standalone Windows
installation and you can use Petrel or EPP to run the simulation job on your PC. However, ECLRUN is
configured with extra connection options to access the Windows HPC network to run simulation jobs. The
simulator is typically installed in a file system on Network Attached Storage (NAS) but can be installed on
the cluster head node. The simulator must be accessible from the PC and the cluster. To run a simulation
job, ECLRUN sends job requests to the HPC Client Pack. The simulation dataset is on a shared disk visible
to both the workstation and the Windows HPC cluster. The cluster head node allocates the job to the
compute nodes which process the job in parallel. On completion, the simulation results are sent to the
shared disk and the HPC Client Pack sends ECLRUN information about the job's status.
On the cluster, the Infiniband network provides high throughput communications links between the cluster
nodes. The private network contains data for the operation of the cluster.
The workstation, cluster, license server and shared disks are on a network, typically ethernet, which they
use for communication and to share data.

The HPC installation procedure

Before installing the simulation software you must have met the following prerequisites:

Install the simulator on Microsoft HPC

11
ECLIPSE Installation Guide

• Installed and configured the HPC cluster according to Microsoft's instructions.

• Set up the user accounts and cluster access.
• Run the cluster diagnostics and corrected any issues.
• Ensure that Microsoft MPI version 5 or later is installed.
The instructions in this guide describe how to:
1. Check the Infiniband interconnect in the cluster.
2. Configure the cluster nodes.
3. Install the simulation software on the cluster head node.
You must install Schlumberger licensing if this is not already in your network.
4. Configure client PCs to access the cluster.
To complete the installation:
• (Optional) Install and set up the license-aware activation filter to manage job submission to the cluster.
The software and documentation for this is supplied as a third party PC resource on the installation
DVD.
• Set up ECLRUN to submit jobs to the cluster.
You can use ECLRUN's command line options to debug and test the setup during installation and
configuration.

Check the Infiniband driver firmware

The current Infiniband (IB) drivers released from Mellanox require the firmware on the cards to be at the
latest level in order to work with HPC. In most cases you can download the firmware from the Mellanox
website http://www.mellanox.com but you may need to determine what type of card you have. To do this:
1. Download the Mellanox Firmware Tools (MFT) from http://www.mellanox.com/content/pages.php?
pg=management_tools=34
2. Follow the instructions to get your card type.
3. Then download the correct firmware for your card.

Note: It is important that you download the correct firmware or you can risk damaging your card. The
firmware needs to be installed before the IB drivers. The current generation of IB drivers (2.2.1) are
available from http://www.openfabrics.org for the generic drivers. There may be specialist drivers for
certain modern IB cards that you obtain directly from the Host Channel Adapter (HCA) manufacturers.
You need to check on this with your card manufacturer.

Once the drivers are installed check that you can see the IB HCAs in the device manager. If they are not in
the device manager, the HPC installation will not be able to use them.

Install simulation software on the HPC cluster

Be aware of the following when installing the simulation software on the cluster from the installation DVD:

Install the simulator on Microsoft HPC

12
ECLIPSE Installation Guide

• Only install simulation software. If you are installing ECLIPSE, clear the EPP components option,
when prompted for the components to install.
• Do not install Intel MPI as Windows HPC has its own MPI.

Set up the license server

Note: For the 2015 or later versions of the simulators you must install or upgrade your license server to use
Schlumberger Licensing 2015, or later (2017). This is available on the installation DVD for Windows-
based licensing and is in the Tools directory for Linux licensing.

For a new installation, you must set up and configure a license server to provide your users with access to
the software (if you are updating an existing installation the license server will already be configured).
Typically, you use Network Attached Storage (NAS) for the license server and choose a location that is
accessible by the cluster and your users' PCs.
The licensing software is on the installation DVD (insert the DVD, select Install Products, then
Schlumberger Licensing and follow the installation instructions).
After installing the software, open the "Schlumberger Licensing User Guide" which is located in the
directory where you installed the licensing software (for example C:\Program Files
(x86)\Schlumberger\Schlumberger Licensing\<version>). Use the guide to configure
the licensing server and to provide your users with access to it.

Configure the HPC cluster

HPC requires specific settings in the cluster and on PCs connecting to the cluster to run the simulation
software. This section describes these requirements.

Configure cluster information on the head node

You must set some environment variables in order to use the simulation software with the cluster. Do this
with administrator access rights on the cluster head node.

Note: The use of Universal Naming Convention (UNC) paths is vital to the current implementation of HPC
as the simulator is installed on a file system that must be visible to the head node and the compute nodes.
All of the required DLLs are installed not only in the lib directory as previously but also in the bin
directory. This means you do not have to install the simulator software on the compute nodes.

For Windows Server 2012, set the following:

cluscfg setenvs SLBSLS_LICENSE_FILE=7321@licenseserver

cluscfg setenvs ECLPATH=\\139.10.10.1\ecl
cluscfg setenvs MPIEXEC_AFFINITY=1
cluscfg setparams AffinityType=NoJobs
cluscfg setenvs MSMPI_CONNECTIVITY_TABLE=1
cluscfg setenvs MSMPI_ND_EAGER_LIMIT=2000000000
cluscfg setenvs MSMPI_SHM_EAGER_LIMIT=2000000000

The first line sets the location of the license server for the simulator and the optional License Aware
Activation Filter. The second line identifies the location of the simulator installation which must be shared

Install the simulator on Microsoft HPC

13
ECLIPSE Installation Guide

and accessible by all users and machines that submit simulations to the cluster. The remaining
configuration parameters provide information for the Message Passing Interface (MPI).
You may have to set other environment variables depending on how you provide access to your cluster:
• If your users are going to submit Multiple Realization (MR) jobs to the cluster, you need to set the MR
scheduling environment variable:
cluscfg setenvs ECL_MR_SCHEDULING=true
You must also add the following to the eclrun. config file on each client machine:
<LSFLicenses><True></LSFLicenses>
• The DefaultHoldDuration parameter specifies the time between license checks. By default, if
no licenses are available, the scheduler waits 900 seconds before checking again. A setting of 90 is
recommended:
cluscfg setparams DefaultHoldDuration=90
• To check the environment variables settings, use
cluscfg listenvs
This sets the variable for all users and all compute nodes. If these are not set, ECLRUN will return an
error. These settings require HPC Service Pack 3 to work.
• If you want to alter any of the system parameters you can find those that are available using the
command:
mpiexec –help, mpiexec –help2 or mpiexec –help3
Be careful when changing settings as they can cause unforeseen problems.
• Set the license server by IP address rather than name, although both methods should work.
• Wherever simulation data is stored it must be visible to the compute nodes for both read and write
operations. The ECLRUN program checks this before it submits a job.
For HPC the eclrun command takes the form:
eclrun –s localhost –q <cluster name> -u <user on cluster> eclipse
<datafile>
You can also use the debug option when you are testing the connection:
eclrun –s localhost –q <cluster name> -u <user on cluster> --
debug=both eclipse <datafile>

Note: The server for submitting to HPC is always localhost. This option can only be used on systems
where the HPC client pack has been installed.

Configure the HPC web services interface on the head node

This in an optional procedure to allow your users to use the Web services or REST interface to monitor
their jobs through a remote interface, as well as monitoring the cluster load through the basic Job
Manager interface. You can also set up a job submission web page for the simulation software similar to
that provided by EnginFrame. To install the web interface:
1. Install Web services components from

Install the simulator on Microsoft HPC

14
ECLIPSE Installation Guide

Program files\Microsoft HPC Pack 2012\Bin\HpcWebComponents

2. Run a Powershell Window as administrator.

Note: Powershell scripts and arguments are case sensitive.

3. Set unrestricted policy executions so that you can run scripts:

a. To check the current execution policy settings, run the command
Get-ExecutionPolicy -List
b. If the execution policy is restricted, run the following command to set it to unrestricted:
Set-ExecutionPolicy -ExecutionPolicy unrestricted
4. Change to the \bin directory:
cd program files\Microsoft HPC Pack 2012\bin
5. Enable the service portal.
If you want to use the default port number of 443, type:
.\Set-HpcWebComponents.ps1 -Service Portal -enable
If you want to use a different port, type:
.\Set-HpcWebComponents.ps1 -Service Portal -enable -Port
<web_port_number>
6. When prompted for the certificate, enter 0 (zero) for a self-signed certificate.
7. Then enable the REST interface making sure that you use the same port number as for the Enable the
service portal step:
.\Set-HpcWebComponents.ps1 -Service REST -enable
or
.\Set-HpcWebComponents.ps1 -Service REST -enable -Port
<web_port_number>
8. To restart the scheduler, type:
Net stop hpcscheduler then
Net start hpcscheduler
9. To test access to the web interface, go to another machine and type
https://<machine_name>/HpcPortal
If you cannot access the web service, repeat the setup procedure.
10. After setting up web access you should restrict execution policy again by running the command:
Set-ExecutionPolicy -ExecutionPolicy restricted

Configure nodes
To set up the nodes to run the simulation software, you need to:

Install the simulator on Microsoft HPC

15
ECLIPSE Installation Guide

1. Install the VCC runtimes.

2. Define the file system parameters (although this is not always necessary.
3. Set up the power configuration.

Install the VCC runtimes on compute nodes

You need to make sure that all the required VCC runtimes are installed on all compute nodes. These can be
found on the installation DVD in the following directories:

3rdparty\PC\resource\vcredist_vs9
3rdparty\PC\resource\vcredist_vs10
3rdparty\PC\resource\vcredist_vs11

Define the file system parameters

Windows Server 2012 (and Windows 8) use the new SMB3 file system. However, the SMB2 file system is
still widely used. If you are using this file system, and want to continue using it, run PowerShell as an
administrator and run this command (a single command line):
Set-ItemProperty -Path "HKLM:\SYSTEM\CurrentControlSet\Services
\LanmanWorkstation\Parameters" RequireSecureNegotiate -Value 0 -Force
You must run this command on all compute nodes and the head node.

Set the Windows power configuration

On Windows HPC you must set the power configuration to high performance. An incorrect setting can
have a 15% performance impact on simulation runs.
To do this, you must have administrator rights on the head node and all of the cluster nodes.
To see what configuration you have, type powercfg -l in a DOS prompt on the head node. It should
return:
Existing Power Schemes (* Active)
-----------------------------------
Power Scheme GUID: 381b4222-f694-41f0-9685-ff5bb260df2e (Balanced)
Power Scheme GUID: 8c5e7fda-e8bf-4a96-9a85-a6e23a8c635c (High performance) *
Power Scheme GUID: a1841308-3541-4fab-bc81-f71556f20b4a (Power saver)
If you do not have high performance set, you will not see the entry marked with an asterisk (*). To add this
setting, use the clusrun command in node management to run the command on all cluster nodes:
clusrun /all powercfg -s 8c5e7fda-e8bf-4a96-9a85-a6e23a8c635c

Configure PCs to access the cluster

After installing the HPC client pack on each PC/workstation, you must add an environment variable so that
Petrel and the Simulation Launcher can work. On each client PC, add:
CCP_SCHEDULER=<name of head node>

Install the simulator on Microsoft HPC

16
ECLIPSE Installation Guide

If you do not set the CCP_SCHEDULER environment variable, you see the error:

D:\>cluscfg listenvs
No connection could be made because the target machine actively refused it
127.0.0.1:5800

If you do not set the scheduler, ECLRUN and remote job submission does not work.

Complete the HPC installation

After installing the simulation software, you can optionally install and set up the license-aware activation
filter to manage job submission to the cluster. The filter software and documentation are in a zipped file on
the installation DVD in the 3rdparty/PC/resource/HPCUtilities directory. Use the
documentation to set up and configure the filter.
You must set up ECLRUN on all of the PCs that can submit simulation jobs to the HPC cluster. The
"ECLRUN User Guide" is part of the installed documentation set. To get the guide, start Simulation
Launcher, select one of the simulator manual sets and open the PDF documentation. The guide is
available from the documentation bookshelf. Use the configuration information to provide your users with
access to the cluster.

Install the simulator on Microsoft HPC

17
ECLIPSE Installation Guide

4
Install the simulator on Linux
This section describes how to set up and configure a Linux cluster to run simulation jobs. This involves
some cluster-side and PC-side configuration. An example network is shown in the following figure.

Figure 4.1. Linux installation

The installed software on the personal computer or workstation is the same as for the standalone Windows
installation and you can use Petrel or EPP to run the simulation job. However, ECLRUN is configured to
access the Linux network to run simulation jobs. ECLRUN is installed on both the workstation and in the
Linux cluster. It works with various scheduling systems to manage job submission and completions. LSF,
PBS and UGE. ECLRUN and a scheduler are installed on each cluster node. The cluster head node
allocates the job to the compute nodes which process the simulation the job in parallel. On completion, the
simulation results are sent to the shared disk and information about the job status is sent to the workstation.
The simulator is typically installed in a file system on Network Attached Storage (NAS) but can be
installed on the cluster head node for example.

On the cluster, the Infiniband network provides high throughput communications links between the cluster
nodes. The private network contains cluster data.
The workstation, cluster, license server and shared disks use the ethernet network to communicate and
share data.
For the shared disk configuration to work, the shared disk must be accessible by both the Linux and
Windows file systems. That is, the disk must support both the Windows and Linux file naming
conventions. The user id and permission mapping must work on both Windows and Linux. Where this is
not possible, there are separate data disks with the PC/workstation transferring data to a Windows disk and
the Linux cluster using a Linux disk.

Install the simulator on Linux

18
ECLIPSE Installation Guide

The Linux installation procedure

Before you install the simulator software you must already have:
• Installed and configured the Linux cluster according to the manufacturers' instructions.
• Set up storage, user accounts, and cluster access.
• Run the cluster diagnostics and corrected any issues.
To set up the cluster to support the simulators follow the steps below:
1. Configure the Infiniband interconnect in the cluster.
2. Configure the job scheduler and install the job scheduler integration kits.
3. Install the simulator.
You must install Schlumberger licensing if this is not already in your network.
4. Configure MPI settings.
5. Set up ECLRUN to allow your users to submit jobs to the cluster.
You can use ECLRUN's command line options to debug the setup.

Configure InfiniBand drivers

You must install the OFED drivers if you want to use InfiniBand. See http://www.openfabrics.org for more
information.

Note: You can also install the drivers from the Redhat installation disk. Depending on the type of hardware
in your system, you may need to download and install a later driver from the OpenFabrics web site.

General InfiniBand notes

To use InfiniBand (IB) cards, you must assign an IP address to the card. This allows you to "ping" the other
computers via the IB interface. If the ping succeeds then you know all the hardware is physically connected
properly.

InfiniBand drivers
Follow the installation instructions from the InfiniBand supplier. You may find that IP over IB is working
properly but the simulator is not. If you are using Red Hat 5 or later, you can correct this by editing the
file /etc/security/limits.conf and adding the following two lines:

* soft memlock 102400000000

* hard memlock 102400000000

The value represents the number of kilobytes that may be locked by a process. The file above contains
further documentation.

Install the simulator on Linux

19
ECLIPSE Installation Guide

Note: The steps described above will allow any user in the system to lock as much as the whole memory
set in the configuration files.

Infinipath drivers
To obtain drivers and instructions, visit this site: http://www.qlogic.com.

Configure job schedulers

This section describes how to set up and configure the job schedulers on the Linux cluster. You can use
LSF, PBS or UGE. This section:
• Gives an overview of schedulers.
• Describes how to configure LSF to work with Infiniband and Intel MPI.
• Recommends what to do if you have LSF HPC set up.
• Describes how to install the integration kits.
• Describes how to use a cluster without a scheduler.

About job schedulers

It is common practice to run simulation jobs using a commercial job scheduler (also called a queueing
system). If a single user is using a machine, it may be adequate. However, the main advantage of using
parallel processing is speed. This is lost if a machine is overloaded, as all parallel tasks slow down. This
can often create run times longer than for serial simulation runs. Job schedulers allow greater throughput of
jobs by using the available resources more efficiently.
Using the eclrun macros with a job scheduler allows you to make sure that the appropriate control
parameters are used when your users submit simulation jobs to the Linux cluster for processing.

Configure LSF to work with Infiniband

If you are using LSF, you need to edit the lsf startup file /etc/init.d/lsf to add an entry for
ulimit so that both Intel MPI and Platform MPI can run when under LSF control. To do this, open the
file and add an entry for ulimit, for example as shown here in the in the two lines starting #line
added so simulator:

#!/bin/sh
# $Id: startup.svr4,v 1.10 2008/04/08 06:13:09 xltang Exp $
#
# Start and stop LSF daemons, System V / OSF version
# Make sure we're running a shell that understands functions
#
# The following is for the Linux chkconfig utlility
# chkconfig: 35 99 01
# description: Load Sharing Facility
#
# The following is for the Linux insserv utility
### BEGIN INIT INFO
# Provides: lsf
# Required-Start: $remote_fs
# Required-Stop: $remote_fs
# Default-Start: 3 5

Install the simulator on Linux

20
ECLIPSE Installation Guide

# Default-Stop: 0 1 2 6
# Description: Start LSF daemons
### END INIT INFO

#line added so simulator can run over Infiniband when under LSF control.
ulimit -l 102400000000

if test "$SH5" != "/bin/sh5" -a -r /bin/sh5

then
SH5=/bin/sh5
export SH5
exec /bin/sh5 "$0" "$@"
fi

check_env () {
if [ x$LSF_ENVDIR = x ]; then
# Using default path of lsf.conf...
LSF_CONF=/lsftop/lsf/conf/lsf.conf

Note: LSF by default saves temporary files in a hidden directory called . lsbatch which is inside the
user’s home directory. This can cause problems if the home directories don't have much free space, or
quotas are enabled. This can be avoided by adding the following setting in lsf. conf:
LSB_STDOUT_DIRECT=Y
You must restart LSF for this change to take effect.

Configure LSF to work with Intel MPI

The introduction of the HYDRA loader from OpenMPI, caused some changes to the way jobs launch, and
this can cause issues.
To go back to the old method of using ssh you need to set the environment variable
I_MPI_HYDRA_BOOTSTRAP to ssh. This can be done in one of two ways:
under csh/tcsh using setenv I_MPI_HYDRA_BOOTSTRAP ssh in your . cshrc file, or
under bash using export I_MPI_HYDRA_BOOTSTRAP=ssh in the . profile file

Configure LSF HPC

Note: We highly recommend that you do not use any LSF extensions for LSF HPC. These can cause
problems if the versions of LSF HPC and Intel MPI are not compatible.

Recommended disabling of LSF HPC

The use of LSF HPC extensions is controlled by an environment variable ECL_LSFHPC. If the variable
exists, the extensions are invoked. It is recommended that you removed the ECL_LSFHPC variable from
any settings. This causes simulation jobs to launch using the standard and more flexible method.

Configuration changes if you require LSF HPC

If you require the extensions to support other cluster operations, you must edit the MPI location in the
wrapper script, for example wrapping the Intel MPI in the intelmpi_wrapper script. The following example
assumes that the simulator has been installed to /ecl and lsf has been installed to /lsf. It also assumes
that you are using Intel MPI. If you have installed elsewhere, please use the appropriate path.

Install the simulator on Linux

21
ECLIPSE Installation Guide

1. In the LSF installation directory (/lsf in this example), edit the file
/lsf/9.0/linux2.6-glibc2.3-x86_64/bin/intelmpi_wrapper
2. Search for the line MPI_TOPDIR="........"
3. Replace the line with the correct location of the Intel MPI. If the default settings have been used this
line should look like this
MPI_TOPDIR="/ecl/tools/linux_x86_64/intel/mpi/5.0/"
4. Find all occurrences of "$MPI_TOPDIR/bin" and replace them with "$MPI_TOPDIR/bin64"
5. If you wish to use SSH to start the MPI daemons:
a. Search for the line
MPDBOOT_CMD="$MPI_TOPDIR/bin64/mpdboot"
b. Change it to
MPDBOOT_CMD="$MPI_TOPDIR/bin64/mpdboot -r /usr/bin/ssh"

Install and configure the integration kits

The installation DVD contains an integration kit for each of the schedulers for: PBS Pro, LSF, and UGE.
Each integration kit enables jobs to be started only when licenses are available and to queue jobs when no
licenses are available. The kits are in the 3rdparty/PBSPro/resouce, 3rdparty/LSF/resouce
and 3rdparty/UGE/resouce directories on the installation DVD. Follow the instructions to set up
your chosen kit.

Using the cluster without a scheduler

If you decide not to use a job scheduler/queueing system, you can use a host file to determine how requests
are made to computing nodes to run simulation jobs. If you do not specify a name for the host file, a host
file containing the name of the current host is auto-generated. This file can then be updated using the
relevant formats:
• The standard hostfile format is of the form
Node01
Node02
Node03
Node04
This runs a process on each node in turn.
• In the case of nodes with twin CPUs, to specify that a process is to be run on each CPU, use a hostfile
of the form
Node01
Node01
Node02
Node02

Install the simulator on Linux

22
ECLIPSE Installation Guide

Install the simulation software on a Linux cluster

The software can be installed independently of previous major or incremental releases. It can be installed
over existing installations. For the installation please ensure the following prerequisites are met:
• Only use static mounts for any directories holding home directories, simulator data, or simulator
executables. Problems have been noted with auto-mounted file systems.
• Reservoir simulation software should always be installed in version order. Do not install the current
version followed by an older version.
• If you are installing both ECLIPSE and Intersect, install ECLIPSE first. For example, install ECLIPSE
20XY.Z and then Intersect 20XY.Z.
• If you do install an earlier version of the software, do not install the macros. To do this, clear the
macros option during installation.

Note: It is assumed in this chapter that the software is installed in the default location /ecl.

After installing the simulator software, you must install the licensing software if you do not already have a
licensing server in your network. You then configure the MPI settings to support the simulators.

Install ECLIPSE in a Linux cluster

1. Insert the installation DVD into the disk drive.
The DVD should mount automatically and you can continue with the installation.
2. Run the install C-shell script which is supplied on the disk:
/media/cdrom/ECLIPSE/UNIX/install/cdinst.csh
If this is successful, you are prompted to continue the installation.

The following versions were found for installation

0) [ ] Exit without installing

1) [ ] documentation_and_data 2015.1
2) [x] linux_x86_64 2015.1

Choose a version to install (1 - 2) [default : Linux x86_64] :

If this is not successful, it means that the DVD has not mounted automatically and you have to mount
it manually. To do this:
a. Unmount the installation disk using the command: umount -k /dev/cdrom
b. Mount the installation disk use the command mount /dev/cdrom /tmp/a.
Please note that /tmp/a should be an existing directory.
c. To start the installation, type /media/cdrom/ECLIPSE/UNIX/install/cdinst.csh
You are prompted to continue the installation.
3. Press enter to select the default Linux installation.
This displays a list of options and prompts you for the installation directory.

Install the simulator on Linux

23
ECLIPSE Installation Guide

Options for linux_x86_64

1) E300
2) ECLIPSE
3) Expand
4) Extract
5) FrontSim
6) ParallelE300
7) ParallelECLIPSE
8) Rename
9) Tools

Select A - Install all of the above

N - Install none of these
S - Select a subset

Please enter choice :

Enter proposed location for installation (default /ecl) :

4. For the choice, type a and for the installation directory, type /ecl. Then press enter.

Note: Always install ECLIPSE to the same location as any previous versions. There is an internal
directory structure to prevent any old versions being overwritten.
ECLIPSE for different architectures should also be installed into the same location

5. Select the install directory.

The software then installs:

Enter proposed location for installation (default /ecl) : /ecl

Installing E300 (size 15877259 bytes)
Installing Eclipse (size 16229034 bytes)
Installing Expand (size 356035 bytes)
Installing Extract (size 572208 bytes)
Installing FrontSim (size 25773064 bytes)
Installing ParallelE300 (size 33849984 bytes)
Installing ParallelEclipse (size 30258247 bytes)
Installing Rename (size 336972 bytes)
Installing Tools (size 155826405 bytes)
Installing Resource
(size 20530476)
(size 67967635)
Resource installed (total size 88498111)
Do you want to install the macros [default n]?

6. At the prompt Do you want to install the macros [default n]? type y and press
the return key.
Always answer y (yes) to the prompt unless you are installing an earlier version, in which case, do not
install the macros. The current macros directory, if any, will be backed up with a suffix consisting of
the date and time of install, for example macros. backup. 13: 24: 52. 230209.
7. If you want to install the ECLIPSE documentation and data, rerun the cdinst.csh script and select
the option when prompted.

Install the simulator on Linux

24
ECLIPSE Installation Guide

Setup the license server

Note: For the 2015 or later versions of the simulators you must install or upgrade your license server to use
Schlumberger Licensing 2015, or later (2017). This is available on the installation DVD for Windows-
based licensing and is in the Tools directory for Linux licensing.

For a new installation, you must set up and configure a license server to provide your users with access to
the software (if you are updating an existing installation the license server will already be configured).
Typically, you use Network Attached Storage (NAS) for the license server and choose a location that is
accessible by the cluster and your users' PCs.
The licensing software is installed along the Linux tools. You can get the software and the "Schlumberger
Licensing User Guide" from the tools directory /ec/tools/linux_x8664/flexlm1112/.

Configure Intel MPI or Platform MPI for Linux

Intel MPI and Platform MPI are installed with the simulator installation. No separate installation is
required. Please note the following:
• Both of these MPIs use ssh to set up the communication between nodes, so please make sure that
password-less ssh is set up between any computers that may be used to run a simulation job in serial or
parallel.
• Add some tuning parameters to the Intel MPI configuration if you are using this MPI.
• You must set Platform MPI interconnect settings if you are using this MPI.

Check and configure ssh for use with the MPIs

To set up the ssh connections, first check to see if the ssh connections are working. If they are not, generate
the public and private keys for each user and make sure that the authorized_keys file exists and has
the correct keys inside. You then update each user's ssh configuration files with the correct settings and
access permissions. Using this ssh setup means that connections to new machines are automatically
accepted (if you don't do this it means that you have to log on to every machine in the cluster before you
could guarantee that a MPI job will work correctly).

CAUTION: If you support other applications which rely on ssh to work, you may need to re-evaluate these
instructions as they could affect the other applications.

Check the ssh connections

Check that the connections are working by trying to connect to other machines in the cluster. For example,
try to connect from comp001 to comp002 using ssh with no password:

comp001:/home/user>ssh comp002
Last login: Thu Apr 9 10:37:03 2009 from comp001.geoquest
comp002:/home/user>

If the connection is successful, follow the instructions in Update the user configuration information to set
up each user's configuration files. If not, follow the instructions in the flowchart to generate the
authentication keys.

Install the simulator on Linux

25
ECLIPSE Installation Guide

Figure 4.2. ssh setup

Update the user configuration information

To create or update each user's configuration files and set the correct access permissions:
1. Either create the config file in the user’s . ssh directory, or add the line if it already exists.

comp002:/home/user>cd .ssh
comp002:/home/user/.ssh>cat config
StrictHostKeyChecking=no

2. Change the file permissions for the config file to 400 if necessary.
3. Change the file permissions for the authorized_keys to 600 if necessary.
4. Set the permissions of the $HOME and $HOME/. ssh directories so that other users do not have write
access to them.

Intel MPI settings

There are some tuning parameters which you can add to the mpiexec.conf file which can improve
performance. You can find this file in the directory
ecl/tools/linux_x86_64/intel/mpi/<mpi_version>/etc
and ecl/tools/linux_x86_64/intel/mpi/<mpi_version>/etc64

Install the simulator on Linux

26
ECLIPSE Installation Guide

or in the local work directory. For the <mpi_version>, see Linux cluster. Add the following tuning
parameters to the file:

-genv I_MPI_PIN_PROCESSOR_LIST allcores:map=scatter

-genv I_MPI_PIN_CELL core

The Intel MPI detects and uses the correct interface. Setting an environment variable will force an
interconnect to be chosen. The interconnect settings are shown in the following table.

Device Type Description Setting

OFED OFA interface is used by 10G iWarp, InfiniBand and Infinipath cards. ofa
Standard Ethernet Standard 100Mbs, 1Gbs and 10Gbs network cards. tcp
Infinipath Infinipath cards from QLogic. tmi
Table 4.1: Intel MPI interconnect settings

Set the I_MPI_FABRICS_LIST environment variable as follows:

setenv I_MPI_FABRICS_LIST ofa,tcp
This sets Intel MPI to try the chosen devices in order.
I_MPI_FALLBACK and I_MPI_DEVICE are no longer required and should be removed if set.
To print out the connection or interface type used set the following environment variable:
setenv I_MPI_DEBUG 2
For more information, see http://software.intel.com/en-us/intel-mpi-library/

Note: The latest Intel MPI HYDRA launch mechanism is a 32-bit application that starts the simulator
processes. This means 32-bit libraries must be installed for this to work. On Redhat 6, the required rpm is
called glibc-2.12-1.47.el6.i686.rpm. Follow the on-screen instructions and prompts to
complete the installation.

Platform MPI settings

The Platform MPI should automatically determine and use the best interface. If you wish to set the
interface, then it is possible by setting an environment variable. The search order for the interconnect is
determined by the environment variable MPI_IC_ORDER (which is a colon-separated list of interconnect
names). The most common options are shown in the following table

Device Type Description Setting

InfiniBand (IB) InfiniBand can be SDR, DDR or QDR. IBV
Infinipath Infinipath made by QLogic PSM
uDAPL (10g iWarp uDAPL interface is used by 10G iWarp network cards. UDAPL
Ethernet and InfiniBand) Additionally it can also be used by most types of InfiniBand and
Infinipath.
Standard Ethernet Standard 100Mbs, 1Gbs and 10Gbs network cards. TCP

Install the simulator on Linux

27
ECLIPSE Installation Guide

Device Type Description Setting

Myrinet GM based Myrinet cards. GM
Table 4.2: Platform MPI interconnect settings

Example
To set the MPI to use the DAPL device, add the following line to the user’s . cshrc file.
setenv MPI_IC_ORDER UDAPL
or
setenv MPI_IC_ORDER udapl:tcp
Using upper case for UDAPL sets the instruction to use that device or exit. Using lower case sets the
instruction to try the device and, if it doesn’t work, try another. Shared memory is always implemented if
possible. The standard ECLIPSE scripts will add a few extra lines into the top of the output files and these
will display the interconnects used.
The following extract from the FILENAME. OUT file uses TCP to communicate between nodes.

host | 0 1
=====|===========
0 : SHM TCP
1 : TCP SHM

Alternatively you can override it by using command line arguments to the eclrun script. The following
arguments force the Platform MPI to use the IBV setting, regardless of the environment variable setting.
eclrun -c plmpi -–mpi-args=”-IBV” eclipse DATASETNAME

Set up user accounts in the Linux cluster

After installation, each user's environment needs to be set up to reflect the simulator installation. A sample .
cshrc file is shown below. Edit the paths if required, and copy to each user's home directory.

#!/bin/tcsh
#Edit the line below to reflect the location of the license server.
#LM_LICENSE_FILE was used by pre 2007.2
setenv SLBSLS_LICENSE_FILE [email protected]

#Edit the line below to set the install location.

setenv ECLPATH /ecl

#Set Endian so Linux files line up with other OS's

if ( `uname` == "Linux" ) then
setenv F_UFMTENDIAN big
endif

#Source eclrunsetup if it exists, so eclrun works

if ( -r $ECLPATH/macros/@eclrunsetup.csh ) then
source $ECLPATH/macros/@eclrunsetup.csh
endif

#Source LSF environment script

if ( -r /lsf/conf/cshrc.lsf ) then
source /lsf/conf/cshrc.lsf
endif

Install the simulator on Linux

28
ECLIPSE Installation Guide

Install ECLRUN in the Linux cluster

This section provides some information on how to set up and configure ECLRUN to allow your users to
run simulation jobs on the HPC cluster. For a full description of the configuration options and settings, see
the "ECLRUN User Guide".

ECLRUN passwords
The "ECLRUN User Guide" contains a list of restricted characters which cannot be used in passwords as
ECLRUN cannot pass them to the cluster.

Migrate ECLRUN settings

Migrate any custom settings from the old configuration files into the configuration files in the new macros
directory. Settings may be stored in any of the following files:
• CONFIG. ECL
• eclrun. config
• @eclrc
Do not just copy the old files into the new macros directory. From 2009.1 onwards, eclrun. config
uses an XML format which differs from that required in previous releases.

Configure the schedulers to work with ECLRUN

License scheduling is available with the simulator and PBSPro, LSF or UGE. To install and configure the
PBSPro, LSF or UGE integration kits, see the instructions on the installation disk in the folders
3rdparty/PBSPro, 3rdparty/LSF or 3rdparty/UGE.
Regardless of your choice of scheduler, you must update the ECLRUN configuration settings:
1. In eclrun.config find the setting:
<LSFLicenses>False</LSFLicenses>
2. Change the setting to:
<LSFLicenses>TRUE</LSFLicenses>
The LSFLicenses setting is used by all of the supported schedulers.

Note: Only ECLRUN is supported. The old scripts starting with the character @ are not able to schedule
licenses.

Test the installation

The following information may be useful for a simple tests

Install the simulator on Linux

29
ECLIPSE Installation Guide

Test ECLIPSE with an example data file

To test an ECLIPSE 100 dataset:
1. Create a temporary directory, and copy an example . DATA file from /ecl/<version>/
eclipse/data/ into this folder.
2. Change to the temporary directory and run the example using the command eclrun eclipse
<filename>.
You can perform similar tests on ECLIPSE 300 and FrontSim datasets.

Run INTERSECT from the command line

The Migrator can convert an ECLIPSE keyword file into an AFI file, which can then be read by
INTERSECT.
• The following ECLRUN command line generates the AFI file but does not run or simulate it.:
eclrun ecl2ix basename where basename is the root name of the input dataset.
• To run or perform the actual simulation using the AFI file the following command is used: eclrun
ix basename. INTERSECT will read the AFI file and perform the simulation.
If the input ECLIPSE dataset is altered, the AFI file has to be regenerated from the dataset first before
being given to INTERSECT to be simulated.
An AFI consists of {basename}.afi, and its associated files basename_ECL2IX.gsg,
basename_ECL2IX_IX.ixf, basename_ECL2IX_FM.ixf, basename_fm_edits.ixf and
basename_reservoir_edits.ixf.

Test INTERSECT with an example data file

1. Create a temporary directory, and copy the example EX2.DATA from /ecl/2013.1/ix/
Example/Example_2/ into this folder.
2. Change to the temporary directory and run the example using the command eclrun ecl2ix
EX2..
You should see a simulation run and exit normally, as below:

...
REPORT Keyword migration summary:
+-------------------------+
| Fully migrated | 49 |
| Partially migrated | 5 |
| Not migrated | 6 |
| Ignored | 8 |
| Total | 68 |
+-------------------------+
INFO EX2.afi is created.
ix_tools checked in
INFO Successfully checked in license feature 'ix_tools'
INFO Migration complete. Elapsed time 0hrs:0mins:10secs (10secs)
REPORT Message summary
+---------------------------+
| Message Level | Frequency |
+---------------------------+
| INFO | 16 |
| WARNING | 17 |
| ERROR | 0 |

Install the simulator on Linux

30
ECLIPSE Installation Guide

+---------------------------+
INFO Run finished normally.

Install the simulator on Linux

31
ECLIPSE Installation Guide

A
Legacy macros and scripts

Note: We recommend you use ECLRUN to run most of the simulators and other programs. Please refer to
the "ECLRUN User Guide" for more details. The legacy macros are still released but may have limited
capabilities compared to ECLRUN.

For Linux systems, the software can be run using scripts supplied in the macros directory in place of
ECLRUN. The following scripts are available to run the License Manager:

@flexstart Start up the License Manager

@lmdown Shut down the License Manager

@lmhostid Get FLEXlm hostid for the machine

@lmreread Reinitialize License Manager

@lmstat Check license status

The default location for the following files for use with the License Manager is the ecl / macros
directory:

license.dat License file

flex.log Activity log for License Manager

The following macros are available to run the principal simulator software programs:

@e300 Compositional simulator

@eclipse Black oil simulator

@frontsim Streamline simulator

@graf Post processor

@grid Gridding and mapping

@mpieclipse Run ECLIPSE 100 simulator in parallel mode

Appendix A Legacy macros and scripts

32
ECLIPSE Installation Guide

@mpie300 Run ECLIPSE 300 simulator in parallel mode

The following macros are available to run the utility programs (those marked * are no longer released):
• Batch operations:
@fill Corner point geometry generation

• File format conversions:

@convert Converts formatted / unformatted ECLIPSE output

@convertgrid Converts between .GRID and .EGRID formats

@dos2unix Converts ASCII from MS-DOS to UNIX format

@rescueconvert Converts from older versions to RESCUE Version 19 format

@unix2dos Convert ASCII from UNIX to MS-DOS format

• File manipulation:
@expand Merge INCLUDE files into master file

@extract Extract subset of data from ECLIPSE output

@glkread Checks the contents of a Stratamodel binary . glk file

• File name changes:

@change_prefix Changes prefix of file name

@change_suffix Changes suffix of file name

@lower Lowercase file name

@rename Change ECLIPSE filename case and extensions

@upper Uppercase file name

• Viewing:
@manuals Launches Manuals bookshelf

@pdf Launches Adobe Acrobat

Note: Not all programs are supplied with the 2013.2 release; some older applications have been retired. The
scripts have been left so they can be used to run older versions.

• The following environmental variables can be set to control LSF functionality within the macros:
ECL_LSF set true if LSF is installed

ECL_LSF_BYDEFAULT If this and ECL_LSF are set to TRUE then jobs are submitted using LSF
without the need for the -lsf flag. For compatibility with ECLIPSE

Appendix A Legacy macros and scripts

33
ECLIPSE Installation Guide

Office the -nolsf flag has been added as a new option to stop nested
bsub commands.

ECL_LSFHPC This switches on the LSF HPC extensions, but it is recommended that you
do not use this extension and remove it from all settings. However, if you
must use LSF HPC make sure that you follow the configuration steps in
"Configuring changes if you require LSF HPC".

ECL_LSF_PARALLELQ defines the default queues for parallel jobs

ECL_LSF_SERIALQ defines the default queues for serial jobs

ECL_LSF_LICCHECK set if LSF is to control FLEXlm licensing and the LSF-SIS integration kit
has been installed (contact Platform Computing www.platform.com for
details).

ECL_MR_SCHEDULING set this variable to correctly queue multiple realization jobs. Follow the
instructions in the LSF-SIS integration kit.

Macros: command line flags

The Linux macros are written using c-shell scripts. It is essential that the macros directory is placed on the
user’s path; this is not necessary, however, to ensure their successful execution. The macros will
automatically detect all versions of software installed, and will invoke questions as necessary to launch the
program required.
The macros require little or no configuration to launch interactive products. However, it is possible to
configure the macros using command line flags. This is especially useful for running the ECLIPSE
simulators. The main command line flags for running the ECLIPSE 100 and ECLIPSE 300 simulators are
listed below:

-data /home/user/workarea to set default working directory, or other directory.

-file <filename> to define a dataset file name.

-help to obtain a list of available flags with brief description.

-second <Title> to define a secondary run title. <Title> should be enclosed in

quotes.

-fore or -foreground or -fg defines a foreground run (the default)

-back or -batch defines a background (queued) run

-priority defines priority of a background run

-ver or -vers or -version defines version of software to run, for example:

-ver 2005a _1 -file TESTCASE

-proc or -procs or -processors defines number of processors for a Parallel run.

-local uses the local configuration file ECL. CFG.

Appendix A Legacy macros and scripts

34
ECLIPSE Installation Guide

For parallel ECLIPSE, namely @mpieclipse and @mpie300 the following special options are
available:

-hostfile hosts.txt Configuration file for Parallel run (same as -machinefile)

-machinefile hosts.txt Configuration file for Parallel run (same as -hostfile)

-lsf For runs using LSF.

-ilmpi Option for processing parallel jobs using Intel MPI

-plmpi Option for processing parallel jobs using Platform MPI

-ibmmpi Option for processing parallel jobs using IBM POE MPI

If -procs is not set, the macro will attempt to read the PARALLEL keyword in the dataset to get the number
of processors.

-nolocal Do not run on front end when a parallel job is run.

Examples
@eclipse -version 2017.1 -local -lsf TESTCASE
@mpie300 -lsf -lsfqueue normal -version 2017.1 -local TESTCASE

Appendix A Legacy macros and scripts

35
ECLIPSE Installation Guide

B
Installation DVD content
The software is distributed on a disk which contains:
• The simulators for all x86 64-bit architectures.
• Intel MPI Runtime installation.
• Documentation in PDF (Portable Document Format).
• License server and dongle drivers.
• Utility resources which contain:
• Benchmark and datasets (ECLIPSE)
• PC resources
• Scheduler integration kit.

Appendix B Installation DVD content

36